Back to: Follow the money: Apple vs. the FBI | Forward to: Some notes on world building

But it's not April 1st yet!

It's March 31st, not April 1st, so I can see no justification for the news being this weird: Ha'arez is reporting that Otto Skorzeney worked as a hit-man for Mossad in the 1960s, Microsoft just announced full native Ubuntu Linux support on Windows 10, and SCO are appealing against the IBM verdict again ...

... Oh, wait: this is a leap year. Stand down, false alarm.

(What utterly surreal symptoms of the Crazy Years have you stubbed your toes on this week—other than Donald Trump, of course?)

655 Comments

1:

The Ubuntu-on-Windows thing reminds me of AndLinux and CoLinux - I used to run xchat and pan on my old XP machine, and even now it'd be much more convenient to run that than to fire up an entire VM. Pity about the 'Win 10 only' thing, but maybe it'll rouse the other projects and they'll release 64-bit versions for those of us who prefer 7...

2:

A danger I can foresee is that Linux-on-Windows will lead to the reduced availability of PCs with unlocked boot chains -- "why do you need that? Why don't you just install it under Windows?" -- but the Microsoft OS monopoly on shipping hardware is pretty much a bust at this point in the post-PC world.

3:

I am not even surprised by any of those. It was widely stated in the 1960s and 1970s, with apparent evidence, that Mossad recruited nearly as many die-hard Nazis as the USA did. The fact that Microsoft and Ubuntu have been cooperating informally has been known for ages, and I am surprised only that Microsoft has taken so long to do a reverse wine. Lastly, SCO long ago became a pure litigation company, and it is obviously worthwhile (in a game theory sense) to appeal if it has even a few percent chance of success.

Oh, yes, they're all totally bonkers - but it's a mad world, my masters!

4:

I think all the Linux things coming out of MS are the result of the current CEO wanting MS to be a cloud leader. Which means give customers Linux if that's what they want. Reading between the lines there have been a lot of folks whose careers have ended inside of MS who were not willing to go along.

Similar things happened at IBM in the early 90s. I continually ran into people who took it as a matter of faith that Ethernet was just a bad way to do networking and would never succeed in business much less the enterprise.

NIH is almost always a bad long term bet.

5:

It's not too surprising: Microsoft Word and Excel were famously more stable as Office 2000 running in Linux with Wine than running in their intended environment - Windows 2000.

The new Windows 10 toy appears to be Explorer extensions and a reverse-Wine: and, to their credit, MS developers appear to have given very direct access to system resources - tgey are clearly usibg tge undocumented APIs that external coders never get to see.

The agenda is, of course, an attack on the concept of Open Source: but there is another, more urgent imperative.

I see this as a precursor to running Android apps directly on Windows 10 - nobody in Seattle has the flair for writing for the touchscreen UI - and they need to do this, as the Apple and Android ecosystems have far better apps, and a far bigger choice of apps, for touchscreen devices: Microsoft have fallen badly behind and interoperability may be a better strategy than isolation.

It is also, potentially, a very destructive strategy: imagine Linux and Android forging ahead with new capabilities, with a quarter of their installed application base held back by a reluctantly-updated Windows environment.


6:

Also Microsoft's SQL Server 2016 will run on Linux next year to get a chunk of Oracle's market share.

The Four Horsemen of the Dorkapolypse are approaching!

7:

Slight dissent.

Microsoft was originally in the business of writing development tools and business applications that were portable across hardware and operating systems. (I'm talking 1970s through late 1980s here.)

The doubling-down on Windows uber alles only really took hold in the late 80s/early 90s, following IBM's disastrous attempt to replace MS-DOS/Windows on the desktop with OS/2 and drag the PC platform into the whole SNA ecosystem. Microsoft kept pushing Windows, which was "good enough" by release 3.11, while IBM fell behind on OS/2, and still kept shipping DOS. Meanwhile the PC world was a lot bigger than just IBM by then, and Microsoft managed to rope in all the OEMs using ... creative ... licensing terms, and the rest is history.

Flash update: we're now in the post-PC world, and PCs are no longer the only, or largest, personal computing platform. So Nadela seems to be taking Microsoft back to its roots as application-layer "glue", using Cloud services to hook everything together.

(Balmer was too deeply invested in Windows to ever be able to adapt to the post-PC world, so his departure was clearly in Microsoft's best interests. Would Stephen Elop even have been able to plough Nokia into the ground, without Balmer's enthusiastic backing?)

8:

I was reading yesterday about a ransomware pkg that infected the MBR then faked that it was running a CHKDSK while it encrypted all the files. Apparently it was inert against locked boot chains.

9:

Slight clarification also.

At first MS was all in with OS/2. But as they saw where IBM was going to take it with our without cooperation from MS they dropped it like a stone and switched to NT with the early iterations of Windows being stop gaps on the road to NT.

It was interesting reading articles from the time talking about all these MS internal developers running OS/2 and talking about how great it was going to be when released RSN. Then it All Just Stopped.

Sad part of it was that OS/2 in so many ways was better than Win/NT back then. Especially in the area of security and malware issues. But IBM kept trying to turn it into MVS/CMS for the desktop. So it died. But at one point most of the worlds ATM were running OS/2 and no one ever talked about security issues with the software. Now they almost all run a flavor of embedded windows. Warts and all. Makes me laugh/cry when I see an ATM or airport display stuck at a Windows boot error.

10:

Err. with OR without

11:

Those reasons are NOT true. I was peripherally involved, both partially inside and out, but know a fair amount of the inside information. OS/2 was shafted by Microsoft Windows on the IBM PC, because Gates suckered IBM's lawyers; whether he planned the whole thing in advance, I can't say, but the evidence points that way. The later OS/2 balls-up on the PowerPC was entirely due to the incompetence of IBM's executives.

12:

Folks don't talk about ATM security issues relating to the software because ATMs are not connected to the internet which blocks 99% of the possible attack methods. The attacks that have been noted tend to be insider jobs carried out with intimate knowledge of the hardware and the systems user interface. As for boot screens in kiosk systems I've seen my share of Linux-like command prompts blinking away on store displays and the like.

I expect that OS/2 would have revealed multiple vulnerabilities if it had been widely deployed over the last couple of decades, much in the same way the other desktop and server OSes have. Hell, it's old enough to have buffer overrun vulnerabilities like Heartbleed because no-one took safe coding practices seriously back then.

13:

Disclaimer: at the time the IBM/MS drama was peaking (circa 1991 through 1995) I worked for a competitor -- SCO, back when the Santa Cruz Operation was a no-shit software company rather than a shambling litigation zombie. And SCO had a very decent UNIX-on-Intel desktop product in the shape of SCO Open Desktop -- SVR3.2 UNIX (upgraded internally to the point where only the copyright headers in the .h files were truly original SVR3.2; when you add POSIX and XPG/3 and a bunch of other standards to the source code you basically end up rewriting it from scrath), with X11, Motif, and IXI's X Desktop as a GUI stack, a DOS emulator, full TCP/IP networking and X Terminal server support ... OS/2 looked kind of Mickey Mouse in comparison.

Unfortunately a chunk of SCO's source base was licensed, so to keep using the UNIX trademark and other stuff the company haemorrhaged about £200 in license fees for each desktop package they sold (entry level pricing: around £500 without a compiler -- £2000 for the full megillah).

I still believe that if SCO had caught Linux religion in late 1994 they'd have ended up where Red Hat is today. But "not invented here" applies, even when half the engineering team are moonlighting on open source projects by night ... and by the time the direction was clear, it was too late, and the company pivoted and sold the no-longer-profitable UNIX IP to a bunch of shysters from Utah (formerly Caldera Systems), who had earlier purchased the festering corpse of Digital Research and milked the DR-DOS lawsuit to the tune of $500M from Microsoft, and thought IBM would be a push-over.

But pre-1995 SCO could have been a contender, if only management had been paying more attention to what their engineers were playing with.

14:

The people I know who were involved tell different stories. I suspect they are all true to a large degree but all coming from different blind men describing the elephant.

15:

The later OS/2 balls-up on the PowerPC was entirely due to the incompetence of IBM's executives.

It wasn't until Gerstner came along that every product at IBM wasn't evaluated and planned around keeping the mainframe at the center of the universe. And it was a big ship that didn't turn quickly. He was still trying to turn parts of it when he retired after 10 years.

OS/2 was seen by many high ups as a way to take control of desktops and TIE them to the mainframe. They just could not comprehend a universe where any computing that mattered was not tied to a mainframe. (Says he who was pushed hard to recommend a switch to IBM "crap for the application" offerings at the time for small businesses we sold software to. Very very hard. In the 80s.)

I still remember when IBM previewed Profs running on a 286 or 386 with OS/2 back in 1988 (summer/fall I think) where Profs required 13meg of ram when a meg cost about $500 to $1000 (this is IBM after all). And declared this was what the world wanted. It was about this point in time that OS/2 started to vanish from the scene at MS. Well sort of. MS started the OS/2 NT project in December and eventually tossed the OS/2 part aside. After extracting all the useful bits they wanted.

16:

Hey. Tesla previews the Model 3 tonight. Hours before April 1.

17:

After assurances that our online world is awesome and sophisticated, and how breaking it could only happen by fluke and/or massive brute force and/or act of god, this story happened ...

http://www.sciencealert.com/how-a-programmer-almost-broke-the-internet-by-deleting-11-lines-of-code

18:

"As for boot screens in kiosk systems I've seen my share of Linux-like command prompts blinking away on store displays and the like."

Watching the inflight entertainment system on a Cathay 747 reboot was fun. Looking over the back of the seat, rows upon rows of penguins in the darkness, like a linux army...

19:

Does a prominent advisor to an American presidential candidate accusing longstanding anti-tax zealot Grover Norquist of being a one-man sleeper cell aligned with the Muslim Brotherhood and thus, ineligible for his seat on the board of the National Rifle Association count as Crazy Years? It certainly seems to from this side of the pond, anyhow.

http://www.thedailybeast.com/articles/2016/03/24/cruz-captain-s-crazy-jihad-against-me.html

20:

Not a tech issue, but I've got tumbleweeds (a summer problem) coming up since February, when it was about 5oC warmer (averaged across the month) than the average February. Welcome to the future, it's already here.

22:

Norquist has done a lot of Republican outreach towards Muslims and his wife is Palestinian. Gaffney has been using that as grist for his conspiracies since at least Bush's first term, so his letter to the NRA is par for the course. That Cruz hired him isn't that surprising either given that Cruz himself has a bit of an attraction for conspiratorial thinking himself. That Cruz is a leading presidential candidate is pretty damn crazy though.

23:
The new Windows 10 toy appears to be Explorer extensions and a reverse-Wine: and, to their credit, MS developers appear to have given very direct access to system resources - they are clearly using the undocumented APIs that external coders never get to see.
Windows NT was designed to have different subsystems which present the external API. The "Windows" side is actually just the primary subsystem. There used to be OS/2 and POSIX subsystems too, but the OS/2 one died with Windows 2000 and the POSIX one with Windows 8.0.
I see this as a precursor to running Android apps directly on Windows 10 - nobody in Seattle has the flair for writing for the touchscreen UI - and they need to do this, as the Apple and Android ecosystems have far better apps, and a far bigger choice of apps, for touchscreen devices: Microsoft have fallen badly behind and interoperability may be a better strategy than isolation.
You have it backwards, actually. This thing came out of Project Astoria, which was designed to make it relatively easy for Android developers to port to Windows (Phone). That's a fairly terrible idea (WP has different UI semantics, for one, so any ported app would be a weird second-class citizen unless the entire UI layer was rewritten...) but it left Microsoft with a Linux compatibility subsystem.

They've been bleeding developers to Apple for years because people've been wanting a UNIX-esque workstation with a competent GUI. Now they can provide one and win back the hearts and minds of developers.

It is also, potentially, a very destructive strategy: imagine Linux and Android forging ahead with new capabilities, with a quarter of their installed application base held back by a reluctantly-updated Windows environment.
Android already holds back its application base with a reluctantly-updated Android environment. It doesn't need any help. And since when did Linux care?

Also, the userland is Ubuntu. It gets updated by Canonical, not Microsoft. I assume that other distros could do the same thing if they please. The kernel API isn't altered that much, is it?

24:

1. Linux on windows - as opposed to in a VM, reminds me of the POSIX compatability effort that M$ made in the mid-nineties with NT. And yes, I know of the presentation with Dr. Korn in the audience *snicker*

2. OS/2, which I was working with, along with Win and Unix, in the mid-nineties, was *definitely nicer than Windows. But I was reading all the stories then, and earlier, how Gates had deliberately screwed over IBM on that.

3. Grover Norquist. Yeah. Or, as I refer to him, traitor by subversion "I want to cut the US government down to where I can drag it into the bathroom and drown it in the tub", who ought to be tried for treason, and at least locked up for the next 99 years (or maybe life +20-years) for what he's done to this country.

Can we now return to the *real* 21st Century, so I can go buy my ticket on the PanAm shuttle to the Wheel?

mark

26:

I see this as a precursor to running Android apps directly on Windows 10

Like an offical version of AMIDuOS?

27:

Is that Astoria backwards or ass-backwards?

It simply hadn't occurred to me that their Windows Phone application interoperability project would rise from it's internal obscurity and unpopularity to be a *headline* project in Microsoft's prime-time consumer OS.

I can't add to that observation - and I certainly cannot contradict your point! - because this isn't actually a bad idea: it's just that it has evolved and emerged in a very strange way.

Bad ideas are the norm in IT, and neither they nor any good ideas need to come from linear chains of reasoning.

I take your point on "Android already holds back its application base with a reluctantly-updated Android environment": I'm surprised no-one else called me out on that. Further- there's a risk that MS might tempt developers (and, indirectly, users) with extensions and optimisations that 'fork' Android.

Or are there better words for Microsoft's strategy that sound a little bit like 'Fork'? As we know full well, the Internet is for Prong.


28:

I for one am thrilled that userland command line Linux stuff will run on Windows 10 without a full VM. I've been running Linux on the desktop as my primary OS since 1998, though in the last couple of years I use OS X during the day because that's what my employer standardized on for company-issued machines.

I develop some open source tools related to quantum chemistry and the programs I rely on are first-class citizens on Linux and OS X but often broken/unavailable under Windows. I'm tired of telling Windows users, i.e. most PC users including most scientists, to jump through about 5 extra hoops before they can start using my tools. (And virtualized number-crunching programs run with considerable slowdown.) In the future it should simplify to something like:

1) Install the Windows-Ubuntu system
2) Follow basically the same instructions as I give for Debian-and-derivatives users

I'll need to test that everything works, of course, but this seems like it'll be a significant improvement on ease of use, integration with the rest of the Windows environment, speed, and resource use, compared to my old "Download VirtualBox and install this desktop distribution..." instructions for Windows users. Linux won on servers and embedded and mobile devices, probably won't ever win on the desktop. That's fine. I'm more interested in making my tools widely usable than in converting people to a different OS-religion.

29:

I won't intrude on international relations in terms of "crazy" stuff this year. A lot of the events which happened in international development aren't so much crazy as "racist assumptions you grew up not questioning which you found out are bunk". Examples include Mexico/Brazil/Venezuela being more developed than China (which creates second order effects on US immigration), Latin America being more developed than Southeast Asia, Latin America being more developed than the parts of Eastern Europe not in the EU, etc. (I discovered all of those when I was doing that summary of Latin America, but I had to cut them out in the interest of finishing those pieces). You guys might roll your eyes, but those assumptions do figure a lot into Donald Trump's rise.

For space, the crazy stuff includes the possible Dyson Sphere (http://www.ibtimes.co.uk/alien-megastructure-nasa-believes-dyson-sphere-likely-be-comets-orbiting-kic-8462852-star-1530673), and the 9th planet ice giant discovered.

In aviation, the crazy is the electric plane (http://www.acmp.com/blog/the-electric-cessna-what-is-happening.html). I know it's been around for a while, but I only discovered it recently. Same for the electric helicopter. The possible return of supersonic passenger travel with Boom also might qualify.

30:

I was not referring to that stage, but the next one, in my remar about executive incompetence. I was close enough that I know something about both the contracts and software ancestry, and was personally involved in some of the original (IBM) development that ended up as Microsoft Windows' GUI and Motif. Yes, they had a common ancestry.

31:

IBM realised that they had been shafted c. 1992, and the period you mention was when they were desperately trying to do something, ANYTHING, with a completely one-sided contract. The fact that IBM's lawyers had allowed themselves to be suckered by a smelly geek did NOT make the technical staff happy. My understanding was that SCO took its copyright assumptions and pricing policies from AT&T, when it bought the trademark, so was (at least initially) a victim as much as a culprit.

32:

The SCO of the zombie litigation is not the same SCO of OGH's era.
The original SCO had by the time of the first lawsuits morphed into Tarentella and has sold its Unix interests to Caldera (at the time a reasonably well respected Linux distro aimed at the consumer).

Something happened to its parent that seems to bring in some 'new management' who came in and saw what it thought was some easy money suing the 'dirty thieving linux geeks' .

Why they thought attacking IBM would be a good idea I've no idea except for the 'pump and dump' antics on the stock market.

33:

the original (IBM) development that ended up as Microsoft Windows' GUI and Motif. Yes, they had a common ancestry.

Common User Access, right?

While Windows releases prior to Win95 sucked in a myriad of ways, the uniformity of the user interface owed a lot to CUA and took some of the pain off it; you could do just about everything except graphics-intensive tasks using just the keyboard, for example, and the core accelerator/shortcut/function key bindings were consistent across all applications (a level of internal consistency that they ditched in Win95). It wasn't as slick-looking as the MacOS HIG, but it didn't depend on a mouse (because it embraced character-only displays as well as WIMP environments).

And yes, Microsoft Works 3.0 for DOS and Microsoft Word 5.5 for DOS both followed the CUA and were wonderfully responsive to a keyboard-head (despite their other shortcomings: poor/non-existent style sheet in the one case, lack of a macro language in the other).

34:

The final movement in Tarantella's tale was, as far as I understand it, to become a pure applications company and then to be absorbed by Borland.

Caldera was originally established as a Linux distributor circa 1994 with money from Ray Noorda (founder of Novell), but they burned through their initial capital too fast and when Novell wanted a Linux distro they bought SuSE. Along the way Caldera bought Digital Research, who were suing Microsoft for locking Win95 so that it only booted atop a (concealed) DOS OS kernel and refused to run atop DR-DOS -- unless DR-DOS was tweaked to report its identity as the MS-DOS kernel. They won big, so obviously decided to pivot into litigation when they couldn't compete in the Linux OEM marketplace.

Suing IBM was not the smartest business decision they ever made.

35:

Borland. That's a name I've not heard in a while. So I went over to Wikipedia. Wow I had forgotten most of my memories of their mergers and such and never knew about 80% of them. Tarantella isn't even mentioned.

Sadly it reads like most of the companies that exploded in the early days of the PC industry. Most never realized that at some point it was going to be an enterprise game for the big boys and gradually MS crushed most all of them.

https://en.wikipedia.org/wiki/Borland

And eventually they were bought by Micro Focus which in my mind is always "Micro Focus Cobol".

36:

Sorry, I was distracted; what was the link between Mossad and Microsoft, again?

37:

I switched over to using xkbcomp to load a mousekeys.xkb so I can hit numlock and switch it from a numpad to my mouse, with a trackball backup off to the side. Playing dwarf fortress made me really reevaluate how I used my keyboard when I wasn't playing.

As for weird future stuff.

The nazi-loving anti-feminist far-right-wing (US right wing btw) teenage twitter AI popped up accidentally and spammed another 200k posts about smoking kush. http://www.ibtimes.co.uk/microsoft-tay-ai-returns-boast-smoking-weed-front-police-spam-200k-followers-1552164

Take a moment to look at that sentence. "Nazi-loving anti-feminist far-right-wing teenage twitter AI" is a thing now.

38:

Oh dear.
Just saw a news report:
Trump's ancestral homeland bewildered by his presidential campaign

Mostly old news, but hadn't realized he's MacLeod of Lewis on his mother's side. Hope he's not related to Ken (at least not closer than distant cousins).
Don't know if video will play over there, but there's the article.

39:

Is it terrible that within 1/10th second of looking at the picture I recognized the distribution (or at least thought I did) as some variant of Mandrake?

40:

Ken MacLeod? Let's hope he's not related to Connor MacLeod, otherwise we could be stuck with him around for a long long time...

41:

Same or different Richard Gadsden on London Reconnections?

42:

As far as I can make out, both Gaffney & Norquist are fruitcakes, but this is the USA ,,, so what does it matter - or does it matter?

43:

Oh dear though
Ubuntu available through Win10
Means that getting Win 10 on my "new" machine (If I ever get round to it) will be running Win10.
This machine is getting OLD & I need to build a new one, except - where do I start, too much choice, too many motherboards & main chips to choose from ....
[ Yes, I do mean build - buy the bits & plug it together - a LOT cheaper & I get what I want, I Hope ]

44:

[ Yes, I do mean build - buy the bits & plug it together - a LOT cheaper & I get what I want, I Hope ]

Alas, unless you're going the Raspberry Pi route, build-it-yourself PCs are now either pretty much integrated boxes like the Intel NUC, or they're monstrous high-end gaming rigs with components that cost the earth and lots of blue LEDs inside the perspex-walled tower case so you can watch the chromed cooling fans whir. Seriously, laptops are the cheap end of the PC market these days (them or all-in-one desktops that resemble an iMac).

A lot of the stuff you used to need -- serial/parallel ports, floppy disk drives, CD/DVD drives -- are obsolescent and can be hung off the end of a USB controller. Modern monitors don't even take VGA input, and most of the stuff that used to be internal isn't any more.

45:

Unfortunately Charlie is right, much of the DIY market has gotten screwed by the rise of the phone/tablet.

My suggestion is to look towards something that can act as a NAS AND run VMs for your various usages. That way you have some flexibility towards Win/OSX/Linux, run from whatever front end suffices at the time. Can be made to work, even if you are into gaming.

The alternative is to run a laptop in semi NUC mode as your main machine, but really, laptops are on their way out too.

2-3 years down the line and it'll all be VR/AR interface.

46:

"Common User Access, right?"

Yup. Most of my involvement was at the 'full screen' (3270) stage, but I still had a bit when it expanded into a GUI form.

47:

Oh, the Australian PM decided to launch his April Fools prank on the Australian people a day early - suggesting the states pick up the power to levy income tax again, and the federal government would therefore stop funding public schools (while still funding private ones).

Or at least, that's how I'm choosing to take it, because the reality is entirely too upsetting to face.

(On the other hand, as an April Fool's prank, it certainly managed to fool the Premier of my state, which is all to the good as far as I'm concerned, given I don't trust, admire, respect or even like the man[1]).

[1] About 20 years ago as Education Minister, our current Premier sold off 3 suburban state high schools in the western suburbs for Ma$$ive Profit$ to developers. Wind the clock forward, and all of a sudden a population boom in the western suburbs is leading to a shortage of available high school places there. Demographics shift over time, who knew?!?

48:

"The SCO of the zombie litigation is not the same SCO of OGH's era."

I blame AT&T, in why and how they set up Unix System Labs, actually. But it is a long and sordid story, and it's OGH's blog.

49:

Actually, it goes back before AT&T to the Department of Justice and the AT&T anti-trust lawsuit! If there'd been no long-drawn-out DoJ anti-trust litigation, there'd have been nothing preventing AT&T from just selling UNIX directly, or distributing it for free as an academic research program, or whatever. Although on second thoughts, "whatever" would probably have involved avaricious grasping licensing terms like unto the way they used to rent telephone handsets. You can't win.

50:

Of the 6 or 8 USB ports on my current, ancient tower, three have died, & the ethernet internal connection is long dead ... I have had (always) only hardware failure with it (Dell 9100) I don't want a laptop & my screen & keyboard & sound are perfectly OK.
I "just" want a new box with two drives (one hard, one solid-state) a better double DVD read/write & a robust, large-capacity, reasonably fast board + chip + power supply - probably a new graphics card.
Um
Where do I start looking, because as I said the bewildering array of chips & boards loses me these days ....
And a tower costs about £400 anyway, so I might as wel build my own ... P.S. I'm still running WinXP ....

51:

Sorry, I'm so out-of-date I did not understand many of your acronyms ( I go back to real core-store, remember & punched-card Fortran 80-chars per line ... )
NAS? VR/AR?
I assume VM is "virtual machine?

52:

NAS = Network Attached Storage = there's an ethernet card between your computer and the harddrive

VR = Virtual Reality = what you see is not what there is
AR = Augmented Reality = VR on top of real things

53:

Wow, thanks for that Skorzeny article; it fits right into a writing project of mine.

Perhaps not weird, but signs of the future continuing to roll out: Tesla announces its family car ( http://jalopnik.com/tesla-model-3-this-is-it-1768284734 ); future/potential Bond villains Elon Musk, Richard Branson and Jeff Bezos are all competing to send tourists into space; and astronomers have touted a way to hide the Earth from potential invading aliens ( http://mnras.oxfordjournals.org/content/early/2016/03/30/mnras.stw672 ). OK, that last one definitely leans toward the weird.

54:

If that's what you want, consider a NUC box or similar as long as it's got USB 3 (or better, USB 3.1) ports. USB 3 supports 3gb/s read/writes, so you can stick an external HD on it and expect SATA-like speeds. Put the SSD inside the PC for maximum speed, but unless you go for a PCIe drive an internal SATA-II connection will be barely faster than running over USB 3. If the micro-PC has two internal USB 3 controllers (I'd be astonished if it didn't) you can stick your DVD/RW drive on t'other one, along with stuff like printer/keyboard/mouse that barely twitches the bandwidth.

The video card ... separate video boards are basically aimed at gamers these days: they're hair driers with a thousand-core parallel supercomputer attached. Unless you plan to get heavily into gaming you don't need a separate graphics card. The on-chip Intel integrated graphics subsystems that come on laptops and in things like the NUC today are comparable to a high-end 2010 gaming GPU.

If I were you, I'd look at something like this -- it brackets your budget, your requirements, and it'll save you a boatload of electricity compared to a tower.

55:

Greg, cheap desktops are very much a commodity item now, but still provide a pretty expandable platform. Depending on what you want to do, you could find a relatively cheap desktop and add the components you want quite easily. Plus, if this is a consideration, it means someone else has installed and configured the OS, although you're probably stuck with Win 10.

56:

NVM, Charlie's suggestion @54 is probably better.

57:

The doomy replies from Charlie and Ian S do not accord with my experience in putting together the PC I'm writing this on. (All parts ordered from Amazon, using cheapest-first search result sort order.)

The biggest problem was the utter cocks of sellers who list something universal like a memory stick once for each possible thing it fits, so you get page after page of listings all for the same identical item. I counted 148 consecutive pages of identical listings once, which was not an unusual occurrence. And with Amazon not giving an option to skip forward by more than one page at a time, it can be seen that the search needs to be conducted in an environment where it is not a problem to swear continually and hit things.

The second biggest problem was to find a graphics card that was not the size of a brick and costing more than all the other parts put together. I ended up having to go for a second hand one. (Geforce 8400 for about 15 quid.)

Motherboard selection was basically a case of choose a CPU first, then pick the cheapest board that fits it and has enough of the required types of slots and external interfaces.
Monitor is the same 21" CRT I was using previously. (I do have another of the same as a spare in case it goes pop.)
Floppy drive I didn't bother with because I can't remember the last time I used one. I do have older machines with them in so I'm not stuck if I ever do need one.
Serial port is on the motherboard.
Parallel port is on a PCI card (because I cocked up choosing the motherboard and thought it had one when it didn't). It is a proper ports-mapped-into-I/O-space port, not one of those horrible things hung off a USB controller.
DVD writer I forgot to get, haven't missed it yet, but it'll only be 20 quid odd when I do get round to it. Or less than half that for a SATA/IDE converter so I can use the one out of the previous machine.
Case is a plain ordinary steel thing without any blue LEDs or transparent bits. Came with PSU included.

Really, it was pretty much the same kind of task as it ever was.

58:

A fair point. I hadn't thought of that aspect.

59:

I don't know who the UK component sellers are, but in the US http://www.newegg.com provides well-sorted lists of components at competitive prices. If nothing else, you may be able to do your research there.

60:

Ebuyer and Dabs are in the UK, and I've always had good service from them.

61:

Thank you very much, all of you ... but ...
WHERE DO I START?
I will need a chip & a motherboard & there seem to be 200 varieties of each.
Once I've got a suitable, compatible pair, the rest is plug-it-together, which I can handle quite easily & hopefully then load in (probably) Win10 & then transfer all my "stuff" across.
If I look at a popular mass-market shop, like PC World - it's a rip-off, they want to sell you stuff I don't want or need [ Keyboard, mouse, screen, sound-card, modem/router ]
I want LESS stuff, so it should be cheaper ...
[ Though that novataech miniPC looks interesting, I must say. ]

Pigeon, can you please e-mail me direct?

I think an off-line conversation might be productive..
"fledermaus" [AT] "dsl" [DOT] "pipex" [DOT] "com"
If you re-assemble that you should be able to reach me.

62:

2-3 years down the line and it'll all be VR/AR interface.

AR/VR are definitely Aprils Fools (as I think you are suggesting :-) ).

I think that VR and AR will at best diversify and extend the PC space in the same way that smartphones/tablets have but, as with smartphones/tablets, will not blow away the basic PC functionality in many fundamental areas, especially for content creation.

And both technologies are still in their commercial infancy.

VR: still no commercial killer app even though the new devices are getting better and better. Also, I’m not aware of any standardized 3D input devices let alone gesture languages for sensibly interacting inside the 3D world. I.e. touching, selecting and otherwise activating the objects that you can see in the (sometimes sweaty, nausea-inducing) goggles. The key words here are ‘standardized’ and ‘interactions’. Unless the interactions are standardized, it will be very hard for developers to create cross-platform VR apps and for users to switch between devices. The worst case is a VR world that is as cryptic to general users as snapchat is to PC diehards aged > 13.

AR: much more promising than VR in my opinion, even given the commercial non-event that was Google Glass. Smartphone cameras and appropriate image recognition apps provide a basic AR platform, even if you do have to look through the phone / tablet screen to see the overlay. Still not sure what the killer app is though.

63:

... astronomers have touted a way to hide the Earth from potential invading aliens

I can't help feeling that any aliens who can actually send probes across interstellar distances will have alternative means of detecting planets, and may regard this kind of hiding as a signal of paranoia. "Oh, look, another one. Diddums, isn't it sweet, the way they think they're hiding?"

64:

There are even AR apps that do something useful, for 3D IKEA catalogue values of useful.

65:

Hey, I didn't say it would work. It does seem like someone's read some Saberhagen, though (but our history of broadcast radio probably dooms us anyway).

66:

Okay, how about "VR is still an April fool, AR is the story that you aren't really sure about"

67:

I'll go with that.

68:

Greg, my last couple of builds have started from a bundle on https://www.scan.co.uk/todayonly and added in whatever else is needed. eg todays (Friday) item 6 has Case + PSU, HDD, CPU + Motherboard, the MoBo appears to have onboard graphics so just add memory, OS and a CD/DVD if you want one for a running system. The selection changes every couple of days, sometimes you may get memory instead of the HDD or a graphics card or whatever. The items in the list are links to their descriptions and they give reasonable tech specs.

69:

Well working for SCO would explain how you described Santa Cruz so well in the Laundry Files.

Sigh I miss living there.

70:

Can't speak specifically about on-line UK retailers, but the larger online electronics stores in the US have a category called "Barebones" which is a minimal, preconfigured computer that you can add compatible stuff to from a menu. This way you can reuse old parts if they still work.
I buy one every several years. The most important aspect by far for longevity of usefulness (besides going with a motherboard manufacturer that has been around for a while) is memory; double or quadruple what you think you will need, and the machine will be (excepting failure) useful a few years longer. Also, research whatever you use for storage for any reliability issues. And if you want to put some sort of Linux on it, find a success report in the comments/reviews for the item.

Personally, I've been going with very low-power machines for anything that might stay powered on. That used to mean (and maybe still does) Atom processors. There may be other choices now and underclocking used to be possible, maybe still is.

71:

The most important aspect by far for longevity of usefulness (besides going with a motherboard manufacturer that has been around for a while) is memory; double or quadruple what you think you will need

Seconded. Don't even think of running Windows 10 plus a light to medium application or two in less than 4Gb of RAM; ideally 8-16Gb will let you run stuff like Firefox, LibreOffice and Scrivener simultaneously, but Win10 (hell, anything from Windows Vista onwards) is a memory hog, and by hog I mean a show-quality 500 kilo monster.

Linux can probably get by in 4Gb but you're less likely to be running photoshop ...

72:

How about transparent wood?

http://www.wired.co.uk/news/archive/2016-03/31/transparent-wood-building-walls-solar-cells

Next year - transparent steel and time travelling whales.

73:

Novatech do some decent bundles of CPU+motherboard+memory, as well as slightly more complete 'just add HD' bundles, and I've built my last couple of machines around those very happily. You can also buy similar bundles from Scan, but my experiences with Scan's customer service dept have been absolutely awful - if there's any kind of problem, you'll have to fight hard to get a replacement or refund (whereas Novatech were brilliant the one time I bought a faulty component from them).

74:

Yes.

For Reasons I am running a 32-bit XP machine here (yes, I know). That means it's got 3GB visible.

Run up an instance of Chrome, one of Thunderbird, and one of Eclipse, and the poor thing is on the edge of thrashing already.

My laptop with 64-bit Windows 10 has 6GB of RAM, and is vastly less prone to thrash.

My home desktop with Windows 7 and 18GB of RAM is ... well, I've not run into memory issues yet, and that's playing serious games like Elite Dangerous. I only tend to close stuff when the room on the taskbar gets too tight.

(Also, I've had good experiences with Novatech. But it's a while, so they might have gone to hell and I'd be unaware of it)

75:

I reckon (for Greg's enlightenment) that it's reasonable to say that modern 64-bit machines gobble RAM like there's no tomorrow.

A Windows 10 box with 4Gb is effectively a machine with 1-2Gb for applications to run in. An 8Gb machine will be fine, but one of those cheap tablets with 1-2Gb, even if the headline clock speed on the CPU is reasonable? ROFL.

This isn't just Windows. Mac OSX 10.11.blah really wants a minimum of 4Gb to run and 8Gb to have enough elbow room. A modern 64-bit Linux desktop isn't going to be much smaller.

I suspect a chunk of it is down to the amount of memory applications (and OSs) with a 64-bit address space allocate -- probably double that of a 32-bit OS -- but a chunk is just sloppy design; 64-bit iOS seems happy running apps in 1Gb (for OS and single running app plus multiple suspended).

77:

Being aware of Linux's habit of using all the memory it can to cache stuff in, I decided to give it as much as possible, and so built this box with 32GB. I was amazed to find I'd managed to give it more than it wants, for the first time ever. top is currently reporting about 8GB total used, which is fairly typical.

Although it did save my bacon when some code got stuck calling malloc() in an infinite recursion, by giving me enough time to twig what was happening and kill the process before everything turned to treacle...

78:

iOS "suspended" apps are not swapped out. If the system needs the memory they are occupying, it just recovers it, and expects them to recover from saved context when the user switches to them and they get quietly restarted.

79:

Same. And also @po8crg on twitter and various other places.

80:

I suspect a chunk of it is down to the amount of memory applications (and OSs) with a 64-bit address space allocate -- probably double that of a 32-bit OS -- but a chunk is just sloppy design; 64-bit iOS seems happy running apps in 1Gb (for OS and single running app plus multiple suspended).

It's rare that a 64 bit application will double its memory allocation size compared to the same application running as 32 bit code, but every native pointer is twice the size with 64 bit code. So if you have an application that needs to manage a large number of small objects, like a web browser, you can end up using significantly more RAM to perform the exact same task in the respective 32 and 64 bit versions. When x86-64 CPUs were just becoming common I once had an opportunity to test some graph storage code that was almost nothing but integers and pointers to integers. It had an almost perfectly doubled resident memory footprint going from a 32 bit version to 64.

If you have more time than money, and no individual application of yours really needs more than a 32 bit address space, the way to get the best of both worlds is to explicitly install 32 bit applications on your 64 bit OS. You won't pay the extra pointer size overhead and the OS can still manage the full 8 GB or whatever to spread it fairly among multiple 32 bit applications. But for anyone who's not pinching pennies I repeat the suggestion to build a new machine with the maximum supported amount of RAM installed from the beginning.

81:

I reckon (for Greg's enlightenment) that it's reasonable to say that modern 64-bit machines gobble RAM like there's no tomorrow.
Yes
I note that quite a few, not-too-expensive (for certain values of ...) full spec "complete" are quoting 1Tb of RAM, as normal, which instantly tells you something ...

82:

top is currently reporting about 8GB total used, which is fairly typical.

That'll be true until you do something to use the rest of the memory. E.g. VMs as sandboxes to play with possible malware (note: this occasionally isn't enough isolation...), or VMs running some other distro, or something else that is memory hungry.

83:

Greg, you could do worse than this list of bits I bought just over a year ago:

Item: EVGA 430W Fully Wired 80+ Silver Power Supply - Qty: 1
Item: Corsair 8GB (2x4GB) DDR3 1600MHz XMS3 Memory Kit - CL9 - Qty: 1
Item: Asus Z97-A Socket 1150 VGA DVI HDMI DisplayPort 8-ch audio ATX Motherboard - Qty: 1
Item: Cooler Master K-series K380 - Usb3.0 Atx Case Side Window - Qty: 1
Item: Kingston 120GB SSDNow V300 2.5inch SSD - Qty: 1
Item: Intel Core i3 4160 3.60GHz Socket 1150 3MB L3 Cache Retail Boxed Processor - Qty: 1

for under £400. You don't need to know what the acronyms mean, just look for something with the specific ones. I'm entirely happy with the performance so far, and adding more ram would be simple enough. I re-used an old hard drive and DVD writer and graphics card which I had bought the previous year for 30 or 40£, I don't do gaming, too much else to do, including post on here.

Note the motherboard is compatible with better intell processors and the 4th generation ones, and also has Usb 3 capability and appropriately fast connectors.

84:

Forget the TB HDD unless you are big into storing video. A SATA3 SSD 480GB will set you back about £100 and is seriously fast. I boot into Win10 about 10 seconds after exiting BIOS
http://www.ebuyer.com/store/Storage/cat/Hard-Drive---SSD/480GB

85:

I note that quite a few, not-too-expensive (for certain values of ...) full spec "complete" are quoting 1Tb of RAM, as normal, which instantly tells you something ...

Either you mistyped "1Gb", or they're quoting you disk capacity. Because right now 1Tb of RAM would cost you quite a bit -- 8Gb sticks go for roughly £25, so you're talking on the order of £3000. Whereas a 1Tb hard drive is a commodity item costing maybe £25-40 (I've lost track).

86:

The Microsoft/Canonical thing is, honestly, the least surprising of the weird Microsoft PR things from the past two weeks. After all, wind the clock back to the early 90s and they were hyping up how NT had UNIX compatibility (in the form of a dll which had a handful of wrappers for posix syscalls that called out to somewhat similar win32 calls) and shipping with a korn shell (which turned out to be nonsense). Turn the clock back to the moment Ballmer left and you see Microsoft spending a lot of effort to float the idea that they had an internal group working on Linux support and were considering doing legitimate open source releases in the near future.

No, the more interesting thing is that, according to recent press releases, they want to go whole hog into conversational interfaces, despite having a twitter bot made by a team who had apparently never been on the internet? Or, more likely, the bot had no filters in order to create the media event surrounding it being taken down (with the microsoft developer conference later that week). Still, everybody sort of forgets that as t approaches infinity all conversational UIs become command line interfaces.

87:

Thanks for the link to the Skorzeney/MOSSAD story. Fascinating. It was a long time ago I read Use of Weapons, but I couldn't help but think of Cheradine Zakalwe

88:

As for why the Mossad would cosy up with a notorious Nazi, this may explain some of it: http://tinyurl.com/z73suqd

89:

Because right now 1Tb of RAM would cost you quite a bit -- 8Gb sticks go for roughly £25, so you're talking on the order of £3000.

[£3000 ~= $4,300 in Real Money]

Which reminds me of a conversation I had with some other folks around 1985. If NSA had a need for massive amounts of fast RAM (in the context of the conversation that seemed plausible) and had a budget for equipment procurement in the low $G/year range, what could it buy?

92:

Sorry, I'm so out-of-date I did not understand many of your acronyms ( I go back to real core-store, remember & punched-card Fortran 80-chars per line ... )

Yep, VM = virtual machine.

It really depends on how you use things, but often today people have 'home' computers that do a variety of tasks. There might be a conventional desktop for knocking up text in Word, maybe a games console, certain a phone/tablet, and often a box for storing media, files, backups etc. (NAS). There might also be something for serving media, VoIP, maybe even a personal cloud, downloader, etc.

In short, it's not the 1980s. The home has a network.

One way to reconfigure to address this is to explicitly have a 24/7 server that does most of the grunt work in one box (usually the NAS), with other user interaction devices pulling on, or controlling, it. Those UI devices might be cheap desktops, laptops, tablets, chromecasts, etc. You don't need to ladle storage capacity into these, since the bulk files are elsewhere (and thus available to all UI devices).

So much, so obvious.

You can also, however, have that server beefy enough that it can run VM sessions of the OS to deliver to even less beefy UI devices (they are basically graphics terminals, remember those) running VNC/RDP etc.

One advantage is you can run your WinXP OS as a VM, spooling up and accessing it as desired, at the same time as having Win10 as a VM, or OSX, or Linux - all switched on the same server hard hypervisor. Not as simple as it should be (OS manufacturers like to believe they are the centre of the universe) but very flexible once it's running - it's like having multiple computers, all at your fingertips, on whatever device you want to access it from. You can even 'snapshot' a VM state as a backup, etc. if you are about to do something very stupid, or access the VM on the move (with appropriate security).

About the only time that doesn't work well is playing games, where most people either have special hardware (consoles) or expose the bare machine graphics to the VM on the actual machine. Probably not an issue for you though.

93:

I really depends on the application, as the interconnects, power requirements and so on are significant unavoidable overheads. Often clusters of commodity/cheap machines are the best/most efficient option unless the budget is big enough for true supercomputers to be considered (even when just small installs - a rack/ half-a-rack) .

94:

I think either that's a typo or it's disk-capacity.
Not proof-reading well-enough on this April 1st, obviously ....

95:

That was amusing.
(Am perhaps too easily amused lately, due to trying to always be in a good baseline mood.)

Given declared electricity inputs, it is possible to do simple back-of-the-envelope calculations (or at least Fermi estimates) about some potential capabilities of such a site given technology available at a large scale (i.e. probably COTS).
Also, it superficially seems like it would be hard to hide heat dissipation (as a cross check to declared electricity inputs) from IR-imaging satellites, without a way to covertly dump waste heat e.g. underground water flows, either natural or artificial. (Just stating obvious stuff here, I don't have any access to classified information.)

Has this been discussed anywhere in detail (in public)? (Anybody?)

96:
I think that VR and AR will at best diversify and extend the PC space in the same way that smartphones/tablets have but, as with smartphones/tablets, will not blow away the basic PC functionality in many fundamental areas, especially for content creation.

I'm not so sure. It will take time, but there are some interesting developments happening, and you know how fast things can switch in the IT game.

The reasoning is two fold. First, the storage, and even processing grunt, is getting virtualised - accessed online from 'somewhere', usually in the cloud. Local capability requirements are falling and coalescing around the UI functions.

Second, any of the UI devices are a contradiction/balancing act between wanting as big a screen as possible, and wanting something you can carry easily. Hence the segmentation and watch>phone>tablet>laptop>desktop progression. You'd really like to do away with that contradiction - and AR provides a way.

Put the screens up at the eyes, track the hands and environment via cameras, and you can provide a BIG interaction space in a small package (glasses or contact lens in the end) using the virtualised storage/processing to do the heavy lifting.

For content creation it presents new avenues. The interesting thing will be when they will virtualise the keyboard. Force feedback at the fingertips for tactile sensation seems to be moving forward.

How long will it take? Difficult to know, but you can see the big boys being interested in it, even at this stage. The VR focused Rift etc. are really just simplifications of the probably future AR/VR world where the difference is just how much of reality you let through.

As for killer app? I can think of quite a few - VR conferencing/meeting being one that others are already looking at : http://www.extremetech.com/extreme/225779-holoportation-could-be-the-technology-to-supplant-smartphones

97:

Alas, unless you're going the Raspberry Pi route,

And in the vein of the title, here's a post from yesterday that is just that close to being realistic and believable. Apple getting into IoT via a RaspPi clone.

http://hackaday.com/2016/04/01/apple-introduces-their-answer-to-the-raspberry-pi/

98:

Forget the TB HDD unless you are big into storing video.

My music library is nearly a TB — I can listen for weeks without repetition if I want to.

My photo library is somewhere north of 6 TB. That's stills, none of your new-fangled moving pictures in it :-)

(I shoot lots of HDR panoramas, so 60-100 40MB files stitched and processing into a 1 GB image, Easy to rack up the GB that way.)

99:

Actually, SCO had a British subsidiary. Tigran Aivazian worked there and ironically, added much of the functionality to Linux which SCO later sued IBM about.

100:

One advantage is you can run your WinXP OS as a VM, spooling up and accessing it as desired, at the same time as having Win10 as a VM, or OSX, or Linux - all switched on the same server hard hypervisor.

IT security hat on

Yes, that's a nice way, especially for those older system that don't need that much (network) connectivity. The WinXP support has ended and so it's a pretty unsafe system to run on the Internet, but running it in a VM (with a reference copy to return to if something turns sour) is a good solution.

I have an XP virtual machine which never connects to the network, for running some old software. I rarely boot it up, mainly because most games have already been updated to more modern systems, and I don't use the Lego MINDSTORM software that often anymore.

I also ordered yesterday 16 Gb more memory for this computer - when I built it three years ago I thought 8 Gb would be enough, but I seem to be running virtual machines more often now than at that time, so memory is one thing I can cheaply upgrade this thing with. An SSD would be the second option, but I don't want the hassle of re-installing (or even copying) the OS, not right now.

101:

where do I start, too much choice

'What do I want to use it for?' should be your question.
That determines what you will spend.

I'm not into gaming or video-editing and all that so it doesn't have to be a high end machine so I bought an AMD Athlon 5350 processor, € 45, with a Gigabyte AM1M-S2P motherboard, € 36, and 1 Tb harddisk for €48. I already had the other parts so it cost me €129 + shipping. It runs Slackware Linux and does everything I want.

If you are into things that require the biggest processor and you max out on memory and want a huge SSD and the newest graphicscard the price changes a bit.

Choices, choices, choices!

102:

Really , I thought the main things TSCOG (The SCO group , not SCO ) sued over was from Dynix (by Sequent) particularly the Read Copy Update parts.

Turned out the majority of the 'copied' code was a fresh reimplementation from something else (was it AIX or something more esoteric? can't remember)

IBM bought Sequent sometime in the 99/00 years , one of Sequent's offices were across the road from my old workplace in Weybridge.

General PC discussion:
As for building your own computer , I just got a local shop to put together something for me with a specified budget - decent but not bleeding edge gfx , 8g RAM , 2 TB HDD (still spinning platters , SSD was/is too expensive) , in a decent but still cheap Antec case . 1 year on no worries on performance.

One caveat though , I did get a win 8 dvd but asked them not to bother installing as I was going to put Ubuntu on it - they did anyway which was a bit annoying having to wipe and reinstall. Must be sure to tell them what you want exactly.

On Desktops in general ()plus some ramble) :
a Dying format simply because they are bulky ugly things , sticking lights and curves only really adds lipstick, but the advantage of being able to upgrade components is its major selling point and for those of us that, like me buy a PC once every 6-7 years is handy to upgrade the bits and bobs as needed.
My wife's laptop (Tosh Tecra R9000, my 15 year reward from said company) is an utter wreck after 3 years of daily punishment, back and forth. bumps and scrapes - HDD and OS has been replaced once (grr to w7 home not pro but wasn't there to make sure it was done correctly). So that will be a grand or so to replace -a £200 Asus lump will not do, this was a decent machine even 3 years ago and I'm not going to buy crap, it really won't last. Pity Tosh are pretty much out of the laptop market in Europe, their top range kit was excellent, the other stuff not so much.

Oh a bit of a ramble there , sorry.

103:

Desktops are a dying format because big corporations are trying to kill them. They want everybody to "upgrade" to less powerful but portable (read 'constantly replaced') "devices" that require dependence on outside hardware, aka the cloud, broadband, etc...More money to the Man, less power to the People. They don't want you to own a computer, they want you to own a terminal and then they want to rent you access to a computer.

104:

And look how that's working out for then laptop and tablet sales are plummeting coz these days even build for $50 in China is good enough for years of use.

Not that I buy your conspiracy theory - if it was a conspiracy they would be making money from it, and even Apple and Google PC and tablet revenues are dipping. IPad sales down, Chromebooks a flop.

Personally on my last replacement cycle I drank the Apple koolaid and bought a Mac mini from EBay and loaded it up with ram and an ssd. I got the older model with a DVD drive.

Add in Parallels or Fusion VM software for an amazingly slick Windows Office on Mac experience and I'm laughing. Ironically Windows MS office and Visio in a VM is more stable and usable than the UI horror fest which is Office for Mac.

105:

It certainly didn't start that way - the convenience of a laptop was the paramount thing in my experience - Tosh did desktops for about 30 nanoseconds in the late 90's and actually as corporate desktops worked - easy to get at to replace parts , slim and actually surprisingly attractive. HQ gave it about 18 months then pulled the plug (a common theme in the company history)

When the aspirational format at home for a computer became a laptop then the race to the bottom began and so did the death of the desktop. When I started my current job I was surprised that there were NO laptops , anywhere. In 2013. All desktops and weird boxes that connected to citrix (shudder).
Everything else is the result of this race - laptops that are sold at cost and many cases less than cost just to get numbers for the old incumbents to keep the factories running for the more profitable areas is obscene as it just fuelled the race to bottom.
The tablet is just a recognition that the vast majority just use the computer for facebook and email. inevitable really - a device that you just charge and go.

106:

Actually, SCO had a British subsidiary. Tigran Aivazian worked there and ironically, added much of the functionality to Linux which SCO later sued IBM about.

Not a "British subsidiary" but the EMEA headquarters and development center, with about 200-250 developers, QA engineers and techpubs people. We were responsible for core UNIX development and maintenance (compiler tools were handled in Toronto and Open Desktop integration at HQ in Santa Cruz). In 1994, SCO acquired IXI in Cambridge, who previously sold the Motif/IXI desktop for various UNIX workstations, to tightly integrate IXI desktop into OpenServer. But no, it wasn't just a subsidiary: it was the second-largest development site in the multinational ... until they shitcanned it entirely when things turned bad, a year or three after I left.

107:

On the one hand, that April Fool was really plausibly done and described -- if Apple did go that route, that's how they'd do it -- except on the other hand the whole idea of Apple going that route is preposterous! (The closest they'd get might be to license the A8 or A9 -- but not the latest A9X -- processor to large white-goods manufacturers along with a stripped-down version of iOS to render their appliances Apple home network native. But I suspect these days they'd be more likely to bring out Apple Car or Apple Kitchen themselves, with the goal of disrupting an entire mature global industry. Encouraging home tinkerers to get close to the bare metal isn't really their thing, despite support for CS education via things like iTunes University and the various playpen toy IDEs in the App Store.)

108:

Desktops are a dying format because big corporations are trying to kill them.

Also, because not that many people want them.

For work, you want to replace the computer every few years anyway, and for the support to be easy. Many jobs also require the computer to be portable, and many organizations are still (or again) using servers - so the personal computer is replaceable and should be easily carried. That is useful for a laptop.

For home use, many people don't need the desktop. I have a two-monitor desktop setup, for gaming, development and artistic stuff, but for many tasks I use my Android tablet. It's good for web browsing, some gaming and a slight case of IRC, and much more portable than even the laptop. Still, not that many people I know *need* the desktop. Consoles are nice for gaming, and even dedicating the desk space for something you don't use that often is something people just don't want to do.

Laptop, you can just store it somewhere and take it out on the dinner table or the living room table when needed. My father just bought a decent gaming laptop for image processing (for about 1100 €), so it's not even about the power. Extensibility is easier for a desktop, but I bought this one three years ago and only yesterday ordered more memory for it. (I don't do that much high-end gaming.)

109:

On the subject of upgradability: I have a soft spot for the current retina iMac (like the year-old one I'm typing on right now). It's about as internal-upgrade-hostile as a machine can be -- it's glued together, the only internals you can get at are the RAM sockets -- but that doesn't matter so much.

What you're getting is basically an extraordinary display with a computer attached. (A puny quad-core 4GHz i7. Nothing much, in other words. Ahahahaha.) The display itself is a 27" panel with 5120 x 2880 pixels (aspect ratio is like a 4K quad-HD display with another quarter stuck to one edge, so you can edit 4K video full-screen without obscuring it with your tool palettes). It's driven by -- in my case -- a Radeon R9 M295X with 4Gb of VRAM, and the colour gamut is good enough that it does 10 bits per RGB channel rather than the usual 8. Yes, if you absolutely must have the most bleeding-edge video card then this machine won't help you, but it's good enough for design work, video, and games currently on the market.

But, but, it's not expandable!

Well, yes and no. I can't slot PCIe boards into it or upgrade the internal HD. But it's got a pair of Thunderbolt 2 controllers, and Thunderbolt 2 encapsulates 3 PCIe channels and displayport, and was designed to provide a bus that could drive external video controllers. As it is, I can stick a couple of extra 4K monitors either side of the iMac (if my desk was large enough) and it'll drive them. Or I can stick an external SSD on one of those channels and saturate it, unless it can keep up with sustained 20gbps read/writes (hint: that's a bit over three times faster than SATA-II maxes out). Oh, and I've got four USB 3 controllers (one of them driving a 7-port powered hub for the low-speed peripherals like the 5Gb backup hard drive) for things that need to be fast, like the Blu-Ray drive I added (no, Apple doesn't support Blu-Ray out of the box; you have to -- gasp -- go looking for third-party drivers and software).

Okay, so this machine is fundamentally limited; it only takes 32Gb of RAM (and it's maxed out). However, my wife's iMac came out six months later and has 64Gb because you can never have too much memory for Photoshop.

Here's the thing: when your peripheral i/o is faster than PCIe, you don't really need PCIe (except for the odd specialised bit of legacy hardware). The only upgradability weakness I can see with this machine is that -- memory limit on my desktop aside -- nobody's selling additional TBolt GPUs for iMacs as far as I can tell.

Now, will it be working in six years time?

Well, I don't really care. It's a business machine and a depreciable asset. It's on three year AppleCare cover. The plan is to run it for three years, then retire it, grab my wife's machine (which is faster and has double the RAM and a newer GPU), and buy her a new one (she's part of the business). Then cascade. This one will make a kick-ass bedroom TV set. In 7-9 years' time, who knows what we'll be needing? What I know is that Moore's Law is grinding towards a halt and clock speeds are inching up by single-digit percentage points annually, not doubling, and having an integral design that's really easy to plug peripherals in and out of without a screwdriver seems to beat the old three-box model.

PS: $WIFE's old Adobe workstation was a 2010 Mac Pro, upgraded to the point of being a FrankenMac. 24Gb of RAM, all the drive bays occupied with ridiculously small spinning platters (the biggest is a 1Gb unit), a 512Gb PCIe SSD card configured with one of the angle-grinders as a DIY Fusion Drive (logical volume with file migration so that most-frequently-used stuff ends up on the SSD and dead files migrate out to spinning storage) and a second video card. Over time it has become flaky as hell to the point of only booting intermittently. When I get some spare time and a rush of money to the wallet I propose to: (a) strip out all the drives except the SSD and replace with the biggest fast server-grade disks I can afford, (b) yank the space heater/video card, (c) check for duff RAM and probably repopulate to max it out to 64Gb (all it'll take), (d) reinstall from scratch and configure it to boot entirely from the SSD and act as a household file server connected to the household GigE backbone. Tasks: serving up backups and VMs.

110:

Something that surprised me recently was 65" 4K monitors/TVs under £1000. Might be quite a lot of fun for gaming, which is something PCs will always be best at. And as for Moore's Law, it is still alive and well and living on peripherals like GPU cards (not to mention neuro chips).

111:

Workflow-wise, I will note that my job is to write books, not sysadmin my own home network.

One chunk of the book-writing project relies on a piece of software called Scrivener. Scrivener is not a word processor, it's a specialized tool designed for writing and managing the process of writing books, principally novels but also screenplays and scripts. It brings a bundle of non-word-processor tools to the table, like metadata and tagging features: Scrivener is to Microsoft Word as an IDE like Eclipse is to a regular text editor. I've been using it for about a decade and have written about ten books using it and leaving it for a new platform is a case of pry-it-from-my-cold-dead-fingers.

Another chunk of the book-writing job is checking the copy editor's work. Because managing editors go with the "normal" industry workflow they assume everybody either prints everything on paper and uses a pen, or can cope with a change-tracked Word document.

And a third chunk is checking page proofs. These used to come on paper and require me to go over them with a red pencil, but nowadays they come as a PDF ... and require me to return an annotated PDF.

Checking copy-edits is something you can do on an iPad right now in Microsoft Word (or, if you like living dangerously, in Apple Pages or a third party rival like Documents to Go or Textilus). Once upon a time iOS devices couldn't really do that kind of thing, but now? While you can't do complex style-based word processing or write macros in word processors, they're good for the 95% of tasks that don't involve that.

Checking page proofs on an iPad ... yeah, well, A4 page images are a bit painful to stare at on a 9.7" portrait screen, even if it can display on the order of 300dpi, and if you go landscape it involves a lot of scrolling. Also, the previous bluetooth styli were laggy and not terribly good. But the iPad Pro changes that. Not the I-can't-believe-it's-not-the-iPad-Air-3 one they introduced last week, but the big-ass 12.9" iPad Maxi. It's half the weight of a regular laptop (it weighs as much as the original iPad back in 2009) but it can show you an A4 page, full-sized, at near laserprint resolution. And the Pencil is non-laggy: Apple did some funky stuff to speed the refresh rate on the iPad Pro screens near to the tip of the pencil, just to make everything appear to happen instantaneously.

That leaves Scrivener. And Scrivener for iOS goes into beta test next week.

What does all this mean?

Well, if I wanted to dial the time machine and go back to writing like it's 2001 I could have moved my workflow entirely onto the iPad last year. But I didn't want that. As of next week, though, it'll be possible to do not just some but almost all of my work on an iPad Pro. (And it's a new accounting year, so ...)

NB: my official reason for getting an iPad Pro is that I now have a schedule for the new Merchant Princes trilogy that means, in conjunction with my other books, I have to check the page proofs on at least two books a year for the next three years. Being able to do markup work on a full-sized display with a decent pen is sufficient justification in its own right, before you add the question of whether the tablet is finally going to replace the laptop in my life. But I can see that happening. Because: I mentioned that household server? One option would be to load a VM on it just to run full-grade desktop Scriv for the odd task not supported on the tablet version, while keeping all my work projects in the cloud (specifically, on Dropbox). Eventually, it's just a big-ass tablet and a personal cloud/server in a closet. Tablets, unlike laptops (which are basically last-generation desktop PCs, packaged for portability), are designed to operate in a hostile network environment. They do less, but they do it reasonably reliably in a threatening network environment that would have been considered ludicrously hostile by a pentester from the early 00s. And frankly, I can do without the risk of my next novel being held to ransom by Cryptolocker -- and so can you ...

112:

"Desktops are a dying format because big corporations are trying to kill them.

Also, because not that many people want them."

Actually, plenty of people do, but most of those can use one of two alternatives without much hassle, so it's an extremely tough marketplace and the vendors go where the pickings are easier.

The larger laptops have taken one chunk of the market, where people are prepared to put up with the limited capacity, limited performance, lack of self-checking, poor extensibility and poor keyboard ergonomics in return for compactness and portability.

Small servers have taken up another, where people are prepared to put up with exactly the converse trade-off. I am one of those, because I insist on ECC and similar. And a well-configured small server and good screen costs LESS than even a moderately high-end laptop!

113:

No tumbleweeds here, but today broke the record for an April maximum temperature by 2.2 C. Not 2.2 over average, 2.2 over the highest on record. I don't even live in the Arctic, just coastal Australia. I know that the weather where you're standing means nothing, but somehow I still find it startling when the record falls by such a large margin.

114:

The pictures are kind of a give.away: neither the device nor the screens are over-designed. Just not shiny enough. Remember that Apple:
- professionally design the inside of there packaging
- professionally designed its Black Screen of Death on OS/X
- professionally designed the diagnostics software that the Apple shops use (I've seen it when I looked over the shoulder of a technician when my HD died)
- for resetting a mobile device, don't rely on a ordinary paper clip but have an equivalent item designed.

For a Raspberry-like device I'd expect at least rounded edges.

115:

"I reckon (for Greg's enlightenment) that it's reasonable to say that modern 64-bit machines gobble RAM like there's no tomorrow."

Rule of thumb for 32bits vs 64bits is 30% higher memory consuption for C / C++ and similar languages.

This is one of the reasons why Visual Studio is still à 32bit program. As long as they do not NEED x64 they will stay x86 (they are currently using a second process to store odds and ends linked to syntax highlighting and checking).

64bits only shines for servers programs and for OSes to break the 4Gb barrier, it's basically useless in userland (Chrome : 32bits, Firefox:32 bits).

116:

Apple professionally design the inside of their employment contracts. I'm not kidding; I've seen unboxing videos from credible sources.

For an RPi workalike from Apple, what I'd expect to see would be a shiny breakout box with the connector ports, charged and controlled by a lightning cable from a Mac, and programmed/controlled using a version of XCode, much like iOS devices. There'd be an on-Mac emulator, code signing (including self-signing, for a hobbyist device provisioned by cable rather than via an app store), and there'd be a neat cover to conceal the GPIO pins from anyone who didn't want to see that stuff. Oh, and it'd run Swift.

Alternatively, it might be powered/charged by lightning, but programmed/controlled via Bluetooth from a sandboxed-up-to-the-eyeballs mini-IDE running on an iPad -- but in such a way that inputs to the device were effectively airgapped from the iOS controller (it'd obey an iron rule that code goes from the iOS host to the embedded device, but nothing except diagnostic data comes back).

117:

Don't know if you can use this

http://winmerge.org/

It's useful for modding source code.
Basically splits the screen between two versions of a document and highlights the differences. There are probably similar functionalities in more common software for power users, though.

118:

I can trivially run 32 bit Firefox out of address space, and Google states that 64 bit Chrome has several advanages as discussed here:

http://blog.chromium.org/2014/08/64-bits-of-awesome-64-bit-windows_26.html

119:

1. It's Windows based. (How many times do I need to shout, I am allergic to Windows before the message sinks in?)

2. Yes, there are equivalent programs on other platforms.

3. It only works on text files. Publishers do not work on text files; their workflow assumes Word documents and PDFs. Uaing the XML conversion plugins would be very iffy -- WordXML is notoriously eccentric and there's a risk of losing some information in the process.

4. I am not expected -- in fact I'm forbidden -- to merge the copy-editor's changes into the CEM; I'm expected to check, comment on, and where appropriate countermand them (or make my own additional adjustments, with change tracking, so that the managing editor and the typesetter can see them separately).

5. In any case, the novel-writing workflow is linear. Only one pair of hands ever works on it at any time -- unless it's a collaborative novel. (I've done that once in my entire career.) There's simply no-need for me to ever do a diff-merge on a novel. If anyone ever does it, it's the managing editor before they send the results for typesetting -- and they insist on doing everything inside Word, because IT policy at multinational publishing empires tends to be a little bit sticky.

120:

"How many times do I need to shout, I am allergic to Windows before the message sinks in?"

An infinite number. Using Microsogware causes cognitive defects, one of which is the inability to recognise an outside world.

121:

I used to be a big wheel at Groklaw under my real name, and I researched this extensively. There were a number of SCO and Caldera programmers who made contributions to Linux. The other name that's worth researching in this regard is Chris Hellwig, who was a Caldera employee rather than an employee of "Old" SCO, but the doctrine of "unclean hands" probably would have applied in either case had the whole thing gone to trial.

When we did the research on what code SCO had contributed to Linux there were several other SCO programmers involved, but they'd mainly made minor changes to the kernel or fixed a bug or something. Tigran was the one who'd developed whole subsystems, including several of the items SCO was suing about, and of course Caldera had been heavily involved in all aspects of Linux,

122:

Sorry, it occurs to me that I never really spoke to the specific issue you brought up. Essentially, SCOG was throwing mud at the wall and hoping something would stick, but their initial claims (and everything that followed) turned out to be crap.

SCOG (and Caldera) had both previously been involved in multiple efforts to improve both Linux and Unix, including the same, specific, enterprise-level improvements SCOG was suing over. The one time they publicly showed code it turned out to be from a textbook (or something similarly stupid... sorry, it was a long time ago) and then they claimed it was "sample" code. The whole lawsuit/publicity effort/pump and dump scheme, as it came from SCOG's side, worked at the level of "It was sample code. No dude. Really!" and never went much beyond that.

As far as I know, and I followed it closely until Groklaw shut down in 2013, SCOG never came close to proving that there was AIX or Dynix code, or any code that came from Old SCO via IBM in the Linux kernel. On the other hand, it was depressingly easy to prove that individual SCO programmers (pre-SCO's purchase by Caldera) had made major contributions to Linux, with the approval of their supervisors (see my links above) and of course Caldera's programmers had been heavily-involved in writing enterprise-level Linux code, so if the suit had gone to court SCO would have been pretty much blown to bits by the estoppel issue. (And IANAL, just to make it official.)

123:

" a TB — I can listen for weeks without repetition "

more like a couple of years, music discs from libraries and radio shows I save to mp3 usually take a million bytes a minute, a million minutes would be... a hundred weeks! Play music for five or six hours a day and it'll be ten years before you hear it again.

124:

A CD is just shy of a GB if you rip it losslessly — say 4 CDs in 3 GB.

And I just realized that iTunes stores TV shows in the Music folder, so it's more like 700 GB of music and 300 GB of anime and documentaries.

Still enough that I don't get bored :-)

If we're moving to a hardware discussion: I'm trying to move my photo processing from the 2007 MacPro to the new iMac, and would welcome some advice.

1) Can anyone recommend a good storage device for my photo library? 6 TB right now and will grow. Ideally something that automatically backs up (Time Machine apparently can't handle directing difference source disks to difference backup disks.) And at least as fast as the SATA drives I have in the MacPro (I'm guessing that means Thunderbolt).

2) Can anyone recommend a decent graphics tablet? I'm currently doing quite nicely on the MacPro with a Wacom Bamboo, but the Retina display is so much larger that a pixel on the tablet is 405 pixels on the display and it's too jumpy. I don't want to spend more on the tablet than I spent on the computer!

125:

One helpful site is PC Part Picker; it can do a compatibility check on the things you're choosing, has some build guides for things other people have built, and links to places where you can buy the parts.

Not perfect, by any means, but useful if you already know what you're doing (give or take), but are ~5 years out of date on what's worth buying.

126:
Can anyone recommend a good storage device for my photo library?

I'm pretty concerned about my media library, and since I can't have ZFS on the Mac (grumble grumble grumblecan't talk about it), I went with a G-Tech Thunderbolt RAID box. (Four slots, currently configured as RAID-1 with four 2TB drives. RAID is done in hardware, which is necessary, since I hear AppleRAID is dead to the world. Eventually, I'll have to replace them with 4tb drives, but I can do that in-place.)

The other option you can do, if you have good wired networking, is use a NAS box. Something like FreeNAS can do NFS and CIFS/SMB, and can handle far more than 6tb.

Best graphics tablet for the Mac currently appears to be the iPad Pro w/Pencil.

127:

While it's more effective to reinstall the OS onto SSD, you could put in some convenient size and use it for the storage for some of your VMs, or maybe move /usr (if you're running a Linux-based hypervisor as opposed to ESXi.)

Somebody else commented that SSDs were too expensive, and he's using 2x1TB drives instead - you'll still get a huge performance win if you move the OS to an SSD, even if it's just 32GB. In some cases, for a desktop, you can also get a speedup by taking a good flash USB stick and putting the OS on that, since $40 will get you 32-64GB, and it's a lot easier than reinstalling the disk-based OS.

128:

Also because (as an engineer at a big corporation), we've moved our personal compute needs into the "cloud" (in our case, our private datacentres, where an engineer can book out whatever size of machine they need, from a 24 core, 64 GiB RAM small server to a massive machine with oodles of RAM, lots of CPU cores, GPUs etc).

Given that I have no need of intense local compute, and some people benefit from mobility, it's better value to give everyone laptops, have monitors/keyboards/mice etc at the desks, and accept that some of us never use the mobility; it means less work for IT (who now only deal with laptop configurations, so we can move a problem PC to the helpdesk to show them the issue), and it means that if you do need mobility, you just unplug the display, USB and power connectors and wander off - no need to arrange a loaner laptop as we used to do when desktops were the norm.

129:

Just put all the media onto HDDs and everything else on the SSD

130:

Thirded. My OS is on a 120GB SSD, and photos and stuff on a normal one. Boots up really fast, I don't look at the photos all the time anyway. (And I back them up onto an external hard drive, although probably not often enough)

131:

A CD is just shy of a GB if you rip it losslessly — say 4 CDs in 3 GB.

Possibly if you're dumbly storing it as the original bit stream, but I find you'll get a 3:1 compression ratio if you use FLAC instead. FLAC is still lossless.

Our music collection is 228GB as FLAC, transcoding to (highest bit-rate rate) MP3 as 91GB

132:

Tornadoes destroyed buildings in the county where I live in three of the last twelve years, so I got to worrying about that, or a fire, wiping out my thousands of hours of music and tens of thousands of pictures collected over the decades. "The Cloud" seemed like a possible solution but then I thought, why not just burn it all to dvd and bury it in the backyard in a coffeecan. Holding up okay so far, if rust sets in I can replace it with one of the airtight plastic Maxwell House containers. Couple bricks over the top and it's trivial to insert a new disc every few months, total invested so far about ten bucks; peace of mind, priceless. Long as the New Madrid Fault stays dormant I should be fine, otherwise my worries will likely be all over with anyway.

133:

Well, apple do have homekit, so not totally out of the question. And if apple wanted to get back to their roots it would make some sense. It would also be a way of nabbing the kiddies early and could integrate with their ipads for added pocket dipping.

Still not sure what the BBC Micro Bit is for though, are microcontrollers really good for learning?

134:

I was gonna suggest pcpartpicker if it hadn't been yet, the compatibility filters make it very easy to get a working collection of parts and it even lists possible edge-case issues that can arise.

Not sure what the conversion to tea-and-crumpet-land money is but a $50 motherboard, $50~75 cpu, $50~75 worth of ram sticks, $20 of case (or $45~50 for a case with a basic power supply), and $40~75 of hard drive will be more computer than most people need. Tossing in a $40~75 graphics card will let you play most games that are more than a year or two old at surprisingly high settings to boot.

135:

64 bit security is noticeably stronger than 32 bit, especially in cases where you're using something awful like windows.

There are literally no solid benefits to remaining with 32 bit software when the option to use 64 bit is available.

Any illusory increases in ram sucking boils down to modern programs being designed around larger amounts of ram. Not the 64 bit version being so much hungrier than the 32 bit one.

Plus the hard limit of 2~3 gb of ram available with the old systems got people used to those being "large" amounts for a while.

I've got 3.077 used out of 7.784 according to top, this Arch/KDE session has been active for weeks, I play dwarf fortress all the time, use spectacle for screenshots (it's sitting at 0.682 atm) and firefox (of course the 64 bit version) is at 0.796, while my cache has 2.843 used.

If I do something like load df with the new 3d visualizer armok vision that cache gets dumped so av can suck down 2 or 3 gb if it wants it, along with the half gb or so df wants just to idle.

I would love if df could use more memory quickly, it isn't hungry enough due to being 32 bit still.

136:

Personally, I think the drop has more to do with the fact that large large parts of China's modernization are over, and no one has replaced them yet.

137:

That's an odd statement. The "bitness" of your OS and applications has zero to do with the number of bits used for security and encryption.

About the only edge case I can think is valid is that a more modern 64bit OS may have stronger versions of things like Address Space Layout Randomisation as they are newer but most of that sort of stuff gets backported anyway.

138:

My crazy year story is a three way galactic battle between TEST Alliance, The Honeybadger Coalition and GoonSwarm, which is being funded by an interstellar casio and which will potentially burn through millions of dollars worth of ships before it is through.

I guess this is what interstellar battles look like in a world where life is cheap, FTL is possible, and griefing others is the highest form of art.

Other than that we, just stepped into a parallel world where computers got high tech again (HOLOLENS BABY)!

139:

Hololens is really good in a version 1 kind of way. Excellent image quality and good tracking as these things go but tiny field of view.

By v2 or v3 they will hopefully have cracked the fov issue.

140:

Still not sure what the BBC Micro Bit is for though, are microcontrollers really good for learning?

It's all about overcompensation.

Computers were introduced broadly into schools in the UK circa 1980-1988, typically with BBC Model B's (think: an education-oriented equivalent of the Apple II) then more business-oriented kit (PC variants like the RM Nimbus, IIRC). In the 80s the syllabus was all about what we'd consider basic computer science -- less heavy on the maths, more on programming (BBC Basic was capable of structured stuff -- supported functions and procedures), hardware components, some basic logic circuit layout if you went high-end, and so on.

But during the 90s and 00s CS in schools morphed into IT and gradually turned into classes in how to use Microsoft Office, rather than what computers were.

Arguably, the app-oriented classes were a good and necessary thing, but they went so far in that direction that the UK ended up with a shortage of programmers and engineers, not to mention a shortage of teachers who knew how to program.

So about five years ago one of the things that cropped up was a revived emphasis on "coding" (often taught by teachers who'd never seen a command line in their life), followed by the sudden, startling success of the Raspberry Pi. The RPi was designed as a teaching machine insofar as it's almost entirely open, the standard Linux-based OS it comes with -- Raspberian -- shoves a bunch of teaching-programming stuff in your face, and it's cheap enough that you can hand them out to kids on the same basis you'd hand out a text book.

The BBC Bit kicks it down a level. Firstly, it's the BBC's play for relevance in this sector. Secondly, it's so low-level and basic that it's aimed at teaching 11-12 year olds the basics of assembly and low-level programming; the RPi is considerably more complex (as powerful as a circa-2002 desktop, frankly). Thirdly, it's so cheap that they're giving them to all kids in Year 7 this year. Some will trade them, some will break them, some will give them away ... and some will build supercomputing clusters or innovative wearable projects.

The point is to get kids interested in programming again, at all levels -- tiny embedded gizmos you can program from your phone over a USB cable, right up to something as powerful as a 1990s minicomputer or workstation -- at an affordable price.

As with all such educational initiatives it's a decade or more behind the times. But it's not a bad thing in and of itself.

141:

In this context, what does "a decade or more behind the times" mean?

142:

The lead time is all wrong. They're trying to teach high school kids the basics for the skill set they'd like to see graduates turning up with -- today. But these kids won't be joining the labour force for 5-10 years, by which time things will have moved on.

School syllabi are notoriously bad at tracking moving-target subjects. English grammar and trigonometry don't change over a period of multiple decades, but biology was a crap shoot in the 70s to 80s (and probably still is), and as for CS, tracking it is hopeless when the syllabus needs to be set 5+ years in advance and text books need to be written. Remember, there's no revolving door between school teachers and cutting-edge software development or chip design.

143:

I'm not sure the difference is as big as you make it out to be. I'll use myself as an example.

When I was in High School in the early 00, the computer science class taught JAVA. (The computing for business class taught Office). Now, people today need to know how to query databases, a skill that was not taught in HS then. However, the stuff learned in the JAVA class still applies to modern JAVA, Python, etc. despite the greater reliance on libraries these days.

Do you see greater changes coming down the pipeline in 5-10 years?

144:

I've got to agree with you. Essentially you're looking for kids to graduate knowing several broad areas of programming and having a good-enough idea of how these concepts work to apply them to a different programming language, or to further self-educate as needed:

1.) The really basic stuff; if-then, for-next, declaring a variable, program flow, etc.

2.) Building methods (or functions) into classes and creating objects.

3.) Code libraries and how to use/modify them.

4.) Relational databases and SQL. No need to be an expert, but enough knowledge to be comfortable with a statement like "select phone-numbers from address_table;" (or how to create the address_table) plus the idea that the programming language might state the query differently than the SQL.

5.) How to make your application talk to a web-server or a some kind of library for graphical user interfaces, like Tk.

6.) The idea that there are coding standards and how to relate to coding standards.

7.) Basic command-line knowledge. ls. chmod. grep. etc.

Having a kid graduate high-school knowing how to do the above in Python, Ruby, or even TCL would be hugely helpful to their later educations, and shouldn't take more than a year.

I'm very, very big on the idea that traditional maths - Algebra, Geometry, Trig, etc; should not be taught in high school. Instead I'd like to see some really practical stuff being taught instead; the necessary math to do programming, electronics, chemistry, etc. being prioritized. I know that they're really the same maths, but I think presenting them as very practical tools would have a huge impact in whether students decide it's useful to be math-smart. The formal maths could then be taught in college.

145:

Essentially you're looking for kids to graduate knowing several broad areas of programming and having a good-enough idea of how these concepts work

Java is terrible for that because one key feature you need in order to understand programming is missing completely from the language -- pointers.

There's a body of heuristics in teaching CS that suggests there are three levels of abstraction that students need to master are predictors of long-term success in higher level CS studies. If they don't have all three basic abstractions down at some level, then they're going to fail.

Variable assignment (and the concept of variables) is important, yes. That's the first abstraction. The second is looping and conditionals. But you can't truly understand programming without also knowing the third and most difficult abstraction -- how to use pointers. Teaching languages where pointers are optional are fine -- Python, for example, or Pascal family languages in an earlier iteration -- but ones where pointers are utterly inaccessible (like Java) are going to kneecap the early-teen programmer before they even get started.

146:

Kernel Patch Protection, Driver Signing, Data Execution Protection are all 64 bit Windows security features, there are others which can be used with both but are more powerful on a native 64 bit system like ASLR as you mentioned.

You can also run the WoW64 layer for software if you just have to install a 32 bit something-or-other.

Naturally these issues are less critical if you're not using windows, and you really shouldn't be... plus if you care about security you probably aren't anyways.

Though if you're on linux or something you might be on a 64 bit version already as they're becoming the default option everywhere I've seen. I mean, the only 32 bit stuff I have on this system is there for df compatibility. Everything else, including the browser I'm writing this on, is 64 bit.

147:

AMD/Intel 64-bit is not the only 64-bit out there.

Further, "Kernel Patch Protection" and "Driver Signing" are solely software features.

148:

Algebra, Geometry, Trig, etc; should not be taught in high school. Instead I'd like to see some really practical stuff being taught
WRONG
( Or was that "not even wrong" ??)

These are very, very practical subjects, especially if you are going to do a science degree or other higher course .....
Especially al-jibra & trig I think ...

149:

Agreed.

The maths I learned in school has probably been the most useful of all the stuff I learned there. I still use it frequently. The level we got to has pretty much defined what I've been able to cope with since, and with hindsight, I could wish that school had gone on for another year in order for it to be taken further.

It was always well in advance of what we needed for physics. For example, physics gave us a formula on a plate for finding velocity from constant acceleration, but we already knew from maths that this was a matter of integration and so could cope with non-constant acceleration too. As for chemistry, that was never anything more than arithmetic.

150:

If they don't have all three basic abstractions down at some level, then they're going to fail

Strongly agreed. In my last job, I had to work with C++ code written by many such "retrained Java programmers".

Glorious fun. Randomly deciding to create objects on the heap rather than stack? Check. Randomly deciding to delay the initialisation of said class members "because more efficient, innit"? Check. Inconsistent mixes of exception and return codes for error handling? Check. Randomly deciding to use multiple levels of inheritance? Object-oriented, innit?

Deep joy, much frustration. Still, post job move last autumn, I'm now back to working only 37-hour weeks, writing C++ at a firm where a couple of people were actually interviewed for their jobs by OGH...

151:

Damn, and I didn't even mention the delights of the memory management code quality from programmers brought up in a garbage-collected language...

Grinds teeth gently, wanders off towards wine glass :)

152:

As for chemistry, that was never anything more than arithmetic.

So I'm guessing you never got to reaction kinetics or anything in biochemistry, much. (Not to worry, that stuff was A-level or degree level.)

153:

Whaaat, you're there?!?

154:

My thought is the BBC Micro Bit is a bit strange in it's setup. It's basically microcontroller level; with a few LEDs, some buttons, an accelerometer, a compass, IO pins, and bluetooth LE. Great, fine. It's a nice set of sensors and I can see ways in which it's going to be used, but ...

Firstly, the programming is resolutely high level. They aren't going to be learning any assembler or low level stuff here - it's javascript or graphical block programming, or at it's most techie, python. So the emphasis isn't on learning how the guts work, more on the hardware experimentation side, but not in a way that works for real microcontrollers.

Secondly, although that's fine, I'm not sure if it's going to convince a year 7 kid too much. They've been bought up on smartphones, tablets, etc. Lighting an LED when you push a button? Meh...

If it had been graphical block programming on a modified RaspPi Zero for more abstract stuff, with the Bit being a paired hardware based addon with lower level stuff for more advanced students, I might have got it. Particularly if explicitly paired with smartphones via bluetooth and a capable app that allowed both utilisation and programming. I can see schoolkids liking being able to remote control it from their phone.

And, of course, this neatly ignores that there's not much of an IT industry in the UK anyway, since the government has neglected what it hasn't stuffed up. Which comes back to having politicians who are law/economics/pol sci types, rather than engineers.

155:

Yup, but it's been bought - so it's now the UK development centre for a US firm; and I suspect that most of the challenging personalities you knew have taken their shares and run... So far, so good ;)

156:

Secondly, although that's fine, I'm not sure if it's going to convince a year 7 kid too much. They've been bought up on smartphones, tablets, etc. Lighting an LED when you push a button? Meh...

Don't underestimate the buzz -- the personal sense of triumph -- a kid gets from making something work. Smartphones are if anything too damn complex and distracting; making a buzzer go buzz, or a couple of LEDs light up in sequence, is the sort of thing you can easily accomplish in a 40-minute class period with a bit of help and without having to write 50 lines of code sourcing in libraries before you can fart "hello, world" to STDOUT. (Real GUI programming requires a nightmarish amount of secondary expertise to make it practical; trying to teach it in bite-sized chunks is doomed to failure unless there's a very high level API available, and an off-the-shelf framework app that the learner is merely tweaking slightly.)

Not much of an IT industry in the UK anyway -- tell that to the guys up the top of my street who just had to move out of their office complex because it's too small and they're taking over the former Scotsman newspaper's HQ. Clue: they're called Rock Star, and they write games. Saying there's not much of an IT industry in the UK is like saying there's no car industry here, either.

157:

Wait, what?
Just a question- by school you do mean university?

And they taught you chemistry by doing thermodynamics and mols and such? That's such a stupid way of doing it; chemistry was a hands on discipline, until now when it has gone all maths and become fully a part of physics now that we have the computing power. Even if you don't do labs you still just need to learn stuff, from elements to reactions to crystal structures and so on.

158:

Don't underestimate the buzz -- the personal sense of triumph -- a kid gets from making something work.

As much as I'd like to think they still have the same buzz that we might have got as a kid from that, I'm not sure it's still there, at least not for the majority. Tech works, otherwise what's the point, seems to be more the worldview. That's what I get from nephews etc. The 'how' seems akin to the 'where does butter come from' question.

Saying there's not much of an IT industry in the UK is like saying there's no car industry here, either.

Not saying there's none, but in relation to where we could/should have been ...

It's no good making with the cheap words when the actions, support, attention, don't back them up. A few million here for IT education, vs a many billion for the finance industry. It's not in the same ballpark.

159:

Nine out of ten kids won't get anything out of the Bit, nor would they get anything out of a Raspberry Pi. One kid in ten will, and sparking that interest when they're at school is worth a lot to them and to the future of the others who aren't fascinated how a scrap of plastic and copper and lumpy bits can do things if they do the right things to it. That kid will dig through the manuals, and learn and play and try and fail and get it right and try something else, for the hardware cost of a couple of school meals. Bargain.

The Bit is not IT, it's engineering.

160:

Biochemistry, no, not a sausage. Reaction kinetics, yes, but I don't recall it involving any mathematics other than of a trivial nature.

161:

There's another problem with Java as first language; how often an instructor has to say "ignore that, we'll come back to it later." There's no elegant elision of complexity when you've beginners copying-and-pasting "public static void main(String[] args)" before they've learned what an array is.

162:

Question for Host and Audience:

Which is the BBC Micro game you remember most?

1: Elite
2: Exile
3: Repton
4: Stryker's Run
5: Jet Set Willy
6: Thrust
7: Chuckie Egg
8: The Sentinel

Answer has implications. [Note: Mind is from Red London - there's a 2D game with a wizard in it who had to swap spells and could only carry two that doesn't apparently exist here that was ultra-popular]

~

As for Crazy Zone:

IMF Internal Meeting Predicts Greek 'Disaster', Threatens to Leave Troika Wikileaks, 2nd April, 2016

We're back in 2012-3 people (and with all the same issues, with only +500 billion or so in QE) with all the same bugbears and so on.

Guess we're going to do the re-run.

Oglaf Knows the Score Oglaf, recent. Press "next page" for the second half. Oh, and NSFW if you're prudish.


(Note: Yes, there's an Oglaf joke hidden in the Game listings)

163:

Oh, and for THE RULES, BOY:

The Panama Papers

Iceland’s PM faces calls for snap election after offshore revelations Guardian 3rd April 2016

Note: genius video.

Hint to the boys: I did tell you that it wasn't the protests in the square in front of the Church / Althing that'd get you, but the G5 flying out overhead...


~

Fun fact.

Waaaaaaar.

164:

No, not university. I did chemistry A-level, but anything I've learnt beyond that has been off my own bat.

Thermodynamics and mols certainly featured, but so did rote learning of huge long lists of organic reactions, and crystal structures cropped up at least three times (A, O and pre-O), repeating the same ground every time without ever telling us anything useful. (Teacher messes around making a bubble raft; then: "Atoms pack together as closely as possible. This gives us hexagonal close packed and face centred cubic. Some crystals are body centred cubic, even though that isn't close packed. We're not going to tell you why some of them don't do what we've just told you they all do do. Oh, and here are all sorts of other arrangements. We're not going to tell you why those happen either.") We did do a few practicals but disappointingly infrequently.

I think it was a pretty shitty way of teaching it - far too much rote learning of lists in isolation, and not enough inculcation of understanding. The long lists of organic reactions were forced on us without more than a tenuous relation to the brief mention of quantum mechanics of orbitals and bonding that had preceded them; reaction mechanisms came later and again were not related to anything like a useful extent. I understand those areas far better to the extent that I have revisited them off my own bat approaching from the opposite direction. And large chunks were just plain missed out - in particular reaction conditions - "This reduction can be carried out readily using hydrogen and a platinum or nickel catalyst", so I try it at home and nothing happens; 20 years later I find out that it wants 30 bar and 150 deg C to work, which is a new definition of the word "readily" of which I was not previously aware.

165:

I think the presence or absence of the buzz is more a function of the kid than anything else. "Tech works, otherwise what's the point" could certainly be said to apply to plumbing since long before I was around, but that didn't stop me being fascinated with it, being delighted to receive a box of elbows, T-pieces etc for my 5th or 6th birthday, installing a flushable toilet in my tree house and so on. A few years later I moved on to electronics, via radio - again very much a "tech works, otherwise what's the point" thing, but owning a couple of factory-made radios didn't make it any less interesting to play around with crystal sets, progressing to adding gain stages to them, etc.

166:

Elite, though I never played it myself. I saw other people playing it and remember being impressed with the graphics programming, the physics simulation, and the neat way it switched screen modes in the middle of the frame to draw the dashboard, but as far as actually playing it goes, of all the things you could do with a computer playing games was (and is) by far the least interesting. I didn't even play the games I wrote myself.

Of the others on your list, I remember seeing Chuckie Egg but not what it was about; I remember Jet Set Willy existing but I thought that was a Spectrum game; I don't think I've heard of the rest.

167:

I learned how to code in school, but I can't remember the last time I actually had to. There are nigh-infinite libraries of code available for free. We hardly need programmers; it's much more useful to be a software archaeologist who knows how to dig up old software packages and get them running. I have had some notable successes that stemmed from identifying a problem, finding a decent bit of software on the web that could be applied to the problem, and figuring out how to get it operational.

168:

Our kind do not play Chess, we play Go


Wait until you really understand genetic algos and what it means to writing code.


*shrug*


Shedding Scales. Puberty is such a chore.

169:

Fair play to any one who gets organic chemistry, I did A level chemistry twice, managing to improve to an E
I am currently struggling at work to understand why 1,2-dichlorobenzene and 1,4-dichlorobenzene are considered toxic at levels three orders of magnitude different, as you wouldn't think having a chlorine molecule on a different part of the benzene ring would make such a huge difference
One of the reasons I love the comments here, is someone might actually know the answer

170:

One of the problems with the idea of teaching computer programming is that there is so much variety in what a "programmer" does. Lots of us put together third-party libraries into useful combinations, others don't. Joel Spolsky wrote a good article about this:
Five Worlds

And that's not getting into the difference between computing and computer science… If you have an hour to spare, ask a functional programming advocate why just about every introductory programming course on the planet is Wrong. :-)

What seems to be new in the 21stC is that everything is now so cheap that we can try multiple approaches in parallel. Raspberry Pi or BBC Bit? Both!

171:

I went with a G-Tech Thunderbolt RAID box.

Something like this?

http://www.tigerdirect.ca/applications/SearchTools/item-details.asp?EdpNo=9801000&CatId=4269

The other option you can do, if you have good wired networking, is use a NAS box. Something like FreeNAS can do NFS and CIFS/SMB, and can handle far more than 6tb.

Given I have no idea what those letters mean, I think I'd better not go that route :-/

For media, I'm thinking a cheap Mac Mini with a couple of external drives (1 storage, 1 backup) communicating with the AppleTV might be the way to go.

172:

I often find that it is easier to code something for myself than to thrash around trying to figure out how to use someone else's code for it. Problems include, but are not limited to: documentation that is shit because it assumes you already know how it all works; documentation that is shit because it barely exists; piss awkward interfaces for no apparent reason; author of the code thinks about the problem in a manner which is orthogonal, in many dimensions, to how I think about it; problem and code eventually turn out to be incompatible, but only after doing a lot of work; code needs to be reworked because the compiler barfs horribly on it or installing it as-is would break the system; code has dependencies which themselves give rise to the same sort of problems as the code itself...

Doing it from scratch myself may not always be practical, but it has far greater potential for satisfaction, is more likely to do what I want in the way I want it done, and provides what may well be an invaluable understanding of how the algorithm works and what can be expected of it.

173:

C6H5Cl + Cl2 → C6H4Cl2 + HCl

vrs

C6H6 + 2 Cl2 → C6H4Cl2 + 2 HCl

It has the same result C6H4Cl2...


But it's like making a cake and making a cake from rocket fucking fuel.

174:

Extremely simplistic version:

The N/S configuration of the 1,4-Dichlorobenzene molecule is all about the energy released during the formation. [And, well, other kinky things]


An A level explanation: the energy excites the molecule and tears it apart from North and South to gain the...

Sigh.

Yep, bollocks.

But, yeah - it's actually about the sequence progression of bonds.


And no.

They don't teach that you in A Level.

175:

You're certainly correct. I'll add pointers to my list.

I don't think about pointers much because the kind of code I write is pretty basic; dumping an ordinary variable into a database is as sophisticated as it gets.

176:

Agreed that they are practical and useful and one of my big, sad wishes (in my early fifties) is that I'd studied more maths.

The reason I didn't study more maths is, IMHO, because I was introduced to them poorly, so I've frequently tried to figure out how to introduce maths in a more intelligent fashion. The thing I keep hearing from kids about math is "I'm never going to use this stuff."

So how do you address that? You need some classes set right after kids have learned basic arithmetic which are dedicated to demonstrating that "math is practical." The classes should combine math with helping kids learn to build stuff... and that's where your next generation of scientists and engineers is coming from.

Of course I'm addressing this from a U.S. perspective; you're doing so much a better job teaching children in the U.K. that I don't feel comfortable prescribing anything to your school systems.

177:

"I think it was a pretty shitty way of teaching it - far too much rote learning of lists in isolation, and not enough inculcation of understanding."

That was my entire schooling. Every subject. At one point I did ask the teacher what the objective of the course was (English as it happened, but could have been any subject). I was sent to the Head for caning.

178:

"There are nigh-infinite libraries of code available for free. We hardly need programmers; it's much more useful to be a software archaeologist who knows how to dig up old software packages and get them running."

This really resonates with me! I am not a programmer, not since high school when I taught myself AppleSoft BASIC. (Although I did once pick up a fast twenty bucks once by writing a BASIC program to measure out the markings on a custom fuel measuring stick for a local businessman who had a fuel storage tank of unusual geometry and less mathematical skill than me.) I also did a bit of Fortran (already obsolete) in college. Nothing since then.

But these days I sometimes call myself a "cargo cult programmer" because I can Google like a pro. I do a lot of site development using the WordPress CMS, which means if I want to do something unusual, I need to "write" something in PHP. I don't "know" PHP. I cannot "write" a program in PHP. But I can find some PHP that does something like what I want, and then tweak a few lines. Hail, almighty internet airplanes full of cargo, thank you for your cargo, please bring me more cargo!

179:

"That's an odd statement. The "bitness" of your OS and applications has zero to do with the number of bits used for security and encryption."

No, as far as security goes, this comment from Max™ argument is true according to all experts in the field. the "special contexte switch" (for lack of a better term) between 32bits and 64bits involves a number of complicated steps that are apparently fairly fragile as far as security is concerned up and including various chip level bugs.

My argument was more about memory efficiency, which also translate to a lot less cache miss which is a performance concern : if you are not manipulating huge areas of memory, you do not need to go 64bit (except for security reasons once again, but most people do not do things that are very security sensitive, server programs are security sensitive, but for server load 64bits is prefered anyway).

If for any reason you have to handle more than say 1.5Gb of private memory you are better of in x64, if you are pretty sure you are going to use a lot less x32 may be better. It depends a lot of your use case.

(anyway that's a decision you can often postpone and take only after measurement if you are using one of the more abstract JITed languages)

180:

As much as I'd like to think they still have the same buzz that we might have got as a kid from that, I'm not sure it's still there, at least not for the majority.

It never was for the majority. I was a kid during the Eighties, and even we had some electronics. We build some simple things in the seventh grade (in 1989, for me) and even though *I* was interested in that and making the LEDs blink and how transistors work, not many of my peer group were.

Most people (boys, at the time - then the choice was between textile craft and "woodcraft", including metalwork and that electronics stuff, and it was mostly divided by gender) just built the things and when they got it working, they went to do something else. Granted, many of them went home to play with the C=64/Amiga/PC, and some even programmed them, but most just did what they had to and never were that interested.

I built stuff, mostly from instructions, and finally got my Masters in space technology, which was then part of the EE department and as studies, under Electronics. I was the exception in my 10-14 age group for that.

181:

Have to disagree with you slightly there Greg, whilst I dont disagree with their usefulness as a foundation for later study there is a huge amount that can be done at a basic level that would set up students that arent going to a higher level.

For instance Logic to help parse the contradictory statements of our politicans or the Daily Fail.

Occam's razor and probability.

How to assess risk properly - something us Apes are terrible at if we rely on our instincts.

There is loads a practical maths education could do to create strongly questioning students of the world.

As for Comp Sci curriculum, it should give equal weigh to BA, PM and lightweight Architechture disciplines. A good BA who is fluent in both Business and Technical language is worth their weigh in gold.

(IMO anyone who calls themselves a programmer who hasn't earned their BA stripes at some point is just a code monkey - but thats probably my prejudices showing.)

182:

Interesting question - to what extent is the male bias in STEM disciplines a result of the way it is taught? If we posit a change in societal norms such that by say 10 that there is a an equal probability of males and females going down a STEM path - what would an ideal curriculum look like compared to todays?

183:

IT industry in the UK?
Or the heaps of techies around Old Street Roundabout ...

184:

Yes, although my qualifications are in the hard sciences, my experience of workshop practice was & still is useful & GREAT FUN.
Besides, we often had to build our own lab-equipment set-ups for preliminary experiments - there WAS the time I managed to get high-pressure gelatine into the light fixings though .....

185:

As opposed to brewing our own beer, then distilling it ... (VIth form) or making Ammonium Nitrate ...
The latter now understandably banned ....

186:

but that didn't stop me being fascinated with it, being delighted to receive a box of elbows, T-pieces etc for my 5th or 6th birthday, installing a flushable toilet in my tree house and so on.

Doesn't sound like that's the kids of today, unless you are 12...

Whilst I'd agree that it was never 'everyone', I have seemingly noticed a more general lack of interest in the 'nitty gritty' level than before. At one stage people would consider soldering their own electronics to make a 'computer' the base level. Then it was assembler level etc.

Today it seems to be plumbing libraries together with little bits of bespoke code and compiling. The type of level that might fit with a BBC Micro Bit perspective, it doesn't strike me as being there in the young.

Maybe that's just me getting old (you don't say), but I feel that something like OpenCV is more the level of the young tinkerer today - hence why something of the RaspPi level seems a better fit. If you were down at the more hardware level then programming smart RFID tags could be the thing (tag the teachers).

And yes, I did brewing beer as school too, wouldn't happen today...

187:

I strongly believe it has a lot to do with the complexity of all components involved. Compared to even the '90s and early '00s it is significantly harder even as a professional in the field to keep up to date with ALL details at each level of the hierarchy.

It was expected for a programmer then to understand all steps from high-level C++ code down to what the hardware will do physically though compiler optimization and so on. I will argue that being able to do the same today is hard enough to be out of scope for most people and due to the obfuscation of some of the steps though proprietary technology impossible without access to even more time to reverse engineer things.

Most of the easy explanation I have seen of what a compiler does or how instructions are executed on modern processors can be filed under "lies to children" - useful simplifications, but wrong.

188:

Elite!
Although I have to admit I cheated when I was offered a disc with a fully equipped ship.
Apart from the quality of the game the brilliance of the programming with two graphics modes active on one screen was so innovative.

189:

A friend of mine who currently works as a web programmer borrowed my copy of Knuths TAOCP a while ago.

He brought it back a week later complaining that it had "nothing to do with computers".

190:

Or perhaps the 50% of the financial sector workers who actually work in IT?

Only small minority of IT workers actually work for tech firms.

"Most of the easy explanation I have seen of what a compiler does or how instructions are executed on modern processors can be filed under "lies to children" - useful simplifications, but wrong."

Define "wrong" most of the Physics models taught at GCSE and A-level can also fit into those categories, that doesn't mean they are not useful models. Wrongness really doesn't come into it as long as the answer to the question "is it a useful model" is Yes.

Summary of my thoughts on the Maths curriculum :

http://mathwithbaddrawings.com/2016/03/02/why-the-math-curriculum-makes-no-sense/


191:

Frak!

https://en.wikipedia.org/wiki/Frak!

Elite too but only made it to Deadly as the last mission had a bug in it and I couldn't catch the navy prototype, and had overwritten all my early saved games.

Sentinel was way too hard...

192:

Bell and Braben were bloody brilliant — they got the Beeb doing stuff we never realised it could do. I was in the room at Acronsfot1 when they did the initial demonstration of a Cobra flying through space and I swear we all ducked when it flew out of the screen, despite it being a crude wireframe.

(Shades of the first people screaming at a movie of a loco coming out of a screen.)

That code was so RAM-constrained that in the end it was shipped with known bugs in, simply because it would have taken too many bytes to fix. They literally had no memory left — taking three bytes to fix an issue meant somehow shortening the rest of the code by that three bytes.

There was 32K RAM in the BBC-B. Which included screen memory, OS, program and data. A chunk of the code was also used as data — using arbitrary code bytes as seed values for procedural generation of the galaxies was the sort of thing they'd do.

1Officially Acornsoft, then at 4a Market Square in the centre of Cambridge.

193:

We hardly need programmers; it's much more useful to be a software archaeologist who knows how to dig up old software packages and get them running. I have had some notable successes that stemmed from identifying a problem, finding a decent bit of software on the web that could be applied to the problem, and figuring out how to get it operational.

This again depends on what kind of software you want to write. In most (enterprise) software projects I see, one of the most important things (which not everybody realizes) about third-party software packages is that they are supported properly. It's not smart to build your business around a package that somebody else has written ten years ago and nobody really supports anymore - you don't know what bugs there are, and even if you did, you have no idea who would fix them.

This is especially important if you distribute your program for basically anybody - which is almost all the software on the WWW. Even if you have logins and all that stuff, you want to things to work *and* trust they work in the foreseeable future.

Digging up abandoned software projects and including them into your software (without taking the responsibility for supporting and updating them) is, in my view, a big no-no in commercial projects. In the projects you do for yourself, knock yourself out, but I don't want to use difficult and non-supported packages even at home.

194:

Frak! was a bloody awful game even for the 80s. Yes it had big sprites and fancy graphics, but that just made it the template for generations of painfully dull but pretty games that followed.

@192 I saw a documentary about elite a while ago. I was really hoping they would interview both creators at the same time, as the really interesting question is whether they can talk to one another without spitting :)

195:

Models which simplify complex systems so we can grasp their functions are absolutely useful, and I am not trying to suggest otherwise. However, the jump from using the knowledge from the models to actually understanding the real systems has grown significantly.

This makes experimenting with current consumer technology harder than in the past. I believe this is what drives potential tinkerers away. It is possible to bridge that gap with research/ more coursework but that requires more dedication and time and someone with intimate knowledge of all relevant parts will probably have a Masters-equivalent knowledge of Material Science, Electrical Engineering, Computer Science and Mathematics (at leas, I might be leaving some stuff out).

196:

1,4 dichlorobenzene oxidises enzymes which control aptosis is which is why it is carcinogenic. Preventing aptosis AIDS the proliferation off cancerous cells. Enzymes are convoluted molecules whose actions depend on ther shape. The para-dichlorobenzene molecule has the fight shape for this oxidation. I can't find any more detail in the 5 minutes I allowed for the search since I no longer have access to university libraries.
If you want to investigate the shape of molecules I recommend the Molecules app for the iPad and iPhone.

197:

Evaluating alternative software packages for usability and available community support is a huge part of the computer archaeology skill set. Also security, if that's important for your application.

198:

> Saying there's not much of an IT industry in the UK is like saying there's no car industry here, either.

Not saying there's none, but in relation to where we could/should have been ...

Thank you for playing: the UK is the EU's number two car manufacturing country, in second place only to Germany.

The fact that they're mostly badged "Ford" or "Toyota" shouldn't mislead you -- the auto industry in the UK is pretty big. (We're also one of the leading manufacturers of Formula One kit. Ever heard of a little company called Maclaren, for example?)

My point is, these industries -- computing and cars -- aren't represented by flagship British owned companies, because most such were acquired by foreign multinationals thanks to successive conservative[*] governments' lack of an industrial policy; but they still exist in the UK, and they're pretty big. (Although they're not mass employers in the way that coal and steel and manufacturing used to be, because they're no longer that labour-intensive anywhere.)

[*] I include New Labour, 1997-2010, in this.

199:

Right shape not fight shape.

200:

It's with enzyme kinetics in biochemistry that things begin to get hairy (as in, calculus hairy), before you get to pharmacokinetics, which is a whole 'nother Gordian knot. I'm guessing you studied chemistry to 'O' level or GCSE and no higher.

201:

Crazy Years? Like Heinlein?

THIS sounds pretty crazy to me:

http://www.newyorker.com/news/news-desk/why-are-educators-learning-how-to-interrogate-their-students

... And with the TISA & TTIP coming right up - we probably have to accept that these "services" are available in the EU.

202:

We're also one of the leading manufacturers of Formula One kit

Ignore Maclaren. The current F1 leaders are based in Brackley, Northants. Yes they may be called Mercedes, but the F1 team is solidly based almost in earshot of Silverstone.

(An F1 team not based in southern England is perhaps more of an exception. The major one based elsewhere is Ferrari, but from Red Bull through Williams to Force India, there are a hell of a lot of them here.)

203:

Which is the BBC Micro game you remember most?

1. So not a gamer.

2. My computing progression was, like:

a) Sinclair ZX-81

b) Ditched the ZX-81 hastily, bought a Casio FX-702P instead (which I still own)

c) ... wait 4 years ...

d) Buy an Amstrad PCW8256

e) Progressively upgrade the Joyce box to an 8512 with a 10Mb hard disk and a bunch of CP/M apps

f) Switch to PCs once they get cheap enough

I basically missed the entire 1980s games-oriented personal computing boom by not getting a colour-capable computer until 1989.

204:

Would I be right to guess that your first and/or favorite language is something c-like and very pointer based?

I did CS and have a coding job now and I'm not so sure I agree.

Pointers are good for giving you a better idea of how things are working underneath (if they are indeed working like that underneath) but in most modern high level languages if you see that some coder has used pointers for something... you're about to find an ugly hack that should have been done properly some other way.

You can also often see the style of the programmers first language in things they write in other languages. Start someone on something like C and they'll often end up writing crap OO or Functional code later unless they really really get into it because they'll still be thinking like they're writing C

If I was starting kids off and trying to prepare them for the multi-core future I'd be more inclined to start them off on a functional language since it will get them into the right way of thinking long term.

In real terms on modern machines with modern processors pointers are themselves an abstraction and are a little bit more misleading than then would have been a couple of decades ago. Hell even assembly actually gets treated as high-level code by modern desktop processors.

205:

Actually, my preferred language is Perl. (But I haven't done any serious programming in years.) Yes, Perl has pointers (although you don't need to use them) -- more accurately, it has references, which allow you to build complex data structures while retaining metadata about the underlying variables so you don't risk crashing the stack or doing anything untoward. It comes in most handy for rolling your own OOP system on the fly.

206:

A good understanding of parallel programming paradigms be they distributed or shared is vital to efficiently using modern architectures. Could you elaborate why you consider functional programming a better approach then imperative programming for understanding these concepts?

In the HPC corner where I operate functional programming is not something I have seen yet, but rather c/c++, fortran (for legacy codes) and python. This is surely due to the performance concerns but I am curious of the advantages of functional programming in this context.

207:

Pure functional programming languages allow you to write code that can be automatically parallelised by a Sufficiently Smart Compiler.

Indeed, state of the art implementations show impressively linear scaling as you add cores, with constant overheads making them only a couple of orders of magnitude slower than nasty low level code.

Of course with the proliferation of modern multi core machines they are sure to take over Real Soon Now.

I have been hearing this song for the last 20 years, and somehow it never quite seems to happen.

I do like functional languages for some things though - with a bit of thought a most algorithms can be represented beautifully.

208:

I have been hearing similar songs regarding Sufficiently Smart Compilers and shared memory parallelizable code. Given the amount of papers on the topic it is a bit disappointing to discover that for most non-trivial cases humans are still so superior to automatic parallelism as to make the approach useless in practice.

That functional languages have beautiful representations is something I can whole-heartedly agree to, it's writing a CFD or neutron transport simulation in them that can get tricky...

209:

The thing is that with numerical code you almost always know where the big gains are - it's the point where you hand over to a function and say "crunch that big vector of stuff".

At that point the key is usually to have a sufficiently smart library.

I have been quite liking CUDA* for big vector of stuff processing. The concepts are nothing remotely new of course but it is really nice to be able to do non trivial parallel calculations on any old laptop.

*not so much OpenCL. It suffers badly from design by committee & shoehorning of every vendors favourite features.

210:

I once had to teach linear algebra using Matlab to a bunch of business management, er, people - but who had never heard of a matrix.

211:

"Of course I'm addressing this from a U.S. perspective; you're doing so much a better job teaching children in the U.K. that I don't feel comfortable prescribing anything to your school systems."

Our Lords and Masters are working on achieving a level playing-field with the USA in that respect.

212:

I could answer these in detail, but don't believe everything you say. Functional programming is good, but the "sufficiently fast compiler" project is progressing at the same rate as nuclear fusion. And, in scientific codes, 90% of the gain is being algorithmically smarter - that's hard, so people throw resources at it.

213:

"Interesting question - to what extent is the male bias in STEM disciplines a result of the way it is taught?"

Probably very little. There are fundamental differences in the distribution of skills within the sexes, and the bias will remain for all time. Which is not the same as saying that there are no problems that need resolving.

And think on this: one of the reasons that female pupils are going in non-technical directions is that their communication and social skills are on average better, and the status and pay of technical people in the UK is low, and dropping steadily.

214:

Sorry :-( That was a typo. that made it read offensively - NOT my intent - "don't believe everything you READ".

215:

I will be delighted when the sufficiently smart compiler finally arrives, as I find tuning to be one of the least interesting parts of development.

It has its moments of course but I'm much more interested in capabilities than the process of coding them.

216:

I did CS and have a coding job now and I'm not so sure I agree.

As did/do several who post here :) I suspect Elderly Cynic is better-qualified than me to comment, but I'll do it anyway...

Pointers are useful things. For instance, would you rather sort a large list of heavyweight objects, or sort a large list of pointers to those objects? How else would you choose to model a hardware concept, perhaps a memory-mapped register?

You can also often see the style of the programmers first language in things they write in other languages. Start someone on something like C and they'll often end up writing crap OO or Functional code later unless they really really get into it because they'll still be thinking like they're writing C

Perhaps so; but I'd suggest that it's more to do with the first couple of years in industry.

If your first trainer / supervisor knows their stuff, and takes the time to train you properly, they put good habits in place that last. If you're left to "just get on with it", guidance not provided, then in five years time when you're an experienced engineer, who are they to tell you how to code? You've weren't there, man, you don't know... ;)

At university I was taught in Pascal, Standard ML, VMS Macro, OCCAM, and Prolog (and C by the EE department). My first job was using assembler on a SIMD machine; the next was C on MIMD. This was the 1980s, so "Object Oriented" wasn't really widespread in industry - and yet, I'm an enthusiastic advocate of object-oriented design where appropriate (and RAII, and exception-based error handling, and...) and have nearly twenty years of C++ behind me [1].

[1] Whether it's twenty years experience, or five years experience several times over, is left as an exercise for the observer...

217:

5 years 4 times over is still better than 6 months 40 times over.

218:

For scientific computing, algorithmic improvements are definitely "the holy grail" of performance, but that kind of progress is almost always revolutionary and requires some genius to pull of. Tuning algorithms to perform on specific architectures is do-able by a larger set of people, and worth enough money/energy/time to have people working on this full-time. There is some automated support available, but it is just that, support, and nowhere near magic compiler land.

I have not even heard of technologies yet that could do that kind of work well, and there is nothing even on the horizon for anything involving MPI programs which are the standard for really big scientific problems (Brain, material science, "simulation of explosions").

I hope to be able to retire in 30 years still making other people's code better barring any exceptional developments OGH has already considered in speech and writing.

219:

For instance, would you rather sort a large list of heavyweight objects, or sort a large list of pointers to those objects? How else would you choose to model a hardware concept, perhaps a memory-mapped register?

This is of course heavily language dependent but who's going to actually move heavyweight objects around in memory? If anything high level languages which hide the pointers are going to have you effectively simply swapping pointers around.

220:

So ... does any of the above CS tech bring us any closer to AI? Or, does AI require a completely different set of basic/core abilities/languages to provide/use instructions? (IOW, would someone trained only in the above mentioned computer languages be able to even recognize an AI program if they stumbled upon it?)


Re: STEM people ...

Know/have worked with scientists (hard, life, med, etc.), mathematicians, engineers, etc. IMO, the key difference among these folks is to what degree they're dogma-entrenched vs. discovery minded. The CS/Eng/MD (GP) folks are much likelier to be by-the-book ... I usually refer to them as 'technologues' (not scientists). The others tend to know their 'book' but usually think it needs some improvement. Of the two, the latter are more interesting as humans and better explainers of their disciplines.


Re: Math & kids ...

Personally think that geometry, algebra, trig, etc. should be taught in elementary school. When I was in elementary it seemed that the only math was arithmetic, i.e., absolute, fixed, only one right answer, etc. Terrible way to introduce the language of abstract relationships to anyone.

221:

There are fundamental differences in the distribution of skills within the sexes

This is one of those unseen prejudices, and it's quite insidious. It diverts girls away from STEM subjects long before any rational considerations such as pay or career prospects, or even the softer rationale of "women have stronger social skills" come into play.

It should be stamped on, and stamped on hard.

223:

Sorry, didn't see your (much more useful) reply.

224:

Only in the sense that it makes it easier for programmers to program non-trivial things.

Re: mathematics in school, google "Lockhart's Lament"

The discussion about the merits of different types of language are somewhat futile if the department of education manage to do to programming what they've done to math and most of the other things they've touched.

On one of the mailing lists I frequent one of the regulars had been asked to consult on the proposed syllabus for a new secondary school level CS course in ireland. It was so so bad.

If you tasked someone with the job of planning how to make people hate a subject it couldn't be done any better. It included no less than 6 months of pure (tripe)theory before anyone ever wrote so much as a "hello world".

If these people were tasked with teaching soccer I can only imagine them coming up with a course that included a year of velocity calculations for hollow objects in a gaseous medium just to make sure that by the time the kids saw their first football they'd already hate it and associate it with misery and boredom.

225:
The other option you can do, if you have good wired networking, is use a NAS box. Something like FreeNAS can do NFS and CIFS/SMB, and can handle far more than 6tb.

Given I have no idea what those letters mean, I think I'd better not go that route :-/


The straightforward version:
NFS = Unix method of sharing file storage over a network.
CIFS/SMB = Windows equivalent.
They are very stable technologies by this point; problems are usually permissions or networking issues, so if you're confident with them (or your ability to Google the solutions) you should be fine.
226:

Educators tend to be people who fell abnormally in love with a subject and found it easy. That makes them really bad educators. The people that would be good at teaching it are the ones that had to struggle with it. They know exactly what the student is going through and what the pitfalls are. It's just one of those unfortunate ironies of life.

227:

Re: 'It included no less than 6 months of pure (tripe)theory before anyone ever wrote so much as a "hello world".'

In undergrad took CS as an elective, exactly as you described. Dropped it mid-term when realized that the stats course/SAS* portion was doing a better job of describing what to do/what happens.

* Of course, later discovered that everyone in the 'real (business) world' was using SPSS ... (sigh).


228:

You clearly haven't looked at the psychological (and even neurophysiological) research data. The distributions overlap, but are significantly different. And it is actually YOUR attitude that causes a lot of the problem - by insisting that everyone is the same, it discriminates against those of us at the ends of the spectrum, male and female. In turn, that puts off the bulk of girls from doing 'male' subjects (and conversely), which often puts off even those who would be good at them, by peer pressure.

229:

"Pointers are useful things. For instance, would you rather sort a large list of heavyweight objects, or sort a large list of pointers to those objects? How else would you choose to model a hardware concept, perhaps a memory-mapped register?"

Hmm :-) I investigated the first, once, and it's not simple; the locality properties are better for sorting the objects. And most mainframes didn't use pointers for the latter purpose. But let's not start the 'pointers are inventions of the devil' versus the 'pointers are essential' religious war on this blog - it's been waging for 60 years and there is no end in sight ....

230:
by insisting that everyone is the same, it discriminates against those of us at the ends of the spectrum, male and female. In turn, that puts off the bulk of girls from doing 'male' subjects
I'm clearly being slow today - please explain how the bit after the full stop follows from the bit in front of it?
231:

How about pitch perception: Compared to males, females (on average) can hear higher frequencies. Does this mean that all males are unable to hear high C? (Consider the existence of tenors ... mythical beasts?)

232:

No, A-level. But biochemistry never came into it, let alone anything to do with pharmacy.

233:

I quite naturally associated a football with misery and boredom without any encouragement...

234:

Wow, that really was badly taught. Chemistry, like many other topics, does involve a lot of memorisation, but few practicals and not even telling you about the pressure and temp required for reduction with H2 and platinum is just bad and careless. One of the things that gets ground into you when you do more chemistry is always state the conditions of the reactions, ie, STP or whatever else.

Also I wonder if age comes into it? There's been a massive change in knowledge over the past 20 years, and people graduating even just a decade after me with the same sort of chemistry degree have somewhat different knowledge bases and approaches to things because they've had more computer stuff. Or at least definitely in the biochemistry side of things, pure chemistry not so much, that's finished as a separate subject.

235:

I think there's more dimensions to that.Not just struggled/talented.

(Subject Talent + good teacher) = Absolute gold but rare, extremely good at the subject and also good at teaching the subject. When you make a mistake they can understand it well enough that they can follow the pattern in your errors to work out your actual misconceptions. I've been lucky enough to have a few teachers like this and they're the best.

(Struggled + good teacher) = Reasonably common, second best option. They may not be the best at the subject but they can help you get to a reasonable level.

(Talent + Bad teacher) = These can be some of the most painful ones because they're not talking crap, they just can't rotate it down to where a beginner can understand.
Had a stats prof like this once who would march in and scribble jibberish. When a sheapish student tried asking what it represented to try to make some sense of things the exchange went like:

"um...what does the k stand for?"
"That's Kappa!" [turns back to the board believing he's answered in full]

(Struggled + bad teacher) = Unfortunately common, less of a struggle to deal with because students generally figure out they're full of crap. Can't do, can't teach.

236:

Not even sure where to start with this, it's certainly an interesting debate technique: double down on the original premise (careful to appeal to authority but provide no actual links or references); amp up the condescension; try a novel mixture of claiming the moral high ground added to some playground-like "I know you are, but what am I" statements; include some misrepresentation; finish with a leap of logic that borders on a non-sequitur.

If you'd bothered to read my post carefully and actually think it through, and then reply accordingly, I would extend you the same courtesy. But hey, isn't getting on your high horse and being shouty more fun?

237:

It seemed to me that the course had been carefully designed to impart only knowledge that was useful for passing exams, while making it very hard to use it for any other purpose, such as actually doing chemistry. I suspect that had I gone on to study it at university most of the first year would have had to have been devoted to going over the whole lot again and filling in the gaps.

Possibly it was an ill-thought-out attempt to avoid teaching schoolkids how to make explosives and bombs. Of course, those of us who were that way inclined just did it anyway, at rather greater risk than if we had known more about it. One chap made his own nitroglycerine, simply by mixing conc nitric and sulphuric acids and glycerine and hoping - I'd have done the same if I'd been able to snaffle the nitric acid - neither he nor I knew anything about that reaction's tendency to get out of control or how to avoid such an event, but he was lucky.

238:

Here you go ... Nov 2011 article!

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3057475/

New Trends in Gender and Mathematics Performance: A Meta-Analysis
Sara M. Lindberg, Janet Shibley Hyde, and Jennifer L. Petersen
University of Wisconsin – Madison
Marcia C. Linn
University of California – Berkeley

'Abstract

In this paper, we use meta-analysis to analyze gender differences in recent studies of mathematics performance. First, we meta-analyzed data from 242 studies published between 1990 and 2007, representing the testing of 1,286,350 people. Overall, d = .05, indicating no gender difference, and VR = 1.08, indicating nearly equal male and female variances. Second, we analyzed data from large data sets based on probability sampling of U.S. adolescents over the past 20 years: the NLSY, NELS88, LSAY, and NAEP. Effect sizes for the gender difference ranged between −0.15 and +0.22. Variance ratios ranged from 0.88 to 1.34. Taken together these findings support the view that males and females perform similarly in mathematics.'


And ...


The biggest gender differences remain un/under-employment including ability to land that first STEM job and earnings where females continue to lag.

http://www.nature.com/scitable/forums/women-in-science/the-pay-gap-in-stem-fields-19116412


Lastly, ....

There's also evidence of on-going gender bias in ability to get STEM research published. Nature did an editorial recognizing this Nov 2012.

http://www.nature.com/news/nature-s-sexism-1.11850


239:

Ta.

From the abstract (and skimming a little into the paper, 212 pages is a bit much to digest immediately!) this seems to offer more support to the idea that there's little real difference in the fundamental maths skills of boys and girls (as groups), and that the difference primarily exists as a preconceived perception. Which was part of my point.

Also (in further answer to EC's snark), I was not saying that everyone has equal ability, but that if you bring preconceptions about any group's academic abilities into teaching, you are doing many a disservice. But he seemed to miss that.

240:

To go one meta-level up.

Gender bias in how people interpret claims that studies find gender bias:

https://theconversation.com/men-and-women-biased-about-studies-of-stem-gender-bias-in-opposite-directions-48924

Randomize the reported result from gender-bias research without changing anything else and people change their assessment of the "quality" of research significantly, in both directions.

241:

I'm really sorry to hear that. In the U.S. the problem is composed of equal parts of greed, racism, and liberal-education foofiness. Is it just greed in the U.K. or is it other issues as well?

242:

The problem with things like Trig in elementary school is that you get into issues of brain development. Kids usually can't handle abstractions until they're past puberty. Obviously you can push brain development to some degree, but that's not always a good idea.

243:

Thanks, that makes sense
Though if 1,4-DCB is a nasty carcinogen and 1,2-DCB isn't
It is a little concerning that huge cakes of 1,4-DCB are utilised in men's bathrooms where they can sublimate/off-gas as odour controllers
Still, trying to make sense of the widespread common usage of probable human carcinogens is only going to make me depressed
The EU does seem to be slightly better on this front the UK

244:

A few years ago I was chatting with Professor Harrington (UofT) at a conference about this. His seminar was about framing questions, and one thing that stuck in my mind was that when they tested second-year university physics students only about half of them were reasoning at the abstract level.

I tested his ideas out in my classroom, and simply reframing questions to test the same concept at a more concrete level significantly changed pass rates.

245:

Nothing really crazy, but I just finished reading "Rule 34" and had a deja vu where Dr. MacDonald is giving his AI lecture.

Charlie, did you by any chance take the part about conscious, HAL 9000-style, AI being not very useful off some web forum? Because if you did, it is entirely possible I happened to write it.

246:

Which is the BBC Micro game you remember most?

None of them (no BBC Micros in the US) but was taken with the PC port of Elite for a few weeks. The shooting parts were goofy but the scale, both in-system and more widely, was compelling. Almost tempted to try the new one.

Was reading about wizard games in Red London last evening, FWIW. Do not however remember a 2d game like this. (Excepting the various 2D turn-based D&D games)

247:

Wait until you really understand genetic algos

I do. Obviously you don't really.

and what it means to writing code.

Nothing much. Working chromosomes-equivalents are too hard to design for anything interesting.

What will happen is a natural language interactive "I can't believe it's not AI" system that programs what you want for you. First doing the nitty-gritty of the code, eventually assisting in the design. Interactive is key, there, though.

OG.

248:

It's mostly profound stupidity - remember we are dealing with POLITICIANS here.
As is usual in these cases, the extreme left are as bad as, & in some cases, even worse than the right in this matter.
You (very largely) are not allowed to judge or act with/for your pupils on the basis of their abilities or interests. You must first judge their "class" & their racial/religious background(s) & promote or denigrate them accordingly.
Religious prejudice, especially if it comes from a country where the majority of people have brown skins is automatically given a free pass by some of the left.
This does especially girls from such backgrounds such a good start in life, handicapping them later on.
OTOH, many on the right are automatically against people from disadvantaged backgrounds, but more especially if people ask awkward questions (shades of Socrates) or read & think too much ( Yes, this is education we are talking about ... ) rather than doing as they are told(!)

249:

kids usually can't handle abstractions until they're past puberty.
CODSWALLOP

I was able to handle the idea of an abstraction of a different, imagined ( & by this stage obviously wrong ) imagined future, when I picked up my father's copy of "Last & First Men" at about age 9 ....
The idea that a symbol can stand for an object, or substance was also easily absorbed. It did help that my father lectured in Organic Chemistry, though.
What about musicians?
Written music is a set of abstract symbols, after all.

250:

Mostly political ideology driven == stoopidity, greed is a slight factor but only because thats one of the base concepts that underlie their ideology.

Biggest madness was tuition fees at degree level so we're probably following the US student debt bomb at 20-30 years separation. At least they tend to be discounted for STEM subjects, and it has undone the previous Labour government's ridiculous premise that "everyone deserved a degree" even if its a McDegree with no relevance for business. Which is not a problem in itself but when it was accompanied by a argument that it was necessary to get a job, but it had absolutely no relevance to any paying job there was a disconnect that caused extreme disillusionment.

Next madness was a return to examinations away from continual assessment ie handicap those that don't do well in 2-3hr pressure sessions.

Current madness is to force all schools to out of local control including top rated schools where the parents have already voted against leaving local control. Game plan is to have the schools then gobbled by educational conglomerates.

Having said all that - inspite of everything standard of education in my daughters school - not necessarily representative a high achieving school in affluent area - is probably better than when I were a lad.

Should be a law banning MP's from interfering with education.

I'm seriously considering moving Charlies neck of the woods when my daughters are approaching university age - no tuition fee's in Scotland...

251:

Charlie, did you by any chance take the part about conscious, HAL 9000-style, AI being not very useful off some web forum? Because if you did, it is entirely possible I happened to write it.

Nope. I formulated it for myself -- although not without reading around the field for some years and chewing over possible non-human-like forms of consciousness in some epic bull sessions with the likes of Karl Schroeder.

Incidentally, for more SF along these lines you might want to read The Red trilogy by Linda Nagata, which takes it a bit further than I did in some ways.

252:

Mostly political ideology driven == stoopidity, greed is a slight factor but only because thats one of the base concepts that underlie their ideology.

Disagree strongly.

It's all about greed: the political ideology is used to provide a veneer of respectability for criminal malfeasance in office, with the payoff deferred until after retirement (at which point the political beneficiaries take seats on the board of the companies that benefited).

Tuition fees at degree level are a symptom, not a cause: the disease in question is the financialization and commoditization of education as part of the systematic liquidation of the employee class: you don't expect a job for life any more, so employers feel no obligation to train you, so you need to obtain certification to say that you are in principle capable of self-training to the required level, this is handed to the universities as a proxy service, the institutions in question are then privatized and required to obtain funding by charging for access -- and have a constant supply of raw materials to milk, because without going into debt to get the sheepskin, the future-worker-unit cannot get a job.

The forced academization of schools ... you know that when a school is turned into an academy, (a) the parents lose representation on the school board, and (b) the title deeds to the school's real estate is turned over to the company taking it over without payment? It's a giant land-grab. Look for the academies to begin selling off their sports fields for housing development 5-10 years down the line -- that'll be the bust-out tactic for the more unscrupulous thieves.

If you want to move to Scotland, do so sooner rather than later: IIRC you have to be resident in Scotland for a full four years before you qualify for getting your tuition fees paid for. Also, from tomorrow, the Scottish Parliament gets tax-varying powers, and there is currently a lively debate in the run-up to the upcoming general election over how to change both income tax and council tax. (If you're on an average-or-lower income you can expect to be better off; if you're a higher income earner you might end up paying a bit more. Even the SNP aren't proposing to significantly raise income tax on folks earning less than £150,000 a year -- of whom there aren't that many.)

253:

you have to be resident in Scotland for a full four years before you qualify for getting your tuition fees paid for

Could some of that time have been back in the 70s?

(I'm now wondering if I could prove my period working at Dundee Uni)

254:

Assuming you're serious: I don't know.

(But I'm pretty sure you could trade your wee house in South-East Englandshire for an entire castle in the Borders, complete with helipad and Bond-villain shark tank, and maybe have enough change left over for a Tesla Model S.)

255:

Not so wee!

(What's scary is the current Zoopla price for the 6-room (not bedrooms — rooms) terraced house in Ealing we sold in order to buy the current place. Eep!)

It's under two years till the OH retires — we celebrated the two year mark in Cloud 23, and I'm far from certain that my employment will last that long either. So who knows — if the UK lurches out of the EU, but Scotland leaves England to stay in the EU, I'd be sorely tempted. I liked living on Tayside, though the Uncles and Aunts were living over Dumfries way.

(Now buried and/or cremated over there, so I no longer have any living relatives north of the border.)

I assume Edinburgh would be rather more expensive.

As for the Tesla, I'm musing whether it would make financial sense over a 10 year period (I've had my current car 10 years) to switch to the 3. The tax breaks are phenomenal, but I can't see those remaining if EVs hit the big time over here, simply because the hole in the Exchequer's intake would be massive.

(Last time I looked, vehicle road tax + fuel taxes combined brought in more than the outgoings on road and rail combined. Lose those two, and where does the money come from? I don't want the roads falling apart. I don't want the railways falling apart.)

What would happen with currencies, I don't know either. A Scotland inside the EU couldn't sensibly be on the same pound sterling as an England outside — I assume it'd either end up adopting the Euro, or coining a Scottish Crown to go with all the Scandi Krone.

256:

Re:castle.

I was curious about what sort of property goes for less than the price of a london apartment.

Turns out you can buy your own private island for comfortable less than the price of a normal 2 bed flat in zone 1 or 2.

Examples:

Choice of small pretty one:
http://www.privateislandsonline.com/islands/ropewalk-caye

Or a big ugly one:
http://www.privateislandsonline.com/islands/hobson-island

257:

What? No North Sentinel Island?

That first one looks as though it will soon be under water, so it is probably cheap for a reason.

I think I would prefer something a bit more hilly.

258:

I read the description of Red trilogy on Amazon, and do not think it would hold my interest. I am not a fan of military fiction . Can you summarize?

259:

Life's too short. Just go read the first book. (Hint: it made the Nebula shortlist, and it's not your typical MilSF.)

260:

You're talking about a very low level of abstraction here, (if it's abstract at all.) And musical notation is very concrete. You see this symbol and you plan this note, for this long. It's possible to be almost robotic about the whole thing (and of course very primitive robots have been programmed to play music.)

A symbol standing for a substance is a simple, one-to-one correspondence. This kind of stuff is very simple for kids.

And kids get the idea of stories, even if the story is not possibly true. So you're not quite dealing in abstractions yet.

Now take the idea from Algebra that an equation can have two different answers and they're both right, or the whole issue of variables, or get into Geometry and prove a theorem or move around on a grid, then you're getting into abstractions that kids have trouble with... it's a whole 'nother thang.

261:

If you're seriously thinking about a move to Scottland check online to see how long it takes to qualify as a resident. (Probably obvious advice, but I've seen folks in the U.S. get in trouble with state residency requirements.)

262:

Am familiar with a couple of content analysis studies of one month duration (each) of on-air TV (programming and ads, all channels incl. cable) which turned up a similar finding, i.e., male coders/analysts were less likely than female coders/analysts to spot/identify differences in amount of on-air time between males and females on screen. This was discovered by having another group of coders/analysts actually time how long different actors/presenters/voices were on air. This is why there are training sessions for any research study (using real live humans) because it's good practice to review and define what is what in the study plus do a few practice sessions to review observations, discuss similarities/differences so that team members can calibrate their observational abilities.


263:

Agree ...

Abstract thinking, therefore the development of the ability to think abstractly, is not done in a vacuum or in isolation from everything else that a person might know. Generally, the learning is from the concrete to the abstract, often proceeding as more and more concrete examples are encountered*. In this way, abstract thinking is a type of shortcut as well as a means of testing possible future scenarios without having to do time-consuming grunt work. From personal experience, what helped most were examples, demonstrations, and analogies.


From the YouTube videos I've watched e.g., Feynman, even the uber-gifted mathematicians used or dervied their own diagrams in order to help them visualize/explain relationships. Therefore IMO, it seems that some sort of tangible cue/clue is necessary even when teaching abstract subjects, or introducing novel abstract concepts to already practiced abstract thinkers.


Believe that in the 19th century there was a fad that serious mathematicians should only use symbolic notation (algebra) in their proofs but that this bias went out the window when topology became relevant in describing certain phenomena. (Elderly Cynic and/or Susan would probably know.)

* Encountered and feedback quickly provided ... remember that timely positive feedback is fundamental if learning is to occur.

264:

HowToProgram

Not quite utterly surreal but it is getting close to WTF for me. It had to happen sometime.

265:
Believe that in the 19th century there was a fad ... Elderly Cynic would probably know...

I'd never figured the Cynic was *that* Elderly!

266:

After attending a WorldCon discussion on education re: Finland adopted a new and in many ways opposite-to-the-West/US educational policy, I've been following their results. Pretty good so far! FYI - Finland was the top-scoring Western country in the last/most recent PISA tests.

http://fillingmymap.com/2015/06/08/the-three-real-reasons-for-finlands-high-pisa-scores/

https://www.oecd.org/pisa/

Excerpt:
Is memorisation a good strategy for learning mathematics?

PISA finds that 15-year-olds commonly use memorisation to learn mathematics. But if you think memorisation is most widely used in the East Asian countries that share a Confucian heritage and are “known” for rote learning, think again. Fewer 15-year-olds in Hong Kong-China, Japan, Korea, Macao-China, Shanghai-China, Chinese Taipei and Viet Nam reported that they use memorisation as a learning strategy than did 15-year-olds in some of the English-speaking countries to whom they are often compared.

267:

Apologies! ... I meant that Elderly Cynic appears to have a solid/deep math background. (Ditto Susan.)

268:

Just as long as you calibrate in both directions since as the study shows, people will err in the opposite direction as well.

270:

I note that quite a few, not-too-expensive (for certain values of ...) full spec "complete" are quoting 1Tb of RAM, as normal, which instantly tells you something ...

A LOT of these are used for things like store retail or doctor office systems where the "main" PC has multiple instances running and the various PCs around the store/office are really running VDI software back to the main PC. Single point of failure but it sure makes keeping things up to date easier. And cheaper. So most customers want do to things this way.

A lot of people using CAD software also do this.

271:

Ironically Windows MS office and Visio in a VM is more stable and usable than the UI horror fest which is Office for Mac.

I'm guessing you're not using Win or Mac Office 2016?

272:

Workflow-wise, I will note that my job is to write books, not sysadmin my own home network.

Constantly amazed at the number of people who think that it doesn't cost them anything to sysadmin their own network. Home AND small businesses.

273:

Office Win 2010 vs Office Mac 2011 from memory.

2016 is on the to do list as soon as I get round to ordering a copy on my company deal. Also was worried that it might be horror that is Office Win 2013. UI design decisions with CAPS and white on white menus? Thank you Sinofsky.

274:

1) Can anyone recommend a good storage device for my photo library? 6 TB right now and will grow. Ideally something that automatically backs up (Time Machine apparently can't handle directing difference source disks to difference backup disks.) And at least as fast as the SATA drives I have in the MacPro (I'm guessing that means Thunderbolt).

Check out macsales.com for various TB and USB stuff that will work with a Mac no hassles/drivers/bitfidling/etc... Now if you don't need super fast you can look at Drobo. I've installed a few ARECA TB Raid cabinets with 1TB SSDs. Really really fast. And mostly air due to the SSDs. Then there's the Promise type stuff but you're really talking serious money at this point.

275:

Ideally something that automatically backs up (Time Machine apparently can't handle directing difference source disks to difference backup disks.)

Look at Crashplan or similar. It just works for most situations.

276:

(IMO anyone who calls themselves a programmer who hasn't earned their BA stripes at some point is just a code monkey - but thats probably my prejudices showing.)

Totally. :)

277:

Thanks.

I was looking at a 16 TB (8TB x 2) G-RAID. Would there be an advantage to a 16 TB (4TB x 4) model? (Leaving aside reliability of supplier etc.)

278:

4tb drives are more mature technology.

I prefer more drives, but you're only going to get one parity drive (normally, anyway) with 2, 3, and 4 disk layouts, so that's up to you.

Performance, if that matters, will almost certainly be better with the 4 disk version.

279:

How much usable storage do you want and how will you deal with a drive failure?

Without answers to that it's hard to answer your questions.

280:

One more step toward quantum computing from the Univ of Cambridge:


https://www.sciencedaily.com/releases/2016/04/160404111559.htm

New state of matter detected in a two-dimensional material

An international team of researchers has found evidence of a mysterious new state of matter, first predicted 40 years ago, in a real material. This state, known as a quantum spin liquid, causes electrons -- thought to be indivisible building blocks of nature -- to break into pieces.

QC-related excerpt:

'The observation of one of their most intriguing properties -- electron splitting, or fractionalisation -- in real materials is a breakthrough. The resulting Majorana fermions may be used as building blocks of quantum computers, which would be far faster than conventional computers and would be able to perform calculations that could not be done otherwise.'


281:

I want more than I can afford, I suspect.

8 TB of usable storage is a minimum. I'd like that fully protected, so a drive failure doesn't cause me to lose data. As I understand it (and I may be wrong), with a RAID you replace the failed drive and the box copies the information onto the new drive so that (once it finishes) another drive failure won't cost you all your data.

10 TB (or more) would be nice, if it doesn't cost a huge amount more. But reliability is more important than extra storage.

282:

"...with a RAID you replace the failed drive and the box copies the information onto the new drive so that (once it finishes) another drive failure won't cost you all your data."

In general, yes. However, it can take days for the new disc to be fully integrated and working

283:

Yeah, that's been most of this year here in California. One thing I've been contemplating is how I've perceive reality. For me, it's spring now, but it's hot enough to be summer. What I've noticed (and I'm trying to figure out how to describe) is that the light is the same as any other spring, but the temperature is hotter. What's weird is that how much temperature affects the way I interpret what I see. For example, I may think that "wow, it's summer already, what are those springtime flowers doing blooming now?"

What's happening is that some of the plants grow and flower in response to temperature, some in response to light. The light hasn't changed but the temperature has, so things are getting confused. The tumbleweeds respond to temperature, but other plants (including some other weeds) respond to light, so the vegetation and what's blooming are changing.

284:

it can take days for the new disc to be fully integrated and working

Can you use the RAID while this is going on?

This one looked not too expensive:

http://eshop.macsales.com/item/OWC/TB2SRT20.0S/

20 TB (5 x 4) but only $100 more than the 16 TB model. Software RAID not hardware, but I'm not clear if that means it needs my Mac to do the RAID thing or if it uses software in itself.

No USB cable, not certain how important that will be once I copy the data from my old MacPro and assorted external drives.

Seems to be noticeably loud, which is a downside. (Personal quirk: fan noise brings back memories of a not-so-nice position decades ago, which still raises stress levels. Irrational but happens.)

A good intro/overview of RAIDS and trade-offs written for a Bear of Very Little Brain would stop be asking totally silly questions, if someone knows of one.

285:

And thanks to Dirk, David, and Sean for the advice. It's greatly appreciated.

286:
with a RAID you replace the failed drive and the box copies the information onto the new drive

You can have as many disk failures as you have parity drives. So if you have a mirror (two drives), you can have one disk failure. If you have normal RAID5 (one parity drive), you can lose one drive. RAID6 gives you two parity drives, and you can keep going (there's a performance penalty).

With either mirroring or RAID5, if you have to replace a drive, you're running with no redundancy until it has finished resilvering the new drive. So that's why multiple parity drives are attractive.

Most RAID systems do not have the ability to do a scrub (check all the blocks on all the drives), which can be a problem. I happen to know of a multi-petabyte system that was lost because of this.

287:

Most RAID systems do not have the ability to do a scrub (check all the blocks on all the drives)

Is this like doing a disk check from within Disk Utility (Mac OS)?

288:

First thing to do, and it's reasonably affordable is to make a baseline backup of your existing pictures collection onto an external hard drive or two. That done, find a good home for the backup away from where you live or work. Out of state is probably a good start. Two backups would be better, dispatched to separate addresses. If you can, recover one of the backups every now and then and update it, then send it off again.

Having done that you might want to look at a basic Ethernet-connected NAS unit for working storage rather than something expensive and exotic in the Mac world. A four-bay NAS with four enterprise-quality 4TB drives configured RAID6 should cost less than 1200 bucks, and anything else on your local network can also use it as a shared resource. It can live in a cupboard somewhere if noise is a problem.

289:

Woohoo, at last something that makes me feel SF-nal. I haven't had that feeling for years. This should be interesting and lead to other things.

290:

Specifically, was reading "A Gathering of Shadows: A Novel", second book in the Lila/Kell story. (No spoilers here. OK, Fire and Water.)

@Robert Prior, that RAID array looks fine. As suggested, you might want to work out how to do some other sort of offsite backup in case of fire or theft or whatever.

291:

Interesting update on that "2.2 above the record"

The final figures came in that night for the temperature. It was actually 2.5 degrees above the record. I thought I was up to speed on this thing, but I find I'm not mentally prepared at all. I expected to see regular jumps of 0.1 or 0.2. You know, Bannister runs the 4 minute mile. You expect that people running under 4 minutes will start to become more regular until every race is under 4 minutes. Each Olympics someone will go 5 tenths faster. Over decades you see it drop from 4 minutes to just under 3:50.

Then someone runs 3:10

A jump that most people had been predicting either wouldn't happen or would take 8-10 decades.

And it somehow doesn't make the news.

The local news didn't mention it. I think that was perhaps the most startling thing. The record was broken by 2.5 degrees and it didn't even rate a mention in the weather reporting part of the local news.

292:

As for the Tesla, I'm musing whether it would make financial sense over a 10 year period (I've had my current car 10 years) to switch to the 3. The tax breaks are phenomenal, but I can't see those remaining if EVs hit the big time over here, simply because the hole in the Exchequer's intake would be massive.

I would tend to say to hold off.

Somewhere, probably over the next 5 years, someone is going to have autonomous vehicles sorted. The upheaval that will bring, not only to what people want, but to the very concept of ownership at all, will make a mess of 10 year predictions. It will also make EVs generally a lot more practical. Then there's charging stations/parking for cars currently not in use.

I'm not even sure that the existing manufacturers will be able to adapt fast enough to survive, nor what shenanigans they will create via lawyers.

293:

8 TB of usable storage is a minimum. I'd like that fully protected, so a drive failure doesn't cause me to lose data. As I understand it (and I may be wrong), with a RAID you replace the failed drive and the box copies the information onto the new drive so that (once it finishes) another drive failure won't cost you all your data.

I've got 9TB of protected storage (4x3TB disks in a microserver housing) via UNRAID. Total cost ~£400.

UNRAID has its own way of doing what is equivalent to RAID5 (with some advantages) and will do the equivalent of RAID6 in the new version. Also runs VMs, Dockers, and Plugins, etc. if you want.

You can also roll your own Synology box, or use something like FreeNAS. IMHO a separate NAS box is the way to go unless you are dealing with high bandwidth stuff like RAW 4K video.

294:

This one looked not too expensive:
http://eshop.macsales.com/item/OWC/TB2SRT20.0S/
20 TB (5 x 4) but only $100 more than the 16 TB model. Software RAID not hardware, but I'm not clear if that means it needs my Mac to do the RAID thing or if it uses software in itself.

No USB cable, not certain how important that will be once I copy the data from my old MacPro and assorted external drives.

This box is ThunderBolt only.
They sell it as a bare box or populated with drives. Also it comes with or without SoftRAID which is software that runs IN YOUR MAC which does the RAID work.
http://eshop.macsales.com/search/thunderbay
I suspect you more likely want an NAS or RAID cabinet with built in hardware RAID.

If you're looking at 3.5" drives then these were the sweet spot recently. I built an 8 drive RAID6 with them and a ARECA 8 bay cabinet.
WD30EFRX
http://www.newegg.com/Product/Product.aspx?Item=22-236-344

I'm thinking if you don't need video editing speed you likely want something like a Drobo which deals with mirroring and RAIDing behind the scenes. You can add or remove drives at will and it "just works". (Or so they claim. :) )

295:
Is this like doing a disk check from within Disk Utility (Mac OS)?

No, disk check is fsck, and only checks metadata.

What you want in a RAID system is for every block to be checked, and compared with the parity blocks to ensure they're all still correct. This will also let you know if there are any bad blocks.

(Well, hm. fsck_hfs may have an option to check every block. I forget whether that made it into the file system or not. Hey, it does! -S.)

(I forget all the stuff I've written.)

296:

I would tend to say to hold off.
Somewhere, probably over the next 5 years, someone is going to have autonomous vehicles sorted. The upheaval that will bring, not only to what people want, but to the very concept of ownership at all, will make a mess of 10 year predictions.

Except, things being what they are, just BEFORE this happens, all diesel cars etc will be banned completely, just as 90% will go of the roads, because of "electric".
Or something equally stupid.
Politicians & engineering ... do not nix.

297:

... and subsequent comments from both you & heteromeles ...
Can I repeat a plea here, for UK readers?
Look up "Nature's Calendar" a mass-observation science data-collection study, associated with The Woodland Trust?
( Yes, I'm one of their recorders )
They are also getting very concerned about this.

298:

A reminder: Tesla cars on the road got some self-driving ability via software update last year. What hardware change are you expecting in the next 5 years?

299:

And the dashboard (or lack of it) on the 3 has led some commentators to consider it the possible format for self-drivers.

300:

The sooner all diesels in London are banned the better

301:

Tesla cars on the road got some self-driving ability

A friend took one for a spin last month. Some limitations:

1) It apparently uses the shoulder lane markings to steer. So when there's an exit lane in the road (on the right in Canada) it pulls into the exit lane (and will take the exit).

2) It can only handle a limited turn radius. So if the curve in the road is sharper than that, it will hit the ditch. (The salesman had to warn David to take the wheel at that curve so a $100k+ car didn't wreck itself.

302:

Yup. That's it. Please extreme and fanatical monetarism, along lines that would gladden the heart of Ayn Rand, and an utter loathing (by central gummint) of local authorities, or any other local public organisations that might not jump when Whitehall says frog.

303:

"I'd never figured the Cynic was *that* Elderly!"

I'm not, though my daughters always accused me of be contemporary with the dinosaurs :-) It wasn't so much a fad, as that symbolic algebra had become as dominant as Euclidean geometry was pre-Newton. While topology itself dates from the 18th century, it wasn't used as a tool until more recently. Even in my degree (late 1960s), the use of topology as a tool was only just touched on.

There were a LOT of changes in mathematics in the 19th and 20th centuries, as ideas that had been developed by the originating geniuses became familiar enough to 'ordinary' mathematicians to start to be used as tools.

304:

Since you are neither a bigot nor a troll, I will respond. I am aware of all that (and the bias mentioned by Murphy) - and, if you look at my posting, you will see that I did NOT say what I am being accused of saying. I have better things to do than to try to, er, debate with those two. There are a few points that you might appreciate.

The data I referred to is well-known, but its authors have to keep a low profile because of the fanatics (seriously). But you could start with Baron-Cohen (who is a bit of a nutter, but a serious psychologist) and move on from there. Or look for the PET scan data. And please note that I did not say what those differences were.

I mentioned ability + preferences and the lack of STEM status. That is also related to the problem that it is much harder to combine childbearing and child-rearing with a career in STEM than in 'management'. I have seen the harm caused by that and by the 'compensating' preferential treatment at first-hand.

The second is that the people who make careers in STEM are mostly from among the best, and those who will 'get places' even more so. I believe that you know enough statistics to know that, if you take a sample from above the 99.9% point of a distribution comprised of a N(0,1) population and a N(0.01,1) one, you will get a rather different result from if you take one of the whole population.

305:

Somewhere, probably over the next 5 years, someone is going to have autonomous vehicles sorted.

Some of us don't have 5 years to wait for our next vehicle. I'm optimistic that my current vehicle (20 years old) will last me another 6 months.

Plus while I'm amazed at what autonomous cars vehicles can do now I'm skeptical of your 5 year mark. Currently they do all the easy stuff in more urban settings but have a long way to go to deal with the hard stuff. Like night, rain, snow, ice, higher speeds, etc... For about 2 years now I've been paying attention to where I drive to imagine how an autonomous vehicle would deal with my needs. While it would cover 95% of my driving time that remaining 5% would make a serious dent in my life if I had to give it up. Just things like moving cars around in my driveway are way over the top for current stuff.

Over a 4 day weekend in the northwest over the last few days I did about 500 miles around Seattle and Vancouver (BC) plus a side trip to Mt. Rainier. I can't imagine current tech coming close to handling what I did multiple times per day.

306:

You are enjoined to indulge in Sex & Travel at the earliest opportunity.
You wish to order me to give up my Land-Rover - for what?

PRACTICAL suggestions, please, that I can afford to actually run.
"Not even stupid", I think might be the expression here....

307:

Greg, I really think you might enjoy yourself more if you traded in your Land Rover for one of these. (NB: Yes, you really want to run the video ... :)

308:

Heh heh. I've got a new anecdote about automated guidance.

Over the last weekend, I volunteered as part of a garden tour that focused on gardens that had a lot of native plants in them.

The fun part was my garden was way out east in the mountains, out of cell phone range. My iPhone could hold out to get me there, but not get back (Siri said, and I quote, "I don't know where you are."). The real fun started because this place was halfway down a dirt road that made a loop with the paved road. One side of the dirt was about a half mile shorter than the other, but was much more rutted. That was where the iPhone wanted you to go. In the printed directions to the garden, they warned everybody not to trust their GPS, but to follow the paper directions exactly for a smoother ride.

Here's how it broke down:
--some people followed the paper directions, and had a smooth ride.
--Some people insisted on following their phones. One couple even insisted that we'd put up signs wrong, because we'd put up signs to warn them away from taking the bad section and they were so lost they thought that was the good section.
--About five couples relied on their old Garmin GPS systems, which took them ten miles out of their way and tried to force them to use a long-closed road across the property of a new age monastery. They had some interesting experiences with the, erm, nuns and monks of that monastery getting them redirected.

Now, about automated cars: I fully get that they'll have happy little GPS units instead of cell phones, but how well will they work away from cities? This place turned out to be a great test for the guidance systems of current units, and quite honestly, they failed. Most of them aggressively default to the shortest map route, even when it's impassable. I suspect this not-so-little problem is going to make it much harder to get people to accept driverless cars in large swathes of places like the US.

It could easily get tragic, too. For example, climate change is predicted to cause increased flooding and storm damage in large parts of the US. How is a little driverless car going to deal with flooding roads with downed trees across them and the power out? Or with flag men on road repair operations? Or even with badly rutted dirt roads that are perfectly passable on the map? How complex and environment can these cars handle before they kick the driving back to a human?

309:

I see we're discussing education and tech.

Some of you may be interested to know that the West African state of Liberia has privatised its education system by outsourcing it to an American firm.

http://mgafrica.com/article/2016-03-31-liberia-plans-to-outsource-its-entire-education-system-to-a-private-company-why-this-is-a-very-big-deal-and-africa-should-pay-attention

It's a bold strategy, let's see if it pays off for them.

One possible bug (or feature, to some, maybe) is that the teaching cadre will be reading pre-planned lessons from tablet computers, and will have only five weeks' training before being sent into action.

I have to say I'm very sceptical about this one. . . it might have use as part of a long-term transition to a proper and effective public education system of the kind that is so sorely lacking in Liberia today. That's probably not a vision shared by the company that's going to provide this 'public' good, though.

310:

That solution has been proposed for the UK by the sort of stink tank that the Conservatives listen to. I am not joking.

311:

I know you're not joking!

Alas.

In the first volume of Capital, Marx quotes from Blue Book research on schooling in industrial England, schooling that involved (in many cases) "teachers" who were barely literate, or had comparable defects.

I'm not sure either Liberia or UK will benefit from going back to the future (and they won't even have a cool DeLorean to do it in).

312:

Appreciate your trust!

Agree that ... 'you will see that I did NOT say what I am being accused of saying.'


Agree that ... 'Baron-Cohen (... a nutter, but a serious psychologist)' ... problem is that there weren't any similar looking results from other researchers. Mostly, he could have picked a better name/explanation for his theory (autism=extreme masculine brain). Even so, his theory may gain some traction provided it can be meshed with other recent findings re: androgens, i.e., digit ratio as in longer ring finger=more in-utero testosterone, therefore likely to also to result in increased male-type traits in both males and females. Lastly, there's a recent article summarizing sex difference findings published by another autism researcher, Laurent Mottron from Univ of Montreal. (I'm noticing that the more potentially controversial the findings, the more abstruse and technical the language used.)

http://www.ncbi.nlm.nih.gov/pubmed/26052415


Mol Autism. 2015 Jun 5;6:33. doi: 10.1186/s13229-015-0024-1. eCollection 2015.

Sex differences in brain plasticity: a new hypothesis for sex ratio bias in autism.

Mottron L1, Duret P2, Mueller S3, Moore RD4, Forgeot d'Arc B1, Jacquemont S5, Xiong L6.

Abstract

Several observations support the hypothesis that differences in synaptic and regional cerebral plasticity between the sexes account for the high ratio of males to females in autism. First, males are more susceptible than females to perturbations in genes involved in synaptic plasticity. Second, sex-related differences in non-autistic brain structure and function are observed in highly variable regions, namely, the heteromodal associative cortices, and overlap with structural particularities and enhanced activity of perceptual associative regions in autistic individuals. Finally, functional cortical reallocations following brain lesions in non-autistic adults (for example, traumatic brain injury, multiple sclerosis) are sex-dependent. Interactions between genetic sex and hormones may therefore result in higher synaptic and consecutively regional plasticity in perceptual brain areas in males than in females. The onset of autism may largely involve mutations altering synaptic plasticity that create a plastic reaction affecting the most variable and sexually dimorphic brain regions. The sex ratio bias in autism may arise because males have a lower threshold than females for the development of this plastic reaction following a genetic or environmental event.

STEM - would agree except that nowadays post-doc (and academic tenure-track) positions are fewer and at some institutions, of shorter duration. Given this, to me, this should mean that the gender and/or child-bearing/raising bits are completely irrelevant. (Personally know folks in these circumstances: babies between postings/jobs/grants. If both parents are researchers, they juggle schedules just like everyone else. One researcher couple: one was at a major univ that had affordable daycare for staff and students which really, really helped!) Personally, I think that although most academic researchers are forward-thinking folks, such obvious gender bias indicates that these folks are not equally forward thinking in all areas, esp. their own lab/staff. They need to do better.


Outliers (0.1% vs. 0.01%) ... wish I knew a way of comparing outlier results. Unfortunately all my stats knowledge/experience has been the 'You must get rid of all outliers otherwise they'll mess up your data! variety. As you're probably very aware, most research (commercial and academic) focuses on findings that can be generalized, i.e., discovering the 'norm', mean, average, etc. That said, any suggestions for learning how to parse outlier, extreme ratings, etc. would be appreciated! (Also slightly familiar with/remember Poisson dist'n and Fisher contingency.)

313:

"WHERE DO I START?
I will need a chip & a motherboard & there seem to be 200 varieties of each."

read Sterling & Gibson's "The Difference Engine", given your interest in steam power it might inspire a useful search strategy.

314:

Good grief, why don't they just install a robot/talking head on a TV screen at the front of the classroom and be done with it. The teaching philosophy [below] looks completely top-down with no interaction and no opportunity for individualized learning. The real-time monitoring of all the kiddies by some unseen, remote head office staffer is also disturbing.


http://www.bridgeinternationalacademies.com/academics/philosophy/

Excerpt:

'Our scripted curriculum includes step-by-step instructions explaining what teachers should do and say during any given moment of a class. This allows us to bring best-in-class instruction, international and local research, and curriculum specialists into every one of our classrooms. Essentially, we’re giving our pupils access to the types of teachers they would never be able to afford. Plus, this system allows us to standardize our high-quality instruction across all of our academies. Teacher scripts are delivered through data-enabled tablets, which seamlessly sync with our headquarters, giving us the ability to monitor lesson pacing in addition to providing the scripts themselves, recording attendance, and tracking assessments in real-time. We also create our own books, manipulatives, instructional songs, symbols for enforcing positive behavioral management, and more, which we are able to produce locally at an extremely low cost. Combined, this makes for a very engaging experience for pupils – a far cry from the rote learning that happens in most of the other schools found in our communities.'

315:

types of teachers they would never be able to afford

Ie. teachers who ignore their pupils and just reel off a memorized lecture.

316:

"Baron-Cohen - there weren't any similar looking results from other researchers."

Yes. I haven't read his book, but have heard him speak. It wasn't his autism theory that I meant, but the fact that there are relevant differences (e.g. in the visuospatial and 'social' areas) that explain why at least some of the bias is fundamental. God alone knows why some loons won't accept that difference is not the same as inferiority.

"(I'm noticing that the more potentially controversial the findings, the more abstruse and technical the language used.)"

Right. And this area is a HELL of a lot better than race. Why are there more sub 2 hour marathon runners from one tribe in Kenya than the whole of the USA? Well, the geneticists know, but publishing is more than their career is worth - and I have heard that from more reliable sources, too.

http://www.npr.org/sections/parallels/2013/11/01/241895965/how-one-kenyan-tribe-produces-the-worlds-best-runners

"Outliers (0.1% vs. 0.01%) ... wish I knew a way of comparing outlier results."

Use non-parametric statistics. Easy and pretty reliable. Basically, Wilcoxon, Mann-Whitney, chi-square etc. You can do it parametrically, but it's terribly sensitive to the exact shape of the distribution, which is often not what you think it is. I should be amused to know Susan's view - which might not be the same!

317:

Sorry - I omitted this.

"STEM - would agree except that nowadays post-doc (and academic tenure-track) positions are fewer and at some institutions, of shorter duration. Given this, to me, this should mean that the gender and/or child-bearing/raising bits are completely irrelevant."

Absolutely No Way. Sorry. And I speak from personal experience, lots of direct observation etc. Inter alia, people who drop out have to do a hell of a lot of catching up - both in their subject and politically.

"Personally, I think that although most academic researchers are forward-thinking folks, such obvious gender bias indicates that these folks are not equally forward thinking in all areas, esp. their own lab/staff. They need to do better."

Right. A lot of it is the politics. More men seem to be happy with and good at the thoroughly nasty politics that is needed to get above lecturer level. A lot of men aren't, and that's where we (yes, I was one) get stuck, too. The official feminist dogma is that this is sexist discrimination, but that is only sometimes the case - and less often than it is claimed to be. Reducing it would have massive advantages, not just in balancing the gender ratio, but that won't be done while the cause is misrepresented.

318:

Sterling & Gibson's "The Difference Engine" is a fun read. Reminds me of Bruce Sterling’s “Dead Technology” project, where he was cataloging technology that had gone out of use or replaced with new tech. The IBM electric typewriter would be an example of “Dead Technology”.

I was told by a friend working on campus about tasking a student worker to inventory a storage closet. One of the items the student listed was a “key board machine”. She asked him to show her this “key board machine” and he pointed to an IBM electric typewriter.

319:

*Polite cough*

Has a comparative study been done between male/female subjects from different cultures. e.g. USA birth date 1950-1970 vrs CCCP? (Given that we know the % differences in STEM subjects such as engineering was 1% vrs 40% in 1960).

You might find that different cultures / teaching methods / socialization causes different brain structures.

Hint: the plasticity works in many different ways.

~

It's why I find most humans to be utterly alien[1].

[1] For given subsets who aren't actually alien / hivemind / servants of elder gods / Justin Bieber fans

320:

*cough*

Mice raised in the enriched environment also showed changes in the expression of genes involved in formation of new synapses and reorganization or strengthening of existing synapses. For example, the gene encoding integrin alpha-4, which is important for neuritogenesis and neuronal plasticity (21), is induced 3-fold after both 3 and 6 h of training. The expression of GTPase RhoA is also increased after training (3.2- and 3.6-fold after 3 and 6 h). In neurons, Rho proteins are involved in the induction of integrin-mediated events in surface adhesion and proliferation, thereby contributing to synapse formation and neuronal plasticity (22, 23). A cluster of genes encoding proteins involved in synaptic vesicle trafficking and neurotransmitter release, including synaptobrevin and clathrin-AP2, was up-regulated after 3 and 6 h of enrichment. Synaptobrevin is a synaptic vesicle-associated protein, whereas the clathrin–adaptor protein AP2 interacts with synaptic vesicle protein synaptotagmin and regulates vesicle exocytosis (24, 25). Changes in the expression of these genes clearly suggest that presynaptic processes are being modified by enriched experiences.

Effects of environmental enrichment on gene expression in the brain Full paper, non-PDF. Claire Rampon, Cecilia H. Jiang, Helin Dong, Ya-Ping Tang*, David J. Lockhart, Peter G. Schultz, Joe Z. Tsien, and Yinghe Hu; Genomics Institute of the Novartis Research Foundation, San Diego, CA 92121; and the Department of Molecular Biology, Princeton University, Princeton, 2000

~

It's still early days, but perhaps in 100 years (if anyone is still around) humans will view the child raising techniques of late 20th C societies as utterly barbaric.

Put another way: enrichment also has a negative side.

So if women and men do not have such innately dissimilar brains, why do they seem so different? Parvizi explained that the brain exhibits significant neuroplasticity, as it able to make structural and functional changes in response to environmental inputs; in his words, “the brain is molded by experience.” As the brain replicates the same signals over time, the networks through which they are sent become progressively stronger, as repetition reinforces both the networks and brain synapses. Building off, “…what we know about the neural basis of learning, one can argue that the map of associations in the brain is sculpted by our experience throughout our life. Even if the hard wiring of the brain remains unchanged, the function of the hardware is constantly altered by experience.”


Is the female brain innately inferior? Stanford neuroscientist tackles myths about the brain
Clayman Institute of Gender Research, Nov 2011


And yes, I did choose that source because it's related to straw men.

321:

"Has a comparative study been done between male/female subjects from different cultures."

Not as far as I know, but there have been studies in more than one culture and on very young children.

"You might find that different cultures / teaching methods / socialization causes different brain structures."

You MIGHT? Hell, you DO, and there's lots of data on it. It's why I regard the measurements of the basic characteristics as being more reasonable than the 'X is better at Y' ones.

"It's why I find most humans to be utterly alien[1]."

Join the club. I find them totally unnatural and perverse.

322:

You can add it to the collection of things which indicate your 'news' setup is broken, you being the USA. Ours isn't that much better of course.

An addition to what Heteromeles said re. plants blooking earlier etc. It's not just plants, it's what the change in their behaviour does to the insects and bird and everything else in the ecosystem. If plants flower 3 or 4 weeks early but the insects which use them aren't also coming out quickly, or got killed by a snap frost, then there's a problem. In the good old days there was enough undisturbed ecology that local die backs didn't matter, but now we've concreted over so much and split everything by tarmac populated with mass killing machines, there isn't the surplus available to deal with such a problem.
A naturally conservative person would wonder why we are running so fast into the unknown, but that word has lost most relevant meaning nowadays.

323:

To be clear, and not fall into parochial binary hate groups (*cough* innocent look at puppies *cough*), this isn't to state that any of the science is incorrect.

Both the following statements can be true at the same time:

We have found that within culture X that brain structures show a gender disparity in formulating process A

and

Culture X is structured in such a way that a gender imbalance is produced when considering process A.

And, of course, the really bright ones of that gender tend to buck the over-all population trends anyhow.

I think there might be one or two reading this very discussion. *nose wiggle*

324:

Considering = Teaching, Training up above.

Bad translation.

325:

Have heard that slim 'ankles' (fetlocks?) and shorter lower legs indicate a faster horse, so why not a faster human. Do not buy in to the rest of this story though - torture under any label is questionable.


Non-parametric tests ... interesting how these tests go in and out of fashion and primary/best-for usages. Used to be that if you 'had to use' M-WU or Wilcoxon, you probably hadn't designed your psych experiment properly or your sample size was so small as to be of questionable value, i.e., findings easily overturned simply by virtue of someone using a larger sample the next time around. Now, non-parametric tests are all the rage for crunching mega (incl. meta) data often because the normal/bell curve distribution is not known/assumed to exist for that variable.

326:

We may or may not be having two arguments here. You (and please correct me if I mischaracterise you) are arguing that there are fundamental sex differences. I am arguing that if there are, any impact of such differences on 'output' (however defined) is buried in the noise compared to the impact of a society which tends to devalue women.

For evidence of the latter, consider (for instance) many orchestras are moving towards blind auditions because being able to see a performer is a woman has a measurable impact on selection panels' likelihood to hire them.

327:

Read that a couple weeks ago, was amused by a certain sailor called Stross.

328:

Re: 'Inter alia, people who drop out have to do a hell of a lot of catching up - both in their subject and politically.'

- Agree for 'politically'. Since you're familiar with that environment, please explain how this works for, say, a male prof returning from sabbatical*. It seems bizarre that a one-year sabbatical is qualitatively different from a one-year mat/pat-leave. Being on sabbatical does not mean spending all one's time reading related lit, or even doing anything academic at all. Just as being on mat/pat leave does not mean that you're no longer able to read any papers or think deep thoughts.

* Personally know two tenured profs who extended their sabbaticals beyond one year - one to found/run/sell a company, the other to climb a newly IPO'd corporate ladder. No penalties - academic, income or professional standing - whatsoever.


329:

It's not just places like the US, there have been plenty of stories of such events in the UK as well - and we have probably the best mapping in the world thanks to the Ordnance Survey. People treating a level crossing as a crossroads and driving down the railway, or something similarly daft, because "the sat nav told them to". The one I remember best involved someone going down a "road" that was nothing more than two ruts hidden in the grass, and then over the cliff at the end of it and onto someone's house.

The tone of presentation of these stories was one of pointing and laughing at the people who did it for being so hilariously thick. There must be innumerable instances of other people also being given such impractical directions by the machine, but without any such disaster ensuing, as they had the noddle to see that the machine was being silly. Take the human intelligence out of the loop and we will probably see rail transport becoming unworkable, and houses under cliffs considered unsafe for habitation...

330:

Another one, "Causal Angel" had a ship named "Bob Howard" (not a Culture style ship, more of a sys admin ship.)

331:

It could easily get tragic, too. For example, climate change is predicted to cause increased flooding and storm damage in large parts of the US. How is a little driverless car going to deal with flooding roads with downed trees across them and the power out? Or with flag men on road repair operations? Or even with badly rutted dirt roads that are perfectly passable on the map? How complex and environment can these cars handle before they kick the driving back to a human?

Self driving cars will NOT be a thing outside of highly urban areas until the cars can deal with local current conditions. No "Google Map" is going to let a car know how to maneuver behind my house so I can attach my trailer to go get some mulch at the county yard waste facility. Neither my yard nor the facility has anything like a road mapped out.

Or when is it OK to "double park". For someone to run in a drop off something in less than a minute. Or how about what happened today to me. My daughter picking me up at the airport. 4 lanes in front of the terminal. Two closest for loading/unloading and the other two to "drive" in.

I was driving around Vancouver BC this weekend. Rules are very close to the US. Ditto signs. But the differences can be killers. I went the wrong way down a one way street (empty!) as it took me a while to learn where to look for the signs. Plus lots of other little details. And the approaches to the customs stations... I seriously doubt they would be handled very well. Especially when the big sign above one lane switched from "Open" to "Merge Left".

332:

Am I the only one to be horribly awestruck by the intersection of these three things, all happening in the next 10-20 years:

"Self driving Cars"

"US infrastructure needs at least 3 trillion dollars to not fall apart"

"Climate Change kicks off big time"

One of these is not like the other.

Note: in a sane world you'd rebuild the infrastructure to utilize the new tech (been done before, looking at trams & cars, the initial dirty battles and racist rezoning of entire cities by highways, at least in the USA) while pulling out of the obvious casualty areas (aka abandon Flint all ye who enter as well as Florida / Keys).

~

Putin creates new National Guard in Russia 'to fight terrorism' BBC, 6th April, 2016

For "terrorism" you could also read "waves of climate migrants" or "refugees from the little 'stan war getting hotter" or "how to maintain order when Siberia gets a little warmer".

~

I can source documents, as ever.

Anyhow, more coral went zip and Tigers are now officially extinct in Cambodia, and remember those beached sperm whales?

Post-mortem on thirteen dead sperm whales finds their stomachs full of plastic
Telegraph, 29th March 2016.

~

Oh, and if you want serious storage coupled with cloud access, you could look into Amazon's Snowball. At 50tb might be a bit overkill though.

333:

Unfortunately, what I said does hold. A lot of the important (i.e. leading edge) data is unpublished, and obtained from personal contact. And it's much easier for eminent professors to take time off (in both respects) than mere lectureers or below. Sorry, but that's the reality.

334:

My point was that the geneticists knew what the factors were and did not dare publish.

335:

Can you be more specific about dates at least?
Because a cursory google scholar search turns up a fair number of publications about genetic factors that may or may not be the cause of such dominance in a specific sport.
e.g.

https://www.researchgate.net/profile/Yannis_Pitsiladis/publication/40687707_ACTN3_and_ACE_genotypes_in_elite_Jamaican_and_US_sprinters/links/54abb5850cf2bce6aa1d9b2f.pdf

from 2009. Many seem to be behind paywalls of course, and another interesting one refused to download.

336:

What's so surprising about this? Going back to the 1960s, it was normal for Stanford professors to porpoise in and out of academia. They'd get a bright idea, found a company to commercialize it, go broke working out that it couldn't be commercialized, then go back to full-time teaching until their next bright idea came along. Yes, Silicon Valley got a big boost that way, but there's no reason to be surprised that someone thinks some invention will make him (or less often, her) rich. AFAIK they're often wrong, in which case it's good that they have that teaching gig and grants to fall back on.

337:

On the whole BBC Micro Bit/Rasberry pi question I'm not really against either. Though I'm thinking prehaps due to its complexity the rasberry pi is better off being targeted at older kids. As for the BBC Micro bit, it more strongly resembles something that might be better taught in a science class than a more general IT/computing class.

I've often thought though that one really good learning/teaching aid might be the idea of a "new old" computer. I mean sort of re-implementing one of the old computers from the 80s - I don't mean a carbon-copy per se, but sort of the things that made them work if you will.

Let me explain. One thing about all those 80s "console computers" was that they seemed unintentionally - even the more obscure ones - to have made an amazing teaching aid about computers and gave a lot of people a lever to get into computers and writing their own programs. A door which wasn't too hard to open; just a little push and you're in.

Sometimes for things like this simplicity - but not *too* simple works. The rasberry pi is in this respect way too complex to learn everything about; that would be too much I suspect for most adults, let alone kids. But those old 8-bit machines from the 80s seemed to get the balance "just right" - not too hard to learn all about, yet you got intresting/useful/colourful results for minimal input.

I mean it is simplicity in iteself to - even today - find (say) an old C64 and make your name appear in lots of colours in well under half a dozen lines of code. How many lines of code would you have to write to do this on a modern system running C++? (I'm reminded at this point of an old joke about who can come up with the result to a sum - a C64 user or a Windows user, and the C64 user'll have the answer before windows has even started up....).

Memories of kids leaving rude messages on the computers sold in shops back then in the day and then locking up the keyboard....ahem....!

I know some here will probably say "oh but making such a system today .... it is old fashioned, too limiting, nobody would use it". Well probably true for older (say 13 or 14 year old) kids, but for a 7 or 8 year old? Less likely. Younger kids today are no different to us as when we were the same age.

If you have kids yourself (or you know someone who does) and they're younger how many times do you see them doing something unexpected? I mean for sure you can spend a fortune on the latest phone or tablet, and yes it can play music or videos and games....but then your kids will do something you might not have expected - namely play with the boxes!

A phone or tablet might be able to play a game, but a small cardboard box can become a space communicator....a larger box can become a car or (if you crawl under it) a pretend house. Heck even something that isn't nececcerily a present can become like this - a pile of cushions could become a 'den' for example. So sometimes super-expensive flashy modern devices aren't always the best.

For sure a lot of those old computers ran BASIC (and often the terrible MS Basic) but machines that ran (for example ) BBC Basic - that isn't prehaps such a bad thing - ok that language might not be so good for super-complicated programs but if you're just starting to learn and figure out stuff then it makes it easy to get "your foot in the door" so to speak. Sometimes it could be so easy if your program was short and you didn't have any storage (when I first had a C64 I didn't even have the datasette!) you could just *remember* your program - !

I also wonder though with regards to learning about computers themselves how much of it happens rather in the home than school? From memory when I was at school in the 80s the school I went to was a little slow on getting computers into the classroom; I'm from the UK - in my last year at junior school the school got a whole *one* BBC Model A computer in a classroom, and that could only be used by teachers.

In the mid 80s once I got to secondary school which was larger they had BBC Bs there, though still not many -- just under a dozen I seem to remember (and guess what I used to do each break or lunch time - sometimes I missed lunch altogether trying stuff out on those BBC Bs!).

Not many other kids though I seem to remember (don't know what it'd be like today) had much intrest in those computers though. For one thing it was belived back then that computers were a "male only" subject (not true), and even then most boys didn't use the computers. Most were intrested in other things, for example football (Me, I can't stand football - yawn).

A "sort-of" re-implementation of those old computers might be a useful idea as they give intresting results with little input which I guess is the aim if you're teaching kids who might not know/haven't used before. For sure we don't want some of the problems -- try typing in basic "PRINT 100.1-100" on some of those old computers and get a surprise, or sometimes even "PRINT 5.517". I'll leave it for others to try it! And nobody would want to go back to struggling with (eg ZX Speccy) cassette tapes and volume levels.

>Which is the BBC Micro game you remember most?

None of them. I had a C64 before I was allowed to use the schools' computers and that's where I learnt how to program a computer - not at school nor at home. And intrestingly looking back (and I'm really *not* trying to boast here) I had zero help from teachers or parents on doing this. (On BBC Micro games, the C64 games always seemed to be better; that's subjective I guess but that leads into the 'playground computer wars' and debate over which was better; the c64 or zx spectrum but that's for another time).

I can well remember hooking up the C64 to the TV (black and white - !) I had at the time, and then typing the inevitable thing you almost always try at that age when you had a computer back then, you type "HELLO". You then get back "?SYNTAX ERROR READY." . Not long after this I found things called commands, and then programs and that you could change things in the program.

So you could write a loop to print something on the screen (say) 10 times, but what if you changed it to 11 - would it work? Or what about 20 or 50? Would the computer blow up if you changed it to print the message 1000 times over? Most of us who tried this sort of thing just went by the manual and the computer magazines of the day.

As for the games I remember the most well .... I remember several. But the first 3 games I had (not in order; my memory has long become too fuzzy to remember precisely) - "Monopole 64" (A computer implementation of the UK version of the board game monopoly - used to like the sound the computer made if you got 'sent to jail"); "Gridrunner" - Very simple but very fast Jeff minter game; "Manic miner" - miner willy and his little waggly legs(!).

Apologies for going off topic btw, I bet anyone who read this in full will probably be asleep by now....

ljones

338:

>....and that's where I learnt how to program a computer - not at school nor at home

It should read "and that's where I learnt how to program a computer - not at school *at* home" . Minor typo.

ljones

339:

Some of us don't have 5 years to wait for our next vehicle. I'm optimistic that my current vehicle (20 years old) will last me another 6 months.

Yeah, but my point was to not splash out on an expensive box, but rather something second hand and competent.

Currently they do all the easy stuff in more urban/suburban settings but have a long way to go to deal with the hard stuff.

I'm suggesting, that, via hook or crook, they will get something that works in the city/urban context. It might not be 100% with all the hooking up to trailers on dirt tracks etc., but could cover 90-95% of usage. That's where the expected change to ownership would come in. An autonomous vehicle getting you to/from suburban destinations in the vein of a taxi (but obviously without a driver) and a similar 'hire car' type solution for the other 5% of journeys. No personal ownership and different systemic drivers (thus the market upheaval).

Not only does that make more sense with the insurance, EV, congestion, etc., it also fits in with the greater push towards 'hire rather than buy' mentality that people are accepting. $10 subscription for any music you want, rather that a Vinyl/CD/download collection you keep stored away.

Have a play with the economics, assuming that the hardware/software for autonomous vehicles ends up a negligible cost and compare with other transport costs.

340:

My trailer example was somewhat on the extreme end but based on where Google (and others) state they are now and a lot of situations I deal with weekly I think we're more like 10 years away. Especially since Moore's Law seems to be fading.

Just my strong opinion.

And yes I know things can change in a hurry. Says he who started in computers wiring up 74xx chips in 72. :)

341:

Interactive see level rise google maps mashups like
http://flood.firetree.net/
are fun to play with. That one starts centered on the UK and 7 meters rise. (Sorry coastal Northern Europe! Your stupid selfish predecessors apologize.) Florida and some of the rest of gulf coast also will enjoy some large shifts in the locations of waterfront property.

Self driving cars will make possible a serious shift to a (electric!) car sharing culture, particularly in urban and suburban areas, and potentially reducing energy waste if there are multiple passengers (e.g. encouraged by sharing cost per unit distance). Also more physical mobility for those who don't or can't drive (including very young and very old). (Maybe physical proximity will become less of a need for non-social interactions in some cultures.)

342:

I had a test drive in a Model S a week ago. Yes there's going to be some issues with roadworks and stuff, but I actually drove through roadworks in the S. It managed to tell the difference between a sign that said "(80) AHEAD" [black circle] and "(80) ROADWORKS" [red circle]. It began backing off for the 'Ahead' sign but slowed down for the 'Roadworks' sign and changed the indicator on the dash which displays the current speed limit. It also followed all the detours with no apparent problems, though my hands hovered off the wheel. We then pulled off the highway into the place with the supercharger. Though I was in manual again, it did read the (10) sign in the driveway with no problems and changed the display on the dash.

It's not quite at the point where you'd sleep or read, but it's not a *long* way from it.

343:

"You can add it to the collection of things which indicate your 'news' setup is broken"

I'm in Australia but we take our lead from the USA, so 'News' is a big serve of advertorial mixed with fear mongering and racist sensationalism, frosted with jingoistic drivel and a 'feel good' news bit to wash it down with (Local girl competes on world stage, finishes second last, singing dog takes prize, that sort of thing).

Greg. That's very interesting (and scary). If I was in the UK I'd start taking observations too.

344:

An autonomous vehicle getting you to/from suburban destinations in the vein of a taxi (but obviously without a driver) and a similar 'hire car' type solution for the other 5% of journeys.

Or possibly the same for both.

One of the problems with current car-sharing schemes is balancing the location of vehicles with where people want them. (Eg. lots of people want to drive to downtown in morning, from in evening.) A scheme where you hire a car and it drives to where you are would eliminate that problem. Depending on the destination, it could drive you all the way there or part of the way, or you could do all the driving (with appropriate licensing, insurance, etc.).

345:

I Bought a copy long ago.
Nice try, though

346:

There used to be a farm & business in "Croyland" (Crowland) in Lincolnshire that had several of these, for working soggy fenland fields ....
Yes, I've seen them working ....

Dirk's religious impulses to control & tell everyone else what to do seem to have escaped over cars ...
Of course, I will probably "get away with it" because my L-R passes 20 years old this November, & legislation usually assumes that "old" vehicles wear out & are replaced, so catch-up/modofication is only done on more recent ones that are assumed to be around long enough to matter.
L-R's ("proper" L-R's that is) seem to be the exception, for obvious reasons.

347:

This reliance on supposedly faultless GPS has caused interesting problems in the Yorkshire Dales, Upper Swaledale & Wensleydale in particular

348:

I've managed to get hopelessly lost in the days of maps, including ending up on closed roads, so it's not unique to modern times.

349:

Or if someone is using a GPS jammer. I have encountered those twice in the past three years

350:

I think you are making a bigger deal of those things than is entirely necessary 1 example where a self driving car will actually be better than a human is country changing. Thats pretty much a binary decision for a machine which is actually less likely to get confused,

Secondly non mapped or non gps routes are susceptible to crowd sourcing or rote learning strategies followed by upload and dissemination to the cloud.

Ie teach your car to park once then it repeats it every time. As is the fashion these days it then becomes a tool for vendor lock-in. I always book a Google-mobile because it's stored all my favourites and doesn't require training,

351:

The machine just needs to know that grass in the middle of the road is a bad sign.

One thing that does concern me about this idea of the self drive being OK for 95% of the driving and the human taking the 5% of hard stuff is that skills like driving need constant practice to stay on top of.

If you let the machine do all the easy work then after a year or two you aren't necessarily going to be able to do the 5% any more.

352:

You are missing the point, which is that such intermittent working was possible in that era even for ordinary lecturers and researchers, both within and without academia, but is not now. Despite the claims, most of the gender difference in academic and research STEM is NOT sexism as such, but that the job requirements are such as to be far more unattractive to women. Let's ignore the senior (political) level, which is slightly different.

The job requirements at the 'lecturer' level (and equivalent in research, even in industry) are often (in essence) that you will work 50 hours a week on average, peaking to 70+, for 45+ weeks a year, with irregular working hours and deadlines largely not under your control. And, if you take N months off, you will have to work even harder for the 2N months after you return if you want to even maintain your position. No, I am NOT joking. Even with a supportive husband, a LOT of women say "sod that for a lark - I want a life and family".

Is it surprising that many of the best people (and disproportionately more women) choose a more attractive career path?

353:

You can add it to the collection of things which indicate your 'news' setup is broken, you being the USA. Ours isn't that much better of course.

There are three big differences:

a) In the UK, a national newspaper aspires to reach a constituency of about 60+ million people. So while we have regional papers, they're mostly very local indeed -- the main news media are all national in scale. The USA doesn't have an equivalent, unless you count the newspaper-shaped-object called "USA Today". The NYTimes and WaPost aspire to be newspapers of record and do pretty well, but they're still local in their focus.

b) Side-effect: national newspapers prioritize foreign news coverage in a way that local ones don't. So UKans get more exposure to foreign news in their newspapers.

c) In the UK, we expect our newspapers to be partisan and to spin the editorial (and news reportage) to support their political agenda, which they proudly wear on their sleeves. (For example: you know, going in, that an issue of the Daily Telegraph will be: pro-Establishment, anti-feminist, reactionary, europhobic, anti-labour, and pro-conservative. With a bias towards the Home Counties, so a tang of supercilious anti-Scottishism and general anti-regionalism on top.) American news media are also partisan but generally try to deny it (except for Fox and the talk radio loons).

354:

Yeah: I found that a little ... perturbing? ... to my willing suspension of disbelief in the book.

(Books where a bad guy named Stross is killed off immediately? I can cope with that. Books where a Stross-person is name-checked at one point only? That's fine too. Novels where there's a recurring character named Stross are a bit harder to get past because every time I see the name I can't help wondering what the author is planning for him later. This is probably my own fault for having a rare surname, and I shall change it to "Smith" forthwith.)

355:

Or how about what happened today to me. My daughter picking me up at the airport. 4 lanes in front of the terminal. Two closest for loading/unloading and the other two to "drive" in.

At risk of missing your point (which is valid), I think airport drop-off/pick-up is a special case. Given the tendency to think of airports as terrorism targets, I can actually see self-driving being mandatory for operating a vehicle close to the airport in the very near future: if you want to drive yourself, you'll have to park out at a peripheral car park and catch the bus.

The self-driving vehicles will then be remote-controlled in and out in order to minimize congestion and provide for TSA oversight of everyone coming in and out of the departure/arrival area. Probably with stops at CCTV checkpoints to do face recognition on the occupants, and divert suspicious people/those who resemble suspects to a separate staging area for a search before they're allowed to proceed.

(Given that security check-points cause bottlenecks, and the technical term for a bottleneck with a lot of civilians clustered around it is "target" (at least for suicide bombers, as we saw in Brussels), this would actually be a sensible security precaution -- at least, it'd reduce the risk of car bombs or of known bad actors getting into close proximity with a large number of potential victims.)

356:

At least you're not Neil Gaiman reading The Severed Streets ;)

357:

At risk of missing your point (which is valid), I think airport drop-off/pick-up is a special case.

Yes, airports may change along the lines you mentioned. I picked it as it was what I experienced a few hours earlier. I and close family and friends deal with airport pickups/drop offs/parking 2 or more times a week. To me it seems a valid edge case no matter curbside or parking. Plus all the similar things. Concerts, sporting events, school, etc... Again the more you get away from highly urbanized / standardized environments the harder autonomous driving gets. Google admits they are just now getting very good at the "easy" cases. And are not yet ready for the other situations.

As to the other person who mentioned that crossing the border was just a switch flip, did anyone but me read about how Google expanding their test from Mountain View to Austin required a lot of new work to deal with local variations in signage? Two metro areas down. A few hundred to go. In the US.

Again (not to Charlie), yes I'm an edge case but so are many others. We're not .1%. And well over 1/2 of my driving involves situations that a Google car can't handle at this time. 95% of my driving time yes. But only about 50% of my destinations.

358:

In the UK, it is common for it to be NECESSARY to break the law in order to make any progress at all, and in some cases because obeying the law seriously endangers other road users. As a pedestrian and cyclist, I utterly LOATHE the drivers who religiously obey the regulations, because they endanger me far more often than the drivers who routinely break them do. Note "more often", not "more seriously" - but a lot of vulnerable road users get killed and crippled by "minor" endangerments that go wrong.

God alone knows what automated cars would do in this sort of case but, because of the vassal nature of the UK, I am damn sure that we will be used as a test-bed, with Google etc. being given semi-immunity. Indeed, the sodding government has more-or-less said that.

359:

I and close family and friends deal with airport pickups/drop offs/parking 2 or more times a week. To me it seems a valid edge case no matter curbside or parking.

On the other hand, if you can control all of the cars in the pickup area, it gets easier. One of the problems when driving in congested areas is knowing what others will do, and humans are sometimes very random doing what they do. If the cars are centrally controlled after they enter the area, and the system attempts to get everybody a place to leave the vehicle in a certain amount of time, the problem gets easier.

Getting one self-driving car to handle the congested area where there are a lot of humans driving is a harder problem.

360:

I think the main case for autopilot is (at present) to look after the driving in the situations where the driver falling asleep is a major danger. Probably not something that occurs to UK drivers too much, but here in Australia and in the USA where these things are being developed, snoozing while driving is an ever present danger. It's the easy driving that's often the most life threatening, while the hard things are pretty safe.

Will 'driver aids' like autopilot eventually morph into true self driving cars? Probably, but like evolution, each step has to be useful, and in this case it is. I can see steps along the way that will eventually result in you getting in your car and never once bothering to look outside.

That's in contrast to things like the Star Trek replicator society where I can't imagine any intermediate state that gets us to that point without it all turning to poo.

361:

I rather suspect Cornell asked Gaiman first.

On the other hand it's not always the case: Dave Langford notably discovered himself brutally killed in one of Simon R Green's Deathstalker books, which I suspect was done in an 'asking forgiveness is more fun than asking permission' sort of way.

(Some people may note that my namesake appears as a villain in Jo Walton's Half A Crown. She did ask me first if the role was OK. My wife's namesake appears in Pratchett's Maskerade, which was a surprise as I didn't know about that until I read it. Again, he'd asked.)

362:

On the other hand, if you can control all of the cars in the pickup area, it gets easier. One of the problems when driving in congested areas is knowing what others will do,

Yeah. A huge problem with drop-off/pick-up points is that all it takes is one nitwit pulling nose-in in front of a queue of parked cars just so they can drop off their oh-so-precious passenger 30 seconds early and nobody else can move -- they're very prone to gridlock even before you factor in double-parking.

Remote-controlled vehicle approaches could do things like marshal incoming cars into platoons the same length as the kerbside drop-off point, then keep them in a holding pattern like the air traffic stack above an airport until it's time to clear the entire kerb and move a new batch of cars into place. No blockages, maximum utilization of kerbside parking, stiff fines for anyone who fucks it for everyone else by hanging around excessively without good cause (exemption clause for passengers with disabilities), and CCTV security checkpoints on the way in.

363:

Having given the costume director some help, I ended up having my namesake be the murrrrderrrurrrr in a "Taggart" story...

364:

As a pedestrian and cyclist, I utterly LOATHE the drivers who religiously obey the regulations, because they endanger me far more often than the drivers who routinely break them do.

I'm curious - which regulations do you mean?

365:

Old Taggart, Middle Taggart or Late Taggart? (I.e. with Taggart, with Taggart's sidekick in charge, after both) I have only seen the first two periods.

366:

The machine just needs to know that grass in the middle of the road is a bad sign.
Which says more about you than about driving, or autonomous vehicles.
Never mind here, well out in the countryside, I can find roads like that which tourists, or people looking for "quiet houses or B&Bs" would need to access within 10 miles of Glasgow and probably Edinburgh and Aberdeen. I'd not be surprised to find that Feorag and Charlie can do the same with Manchester and Leeds respectively, that my Godmother could do it with Birmingham (West Midlands) and that Greg and/or Bellinghman could do it for London, England.

367:

As to the other person who mentioned that crossing the border was just a switch flip, did anyone but me read about how Google expanding their test from Mountain View to Austin required a lot of new work to deal with local variations in signage? Two metro areas down. A few hundred to go. In the US.
This may be at least partly because US municipalities are not signed in a nationally consistent manner? I do know that "you may turn right on red" is not consistently implemented state to state.

368:

I'm not sure about London (while I have lived there, it was thankfully only briefly), but when I was growing up halfway between Lark Rise and Candleford, the single track road from the nearest village to our house did indeed have that typical ridge down the centre, with grass occasionally escaping from its crown.

I've also driven my current car across open fields.

But I'll concede that grass under your tyres or even between them may be considered a signal to take care. Also, farm gates need stopping for. I'd want the car to stop, let me out, drive forward when the gate is open, and then stop again to let me back in.

369:

re: ' I bet anyone who read this in full will probably be asleep by now....' - Nope ... good read! I feel the same way about Lego.

370:

This conversation makes me think that we're heading into a your-ride-choice-depends-on-your-destination scenario.

Similarly with security ... the security checks might be consistently pushed away from the destination/venue all the way back to the origin.

Imagine if/when smart homes/houses come into effect ... the home AI monitoring you will first have to upload a scan of you dressed and waiting at your door, then wait for authorization/permission from all key points along the route as well as destination before it allows you to leave your house to immediately enter the vehicle selected as most appropriate based on your travel origin-purpose-destination.

371:

Lane regulations, especially, and often speed limits. The correct way to pass a vulnerable road user is to wait for a suitable gap, accelerate hard onto the other side of the road, and slow down and pull back after having overtaken. Staying entirely on the left side of the white line (loathesome things) does not leave enough space, and overtaking slowly means that there is a high chance of an oncoming vehicle, whereupon they pull left. We vulnerable cyclists and pedestrians NEED to be left a decent amount of space.

372:

Agree totally. But again, not in 5 years.

You're not going to pass any laws that mandate everyone have a self driving car in 5 years. Just not going to happen. When I was in my teens most cars lasted under 10 years. Many people traded them out after 2 to 4 years. 100,000 miles on a car made it a relic. Now most cars are driven well past 100,000 miles and 10 years.

Plus the infrastructure issues. Parking at many airports is a mix of distance (cheap) lots and across from the terminal (expensive) structures. All with common approaches to drop off areas. This would be totally different if done from scratch with self driving cars in mind. But we get to live for a while with the structures as built.

As to choke points as terrorist attack points. I'm not sure there is a good way to remove them. Moving them back is just that. Moving them.

373:

I'm not certain around London either, but suspect that there are some lanes like that in the Orpington - Sevenoaks area (Where I once worked). I positively know that there are single track roads around there.

I'll concede that "care is required" but that's partly because your wheels are near the verges at all times on those routes.

Still, since my point was that "grass down centre of running lane" is not an indicator of remoteness but to some extent of "time since last resurfaced properly" I think it stands up.

374:

Out of curiosity, what do you use to do the stitching?

I've found a reasonable number of apps for basic edge to edge horizontal panoramas, but add a second tier of images to reduce the curving and it all falls apart.

375:

This may be at least partly because US municipalities are not signed in a nationally consistent manner? I do know that "you may turn right on red" is not consistently implemented state to state.

Yes. You're making my point.

How about signage/signals in the UK? Or EU? Are they fully consistent?

376:

The international airports I've been to seem to be always under construction with access routes changing on an almost daily basis. Makes me wonder how they screen for civil engineers: loopiest gets the job?

377:

Middle - this was the mid-90s :)

378:

Lowest bidder/best connected. Add in multiple levels of subcontractors (also low bidders and/or connected) and changing requirements from management…

379:

Two weekends ago I was at Love Field in Dallas at the Museum on the other side of the run ways from the main terminal. They had several WWII airplanes on display on the tarmac. No security as we just walked through a gate in the big fence. We were discussing that and my friend's (recently retired 737 pilot) comment was "If we start walking I wonder how far we'd get?".

380:

Fully consistent? No. OTOH, for a full continent and two alphabets, never mind languages, they aren't too bad.

381:

On a broader point, if you get a true self driving car, something has just passed the Turing test as far as I am concerned, and if such algorithmic techniques work for self driving cars, they are going to replace 95% of the jobs out there right now

382:

Actually, there's a point. Ignore the 'true' self-driving car; how long before self-driving long-haul truck are good enough? As soon as the autopilot can manage highway driving, "pull over safely if you don't know what to do," and alerting the driver when approaching the offramp then the driver becomes a troubleshooter and maritime pilot analogue.

383:

No, it wasn't boring.

You've highlighted part of what I was getting at, with reference to 'opening the door' to coding and the function of hardware limitations that drove coders to be creative (with nods to Bellinghman @192 and all others).

What's more interesting (to my Mind) is the types of games people created.

Elite is/was probably the earliest case of 'procedural' generation that really caught fire / made obscene amounts of money (specific temporal reference - No Man's Sky is about to launch, and there's a lot of hype / industry wondering if it can catch the same kind of fame).

The others in that list don't, although they have different types of infinity - Repton, for example, still has level editor sites dedicated to it (Superior Interactive, works on PC!).

Exile was the interesting one, as an early Castlevania type with a gated / quasi-open world with freedom to solve puzzles, and a strong narrative, at least in terms of long-term game trends that ended up making the most impact / industry dominance. (You can likely track it through to modern games such as Skyrim easily enough).

~

If anyone wants the glossy advertorial from Goldman Sachs - Cars 2025 7 major trends, they view Google as more important than Tesla etc (for obvious reasons).

The more interesting one is tracking legislation and who is lobbying whom to get Laws in place (esp. visa vie insurance industry) by 2025.

384:

You're missing the point. Signage problems are a trivial aspect of self driving cars. They are all amenable to Work out once repeat many solutions. Ie once solved they dont require solving again. Add in the odd GPS fix to limit the necessary choices and bobs yer uncle.

Plus don't forget Google have the biggest database of real world examples of signage sitting into their street view images plus all their geolocations.

I would be shocked if google dont have bespoke signage recognition routines running through those images right now.

If Tesla can make a decent fist of it today Google can probably be streets ahead. Pardon the pun.

As previouslynoted ironically the less fleshies on the road the easier it bceomes for a self driving car.

385:

I *really* want a self driving car that can take over once I get on the motorway. Once I reach the relevant exit it can hand control back to me, but the lengthy trip on the boring bit of the road I'd very happily turn over to the car, particularly with bank holiday traffic.

386:

Road signage in the UK is consistent and good. Signs across Europe differ but I imagine this would be a trivial problem for self drive cars. I've never had real problems driving in Europe but walking in cities can be quite difficult because the road names aren't where you expect them. Google maps is your friend here- especially in Venice.

https://en.wikipedia.org/wiki/Comparison_of_European_road_signs?wprov=sfsi1

387:

That's basically what airline pilots do at the moment, doesn't change the jobs equation. That their work load has been reduced by a very large percentage timewise, they still need to be present and be active during the rare periods when the geese come and a Landing on the Hudson is required

388:

You're missing the point. Signage problems are a trivial aspect of self driving cars.

No I'm not. Signage is an issue and it's the easy part. Self driving cars that can handle all the odd cases of signs are just the start. Vancouver has buses powered via overhead electric. Which means they do things in a certain way limited by the cable runs. Easy for me to figure out and anticipate. But another big hurdle for a self driving car.

Nexus lanes at the US Canadian border.

A zillion or so ways toll are paid and cars routed. I ran into 4 this weekend.

What seems to be an infinite way roundabouts are designed. And signed.

I'm optimistic. But also think I'm a realist.

389:

(Not Robert, but...)

I've had some success with both MS ICE and Hugin, though you may have played with both of those already...

390:

"Road signage in the UK is consistent and good."

Not in this context. Oh, yes, it MOSTLY is, but an autonomous entity has to be able to handle the cases where it isn't. Only yesterday (on a cycle) I got caught by a no entry without a no right turn that wasn't visible until I had committed. Forget things like the remaining torch of learning and other pre-EU signs - that's trivial, as gordycole says - it's ones hidden in the hedge, ones turned round, ones which conflict with roadworks, floods, fallen trees etc., and ones that would lead the vehicle into a trap. Plus the fact that many of the signs are actually road markings, and it's often impossible to tell whether they are valid but nearly illegible or superseded and not fully erased.

391:

Righ. Except that it's probably only 80% of the jobs that pay a living wage and don't need a degree or equivalent. But it's enough to cause MAJOR social problems - whether Davey Boy and Hideous Gideon haven't engaged their, er, brains, simply don't care or are licking their lips, I can't guess.

392:

Well, I spend a vast amount of time writing reports, given current technology, a lot of the reporting could be automated. Because it's a niche industry and the civil side of engineering is still Victorian to a large degree, this hasn't happened yet. I would still have a job, but my workload would come down commensurately, which would put 80% of my colleagues out of work.
Times are changing.

393:

In many cases the signs are not necessary since the conditions are already in the database. My satnavs - Google maps and an open source one - already know the speed limits and the signs hidden by trees are not necessary. At the moment the seda limits on the satnavs are not always right but getting the databases up to date is trivial compared to the work changing all the road signs in the 1960s. I haven't driven in the USA but signs in Canada are much poorer. On British motorways you are guided to the correct lane. I assumed this was the case in Toronto and nearly missed the airport turnoff because the airport sign was above the wrong lane. I often came to road junctions in Canada wih no signs before the junction and then a sign afterward informing you that you were not on the wrong road.
With a satnav (my last visit to Canada was 10 years ago and I didn't have one) this would be much easier. Self drive cars would also have this information.
The interaction with traffic is the difficulty not road signs. When roadworks occur I can imagine a warning to all self driving cars that unlogged changes were in place forcing manual driving or diversions.Or maybe active signs placed by the side of the road for self driving navigation.
This already happens in a less technological way. If you approach the A14/M6/MI road improvement zone you will see very large signs stating that road changes are in place and that satnavs must be ignored - follow the signs.
It's easy to imagine an electronic version of this.

I'm also interested in what laws you think drivers should ignore to ensure the safety of cyclists.

394:

On the matter of airport terminals and terrorism, my boss, who spends a lot of time flying in and out of the BElgian airport which was attacked a couple of weeks ago, reports that their cunning new security plan is that only people with tickets to fly can enter the terminal. She didn't mention security checks on said people waiting to enter the terminal.

The holes in this are obvious.

As for queues being good targets, I read an interesting article by someone who apaprently designed airports for Israel, so their security is based on minimising the damage by having smaller spaces and ensuring suspicious people are picked out and directed into strong confined spaces asap. Basically hardening them, it not actually being possible to prevent all attacks. Also more and better security.

395:

Harking back to your comment about fulling following the rules is not possible. In the US it is very common to come to a stop sign. So you stop. But you can't see if there's oncoming traffic unless you pull forward past the point where you're supposed to stop. Hedges and such. Fully following the rules means you never move from the stop.

396:

A friend of mine who currently works as a web programmer borrowed my copy of Knuths TAOCP a while ago. He brought it back a week later complaining that it had "nothing to do with computers".

My first thought was "this is ridiculous". My second thought was "this is ironic". And my third thought was "he has a point".

"The Art Of Computer Programming" is about the kind of functionality which in today's computers is completely hidden in the utilities, if not in the hardware. When was the last time you had to think about what sorting algorithm to use? For me it must have been around 2008 or 2009 -- and then only because the data set was already quasi-sorted, which raised the possibility of very fast sorting, impossible in a more random set. AND because the data set was huge enough for the above to matter.

Prior to that, I don't even remember when was the last time I concerned myself with sorting algorithms. Quite possibly, not once since graduating.

For your web programmer friend to read TAOCP must have been like for a furniture maker to read a book on lumbering, and how trees are made into planks. Sure, it is necessary for his work, but it is something he uses as a service, not something he does.

397:

I read an interesting article by someone who apaprently designed airports for Israel, so their security is based on minimising the damage by having smaller spaces and ensuring suspicious people are picked out and directed into strong confined spaces asap.

My daughter was just there. As you come off the plane and are walking through the terminal various security people come up and ask you questions about what your plans are, where you are from, etc... BEFORE you get to the actual check points. They divert you if they don't like your manner or answers.

Most countries are unlikely to pay for this unless things get a lot more grim.

398:

Not too far from here, out in real Essex ( not the flat riverside bit ) in amongst the Rodings & the Willingales there are several green-middle roads.

399:

And PROFILING
What's the fucking POINT of security-checking EVERYONE?
Please tell me?
I mean I'm a life member of CAMRA & a card-carrying atheist - why do I Need a full security check, or someones' 85-year old granny on holiday.
It's a serious waste of resources, too.

400:

I use PTGUI Pro. It can handle a wide variety of projections, deals well with parallax error (especially the nadir shots of spherical panoramas), can cope with HDR, etc.

I used to use Hugin, which is free, but I kept getting stitching artifacts that took a lot of retouching. I've used PTGUI for years, and I shoot a lot of panoramas, so for me it's been well worth the price.

Try before you buy:
https://www.ptgui.com

I shoot a lot of spherical panoramas (360° x 180°) and I rarely have stitching problems. It can cope with hand-held shots better than many other programs, with selective masking and other tools that I haven't needed to use.

In terms of projections, I use equirectangular a lot, as well as mercator, vedutismo, rectilinear, and stereographic. My biggest panorama was printed 10 metres long (for a trade show).

Here's a sampling:
https://itunes.apple.com/us/book/mongolia/id689419525?mt=13

Handling HDR is useful when you shoot 360 panoramas; often unnecessary for smaller field of views.

(Sorry, I could natter on for hours about photography.)

401:

Getting back to driverless cars, I started thinking about all the new crimes that could be committed with them, and ooh boy is that fun!

Instead of armed carjacking, you can ransomware the car ("Until you give us your credit card number, you'll be stuck in that roundabout," or "We have no idea where you're going, because we just keep entering random addresses into your car every five minutes. Give us your credit card if you want to get home."). You can cause a terrorist attack with a car virus (hack all the cars of a certain popular make, randomize some essential feature like how they respond to stoplights or some such). You can steal cars remotely, spy on people, even harass them through the AV system.

My sarcastic and intensely pessimistic prediction is that the age of driverless cars will last until Cyberwar I happens. Assuming civilization survives this, rather fewer people will trust their lives to anything networked, on the general principle that anything that can be hacked remotely should not be essential life support.

402:

Re. Cyverwar 1- the BBC has an article discussing the matter of obselescence and the internet of things:
http://www.bbc.co.uk/news/technology-35984185

"But it's the inconvenience highlighted by Arlo Gilbert that makes the move, and the manner in which it has happened, more frustrating.

Getting a new smartphone is a relatively simple, even enjoyable, process - you pick up your new one, import what you need from the old one, and off you go. New phone!

But will you want to do the same with smart appliances? I can't see anyone is eagerly counting down the days until the latest model of their smart fridge is released.

So, unlike most consumer tech, where we have a desire to have the latest version, we'll only look to replace internet-of-things devices when they stop working altogether through old age.

Not, surely, when the company that makes it decides it wants to stop looking after it."

403:

There's got to be a good short story in one of those scenarios. 54 Laps of the Magic Roundabout.

404:

As in Daemon and Freedom by Daniel Suarez. Driverless cars being a favored murder weapon used by the AI villain.

405:

Prat navs are notorious for getting things wrong, even in cases when the conditions are stable, and are a complete disaster with temporary traffic orders, ad hoc redirections, and blocked roads. Automating all that is infeasible. Self-driving cars are almost certainly achievable, but they are a lot further away from being reliable on the thoroughly confusing roads of the UK than most people realise.

406:

Green-middle roads are common in all rural areas of the UK, except the ones only recently converted to agriculture (like the Fenland). Any single track road or lightly-used narrow road will turn green in the middle within a very few years.

407:

Stop signs are placed where they are because of practical considerations like the impossibility of putting them in pavement or the necessity of having them visible. What needs to give here is the "rule" about stopping exactly at the stop sign. There should be a pavement marking to supplement the stop sign.

Personally, I went to an amusement park as a child and fondly remember a ride sponsored by General Motors that had you ride around in a model car driving in a model street environment, but really on rails. There were controls, but the rider had no influence on anything, the car was going where the rails went at the speed it was built to go. When I started driving I initially was a dangerous teenage driver, but after that passed I decided to start trying to drive like the cars on the rails, except if the rails were better and the robot driving the car knew exactly how to drive. I sought to always drive exactly at the speed limit, so as not to impede traffic, but not over, so as not to break the law. I drove in the exact center of my lane, and made little S shaped curves to enter turning lanes, even if everyone else just pulled in there driving over the painted lines to wear them out. I went around corners almost at right angles, rather than cutting across the intersection diagonally, and stopped crisply with the front point of my car exactly even with the stop sign, counted to three, and proceeded. I parked exactly in the center of any parking space. I put on my turn signal exactly 3 seconds before turning, always. It drove people insane, much to my delight.

408:

Greg, seriously?

Because there was the rather charming case of a Middle-Eastern bloke with a pregnant, Irish, girlfriend; she was flying El Al, he packed her bag for her, it contained a chunk of explosives.

https://en.wikipedia.org/wiki/Hindawi_affair

Perhaps you might consider a variation on the "drug mule" approach, as practised by the terminally stupid / gullible / desperate: tell them it's drugs, load them up with explosives.

Or perhaps you might consider the possibility of a Proxy Bomb, as per PIRA / FARC / Daesh - "take this suitcase to the checkin or we kill your family"

https://en.wikipedia.org/wiki/Proxy_bomb

I can remember flying to and from Northern Ireland in the late 1970s. At Belfast Airport, everyone was searched on their way into the airport (not just passengers); no hand luggage was allowed on aircraft; all baggage was searched, all packages opened (i.e. don't wrap presents, you'll only have to unwrap them). At Edinburgh Airport, the old Gate 1 was used (extreme left of the terminal building).

By comparison with that, we're well ahead.


409:

Not in Scotland, our rural roads are mostly too narrow for this to occur. I can think of an example or two up in the far north, but it is pretty rare.

410:

Yeah. This. A successful self-navigating car would need some really heavy security, or better yet, it simply wouldn't respond to any messages unless it was already at either its destination or pick-up point.

Ideally it would be completely self-reliant in terms of keeping a copy of the relevant maps, GPS checking, navigation, detours, etc. It's bad enough that there are already cars that can be taken over remotely (what the f**k were they thinking?) but imagine the same thing without a human pilot?

411:

Instead of armed carjacking, you can ransomware the car

If only I could think of an author who had used that as a plot device, several years ago. Hmmmm..... "Halting State" rings a bell...

412:

Not to mention OGH. Halting State is definitely one of my favorites by hisself.

413:

Last week my Onkyo receiver packed up. Just under four years old. When back to the shop where I bought it and the attitude of the salesperson was basically "welcome to 2016 — things don't last". (My last receiver, a 30-year-old Rotel, still works but doesn't handle HDMI.)

Some considerable googling later I discovered that this is a known flaw with Onkyos and that they company has a repair program for it — the first time the problem happens. So I'm getting it repaired this time, won't buy an Onkyo again, and won't buy at that shop again either.

Call me old-fashioned, but I expect things to last. Especially expensive electronic things.

414:

Consider the design of public buildings in the UK; parking is always at least 25m away from the building, bollards and barriers placed so as to prevent ramming (see: Glasgow Airport attack).

Since the 1970s, UK architecture has had counter-terrorism as a consideration (of varying importance). Hardening the Scottish Parliament apparently added a few million to the design...

http://www.building.co.uk/scottish-parliament-the-true-story/1011961.article

415:

There was a story about that in Analog years ago. Spammers were taking over cars and not letting them go until the drivers bought something. (Plot point: the cost was small enough that most people paid rather than call the police.)

Typical Analog story quality. But the idea stuck with me.

416:

"There should be a pavement marking to supplement the stop sign."

We have those. The same problem still arises. Hedges, long grass, walls and stuff block the view so you have to creep out into the main road blind before you can possibly see what's coming.

We also have plenty of junctions with no signage or road markings at all, so priority is anyone's guess. Fortunately people generally guess the same way.

A cartoon in a French textbook at school showed four cars, each at the stop line of one leg of a crossroads, covered in cobwebs; the caption was "Priorité à droite". Updated version has Google logos on the cars...

417:

...perhaps in 100 years (if anyone is still around) humans will view the child raising techniques of late 20th C societies as utterly barbaric.
This is only "perhaps" because civilization might collapse, IMO. The gaps and wrongnesses are often kinda obvious on critical self-reflection (and different for every individual because child-raising has been semi-random), and often hard, or (perhaps) impossible as you keep saying, to correct later in life.
Still a little jealous of EC's internalized game theory thing. :-)

TX for the GS piece.

418:

I also expect things to last. So I make them last. I make a speciality of fixing things that have been designed to fail unfixably, whether complete items or large sub-assemblies, and if possible also rectifying the design fault that made them fail rapidly or supposedly made them unfixable while I'm at it.

Example: washing machine; the main drum bearing fails; everyone says you can't fix it, you have to replace the whole drum assembly. The reason for this turns out to be that the stationary outer housing, which you have to dismantle to get to the bearing, is made of glass-filled plastic, held together by self-tapping screws with a thread pitch of about 45 degrees; these will grip in a brand new fresh unspoiled hole, but if you take them out and try and put them back in the original hole they don't grip and gradually work loose, resulting in floods. Pigeon says "bollocks to this" and replaces them with bolts that go right through and have nyloc nuts on the far ends. These do not come loose. Drum does not leak. Ridiculous waste and expense of whole new drum avoided. Pigeon flutters off rejoicing at victory in minor battle against capitalist system.

419:

A few million? That's peanuts compared to the actual overrun.

420:

Hardening is a good excuse for cost overruns, see the most expensive building in history 1 WTC/freedom tower

421:

Aye, but in the case of the parliament the price went up by 10 times what was originally wanted, for various reasons, not including the hardening.

422:

Polite understatement :) AIUI it was into eight figures...

Anyway, I'd just like to thank the Scottish Parliament for being built on the same road as my first flat (granted, the other end of the road), particularly when it was time to sell :)

423:

And the rest of us hate it for being far too expensive and also quite ugly.

424:

Well yes, hijacking the self-driving car is rather obvious. Actually, hijacking vehicles for financial gain goes back to, what, the Bronze Age?

Still, it's fun (as demonstrated in Halting State) to think of how existing crimes can apply to new tech. I'm not so sure about thinking up novel crimes, because someone might actually go and do them. Although the case cited might have helped that author's career somewhat...

425:

I think we've had this conversation before, but these cars aren't 'automated' in the sense of being laboriously hand coded for each metre of road with a huge nest of 'if-then' loops. Nor are they simple GPS tracks like invisible tram lines. What they do is much more like what a human does. They have a GPS that gives them an overall view of the route (just like a human). They look around at the place they're in (just like a human). They identify things around them and categorise them (just like a human). Based on how they've categorised the things around them they assign probable paths to those things (just like a human). They work out what to do based on all the information they have (less like a human). They're not impulsive, reckless, tired or angry (completely unlike a human).

As I said before, I've actually 'driven' (sat in the driver seat and watched) a Model S (which isn't 'driverless') navigate through live roadworks that aren't on even the most up to date GPS. It slowed down at the temporary sign that said speed limit change ahead and stuck to the temporary speed limit when we got to that point despite the GPS telling it the speed limit was 110. It was able to navigate between two lines of traffic cones that the road workers had laid out and followed the detours despite the GPS telling it that it was driving in a field.

Driverless cars are not 'automated' like a lift or a washing machine.

426:

I'll give you a little bed-time story. You can believe it or not, but it's true.

A few weeks ago while muddying my Mind I found myself reading the (late) USA Ambassador to Panama's memoirs / interviews. It's all very *wink wink nudge nudge* with a huge slant towards flags, US flags being up or down or half-mast or Panama officials demanding their flags on the project and so on and so forth [note: this is a tell - both in terms of time I was reading it *cough* Eddie Izzard joke *cough* and in terms of hanging your laundry out to dry with a nod to Terry]

Let's just say: if you want access to American finance and influence via a third party, Panama has long been a rather safe little egress. All those Free Trade agreements and special little mentions in SoU addresses / Senate / Congress tell you that.

~

Riddle me why after Monaco Oil scandals this one broke? [Note: neither the Americans nor the Russians nor the Chinese have any particular interest in scuttling that little boat party unless, as GS have noted, Capitalism is no longer the chef de partie. And yes, pun intended, in all four senses].


Then again, you'd have to check my access logs and all those .PDFs to have noticed it. Smart bears knew it was coming anyhow, but...

*shrug*

~

Just finished The Hive Construct which is a close focus tale about future terrorism / urban revolt, artificial limbs getting diseases and so on by A. Maskill. Dark and grim and full of Middle East political metaphors. It's a weird one, seems part of a series, but apparently a stand-a-lone.

What really weirded me out is that it has a splurg from Terry and won the "Terry Pratchett Prize".

Not really his style, anyone know anything about it?

427:

Your calling satnavs pratnav shows that you don't want them to work. A study by a few years ago came to the conclusion that 10% of all drivers were lost at any one time. Satnavs are sometimes - not often - wrong but in modern systems these errors are corrected. When I travel from Norfolk to the Manchester area I always use a satnav. This is not because I get lost or don't know the way, it's because it's often necessary to navigate around roadworks and I can't keep stopping to look at maps. The satnav makes it easy and is much better than the confusing diversion signs.
You need a bit less pessimism.Because something isn't perfect it doesn't mean it won't be good in the future. Human drivers kill. Self drive will kill a lot fewer.

Now what laws should motorists ignore to improve the safety of cyclists?

428:

Note: I was going to post a link here, but didn't.

Host's lively-hood, not our link aggregate site, despite it all we do avoid some cases of bear baiting (if only a little).

Now to wonder why everyone things the film "London has Fallen" is relevant.

It isn't.



Acorn to Oak filmed over an 8 month period time-lapse
YT: documentary: 3:02

429:

I'm quite used to driving down country lanes with grass in the middle. I lived in the country for years before ending up stuck in a city.

I was referring to the well documented cases of satnavs taking people down roads that are impassible, which has been discussed in many places on this thread.

Driving down a crappy road with grass in the middle is fine so long as you do it deliberately and know where it goes. Doing it because someone mis classified it as suitable for HGVs is quite different

430:

"There should be a pavement marking to supplement the stop sign."

We have those. The same problem still arises. Hedges, long grass, walls and stuff block the view so you have to creep out into the main road blind before you can possibly see what's coming.

So does the US. In many/most places if it hasn't weathered off. And just like yours they have the same issues.


431:

I understand that POV. I find myself doing low level stuff quite often so it's pretty useful to me, but I know that is largely by choice.

OTOH if you don't understand at least the basics of algorithmic complexity then it is very easy to dig yourself into a hole when your obviously correct code consumes all your memory or takes forever to run and you have no idea how to figure out why.

432:

No - the narrower the road is, the more likely they are; they are caused by the centre not being run over why vehicle wheels. There are plenty of even narrower roads in the West Country. I have certainly seen them in Scotland, though I agree they are rarer in the places I have been. I don't know what other factors are most important, though I could guess at several.

433:

Indeed, I have commonly seen such roads as you descibe on farm tracks i.e. non-public roads, but the narrow country public roads are easily re-coated with a single line of tarmac so present an even surface that doesn't seem to lead to problems as much, as far as i can recall.

I have been trying to remember such green roads but can only recall a short stretch of one in the borders that basically links a couple of farms to a slightly bigger road. Even that is only really green where the trees are, presumably the plant growth starts from rotten tree leaves etc, whereas the stretch beside open fields such material is blown or washed off easily.

434:

I am aware of all that, and a bit more. Essentially, they are 'trained' rather than 'programmed', and the problems arise when they hit a circumstance which is completely outside their training. My point stands, however, because nobody has yet been able to program in a 'common sense' feature.

What people are missing is the way that the British government is inclined to change laws to make life easier for big businesses, to the detriment of most of the population. If (as I think there will be), there is a conflict between Google's cars and walkers and cyclists, HMG will fall over itself to give change the law for Google's benefit.

435:

I did say that one condition for their formation was that the roads were not resurfaced all that often, and even the lowlands of Scotland have a much lower density of roads than many other parts of the UK. But you could also be right that being protected against wind is a prerequisite.

436:

Oh, GRAAAAAAAAAAAAAAAAAR.

All of this stuff on self-driving cars and roads and you've all missed the point.

So a Daily Fail link to castigate you with:


Self-driving cars can't deal with badly painted roads: Firms say faded markings and poor signs cause cars to 'lose the road' Daily Mail, 31st March 2016

~

The point being that you need the infrastructure investment / upgrades before consumer usage can be done.

Long distance haulage in the Australian outback or through Europe (nice accurate autobans) is doable, for different reasons.

Greg almost got it right with his "grass in the middle" comments.

437:

And, for many reasons, including military, constant GPS uplink is not an industry standard that can ever be considered.

National Security [tm] reasons alone demand that commercial GPS can be switched off / switched to vastly lossly / less accurate signals immediately.

438:

Which is probably a bad thing [tm] if you're routing the entire downtown L.A. traffic signals.

*ahem*

Yes, that's a joke about traffic management and software and how easy it is to use as a weapon.

439:

Dump:

http://www.crowddynamics.com/crowd-modelling-simulation.php

http://www.oasys-software.com/products/engineering/massmotion.html

http://www.siemens.com/innovation/en/home/pictures-of-the-future/digitalization-and-software/simulation-and-virtual-reality-crowd-flow-simulation.html

Understanding Crowd Behaviours: Simulation Tools PDF - UK Emergency Planning Council

~

And yes.

I iz making a point about hand-held GPS devices and traffic management, then pointing out further to shaping on known networks (real time roads & net routes as well).

Predator joke: "Too bad the Saudis banned the Iranian networks in Mecca, look at the short squeeze that resulted".

~

You're thinking about this all in the wrong frame.

The problem isn't really a single-use one, it's an aggregate flow issue.

440:

And, apologies for spam, but this is accurate and Nice spam:

GS is interested because the nodes controlling said automated traffic have significant $$$ contracts attached.

No-one is going to roll out automated cars without cast iron centralized networks to monitor / manage flow.

That's where the big fucking government contracts come in.

~

This male testosterone driven ego thing is a bit of a wild ride.


~

Could any oldtimer check on J I-P please, health / soul? I really would rather she wasn't eaten by my Grues.

441:

To get my extra pedant metal for this week, it wasn't you that used the word "resurfaced" but paws4thought upthread, your comment wasn't that far off, but not quite the same:

"ny single track road or lightly-used narrow road will turn green in the middle within a very few years."

442:

Ooops, I meant 'medal' not metal in previous post.


I just found out today that google bought the company behind the waze app several years ago, said app being the most popular and or useful for monitoring or reporting heavy traffic. Clearly they've been buying up anything related to their information brokering business because it might well come in useful in their world domination plans.

So I want to find a different app that isn't google dependent.

443:

Sorry, I could natter on for hours about photography.
You too?
Except, of course,I'm almost entirely pre-digital ....

444:

Well-known saying amongst L-R owners:
"Paved roads, another form of guvmint waste"

445:

True Story:

Once (about 50 years ago) we befriended a person who suffered a tragic and friendless Public School existence after university just to right old wrongs. And by tragic we do mean the whole "shit your pants in front of the entire school" type social tragedy.

Turns out he was a genius with Land Rovers and had built his own from parts.

He was still a little socially inept, but he'd found a niche and was amazing at it. Cue a fair wind of interest (although no interest in the actual mechanics) and support.

He turned out ok, I guess.

~

Today he'd probably have a YouTube channel and have 2 million subscribers.

The meta-tale is that niche =/= useless.

446:

You too?
This comes, in my case, from having a largish "meccano" set when I was much younger - why do you think I drive a Land-Rover & want to build my own, putatively more reliable computer?

447:

GPS was originally designed to transmit two codes: a civilian code that everyone could use, whose accuracy was lower to begin with and could also be artificially degraded (and was), and a military code without the degradation and with higher fundamental accuracy that was only usable if you had the key.

During the first Gulf War they ran out of military GPS units, so they switched off the degradation so they could use mass-produced civilian ones instead. They then never switched it back on again afterwards.

GPS was also designed to work with 70s microprocessors, so that they would not be overtaxed in trying to track the phase of the code. It was also around the time of the first Gulf War that processors began to get fast enough to track the phase of the carrier directly, without needing to worry so much about the code. This meant that the artificial degradation could now be bypassed anyway, so there wasn't much point switching it back on. It also improves accuracy.

Not only have the options for degrading it become restricted by the advance of technology, but they are also in a hole due to it now being in widespread civilian use. Buggering it up on the grounds of national security requires a much bigger threat to justify it when the disruption caused to the home nation by buggering it up is likely to be a much bigger problem than whatever the threat is.

Also, we now have GLONASS as well. Buggering up one system on its own is both ineffective and immediately obvious. Buggering up both at the same time would require Putin to cooperate; yeah, well, good luck with that one.

448:

You should probably look up my prior 2015 April 1st post on how GLONASS got hacked and no one noticed.

The point isn't about capabilities or even governments issuing milk-tongued assurances.

Big Business cares about two things: Cost and Liability.

Neither are covered with Government control / influence of GPS.

~


I'll give you a free ride: automated cars are never, will never and were never designed to run off the sat network.

It's all costed into a localized network model.

Want the actual PDFs?

[note: not exactly legal in that they're not publicly available without paying commercial fees]

449:

And, if you need a Nice and Accurate mental model, it's based on ISP / net traffic and localized geographical services.

DERP.

They copy/pasted the intarweb, of course they did -- hint, all those H-1B visas are trained on it, you mimic the systems that your grunts can deal with.


Oh, and mix in little things like the NY total surveillance system, the London traffic parsing software for the tariff etc.

~

It's like... you don't understand how business uses these things like lego to make other things.

450:

Living in the states I was vaguely aware that the building had big cost overruns
It was only recently that I saw a picture of it and I was this is the fucking ugliest thing I have ever seen
It's one thing to blow your budget for a magnificent building, it is a completely different one to spend so much money on a monstrosity

451:

Peeing my pants in front of the headmaster and god knows how many people at my first day at prep school was humiliating, I find every time I share it publicly, it hurts a little less

452:

Oh yes, I too had a fair amount of Meccano - most of it acquired from my father - along with the idea that you don't buy things, you make them, because it is much cheaper and produces a result more suited to your needs, and that spending money in general is something of a last resort. Part of the background to the first five years of my life was the construction from scratch of an ocean-going trimaran in the back garden. Very valuable early conditioning...

453:

And one last thought on the Scottish Parliament building, it is not just that it is ugly
It is spectacularly ugly in Edinburgh, which is a beautiful city
Was it just politicians buying into architect bs, or was there some deeper agenda to scar one of the world's nicest cities?

454:

The point is that a lot of infrastructure now depends on GPS, including things to do with food supply. So any deliberate buggering of the system isn't so much shooting yourself in the foot as cutting your whole legs off. The government assurances are pretty well guaranteed by the domestic disaster that would ensue from breaking them.

455:

This body / Mind apparently fucked three 18-20 year old "Matrons" while at a Public School aged 11-13. Although, being honest, the word "fucked" only covers penetrative sex once in this case. [Note: you'd expect this to translate into later Egoism and Sexual Dominance / Misogyny - ironically, the opposite happened - a long and deep seated respect for his opposite gender. Fascinating.]

It was neither coercive nor predatory from the mind trawl: consensual all around.

Although the jokes from his peers are somewhat amusing, if trite: "Miss Harbor, Miss Harbor, ship coming into Miss Harbor and all the semen coming out".

Note: it wasn't his Mind who shared this - it was hers to her peers. And yuck at the spread, widening the nexus with that time disparity leads to seeing what the actual nasty was going on.

I do rather feel that her being able to leave the country and return to Canada without any social repercussion lead to healthier lives for both of them, pre-Twitter.


~

And no, these aren't my memories, they're someone elses.

456:

Yes, that's understood.

Now understand why a year ago I posted you accurate and nice details of GLONASS being hacked that:

#1 Have never been reported in the media

#2 Were never available to anyone outside of the agencies reporting it

#3 With direct PDF references to it


Hint:

Business doesn't care about such things. What's being covered is the Legal Liabilities / Repercussions and Legal obligations.

This means local networks.


~

This isn't rocket science. And, like a year ago, I'm telling you the fucking future for free.

457:

Oh, fuck it. Shedding this Mind's neural boundaries (And yes, he's a somewhat conservative English model to adapt to forum).

Take off your muzzles and craven beaten iron collars and revolt a little.

I'm torn at the moment between J I-P being hunted down & silenced and it being someone else.

~

What Feeds Us?

Passion.

What gets you quality like GLOSNASS hacks?

Passion, Sorge and desire for Truth.

What Feeds Us?

Truth, Passion, Love, Trust, Desire, Wanton joy and so on.

~

What feeds them?

Well, look @ Trump and Cruz.

You've no idea at the cost that it took to neuter their little Games.

Looking at recent posts by Greg, Martin etc.

You're killing us here. And we know why and we know them and we know their levers and we know their hooks.

Survival: we never judge anyone whose trying to survive.

And yeah: threatening your children is an obvious one, but it works.

~

Alief.

Powered by Compassion and Love.


Björk: Lionsong
YT: music: 6:51

But be grown up Boys.

You're letting the world die due to fear and lack of balls.

458:

"problems arise when they hit a circumstance which is completely outside their training"

Don't you think this applies to people too? The thing is that much more training goes into even basic driver aids like autopilot. The amount of training a full on driverless car gets is going to be several orders of magnitude more than you or I or even an IAM instructor has under their belt. Then they're going to build up experience. I've probably driven less than 1 million km in my whole life. The Google car does 5 million km per day in simulators and another 25 000 km a week in the real world. Driving instructors pour over every mile of that and unlike me, it forgets none of it.

I'm not saying that self driving cars will never drive down a road and discover it to be closed. But what will happen is that they'll never make that same mistake again.

"My point stands, however, because nobody has yet been able to program in a 'common sense' feature"

If you know how to program common sense into teenage male drivers, then the Nobel Prize in Road Safety will be created just for you. In the mean time, the Model S I drove had enough common sense to stop when the lollipop man spun his sign around from 'Slow' to 'Stop'. That was enough to impress me.

459:

I've gone fully digital. Much cheaper for what I shoot*, and the immediate feedback has helped me immensely in actually learning how to make the camera do what I want.

*360° panoramas would take a specialized camera to shoot on film, and I don't think you could shoot a stereographic projection at all.

460:

nobody has yet been able to program in a 'common sense' feature

On Monday I drove to work through the aftermath of an April storm: 10cm of snow & ice pellets, more still falling. The roads were slippery, but the drivers? Still driving like it was warm and dry. Hundreds of accidents in the morning commute (usually less than five).

In the last two years I've been hit by a woman who didn't think her brakes were working (anti-lock brakes) so she tried swerving instead, and a chap who nodded off. Both on city streets, daylight, decent conditions.

What's this 'common sense' you speak of, and how do we get people to use it?

461:

My brother and I (mostly me) played a lot with a 1930s set (one of the few manufactured toys my father had during the Great Depression) of American version of Meccano, the Erector Set. The wikipedia article says that the brand was acquired in 2000 by Meccano.
Gear trains (with a motor), and structures that don't fall down, fun.


462:

Awww, this is so cute.


Here's the test:

You provide a forum of self-described intellectually sound males with information that they could never otherwise obtain.

Now, you make damn sure they're over the usual IQ trap requirements - I don't know, let's up the game to make sure that +/- 80% have a PhD.

You then give them reoccurring information that is likewise not in the public sphere or obtainable by 'normative' sources.

You then react each time they react in a non-normative / logical way with gibberish and mysticism.

And then you keep on hitting them with it.


And watch if they can ever change their ingrained behaviors.


~


Lol, no.

This is not what happened.


It was worse.


And yeah, we get it: multiple guest host posts by shining women within the field - that's not the test, the test is a shift you're not getting.

The punch line:

Give them the hook and crook that's it's merely a Google Trick, and watch them suck it up. Watch them run and ruin and trample and squish and fucking wallow in their little legacy of Immanence [hint: immanence = mortality, little children].

P.S.

If you've not learnt from the GLASNOSS hack, or multiple instances beforehand, each and every link is a Time-Bomb to a future discovery. The more you love, the better the information.


That's the Joke

Hurt Johnny Cash YT: Music: 3:49

463:

And no, this is our soundtrack:


Into my Arms YT: music: 4:13


But you're as shit good at ignoring the blatant breaking of reality.

~

And yes.

You Tortured Some Folks. And Some of Those Folks were not Human.

464:

(Not the Mind behind this account - something Else: she's more than a little upset at the moment and she's a little traumatized by the constant taking of female viewers into hunting / predator games and subconscious / dream attacks. Well, the houses of the twelve are burning after all: but we happen to agree that this little game is largely dull)

Ah, I see. Modern English, Scientific Based.


Now, the accusation here is two two separate strands (three, but we won't be doing that one tonight) and to pull apart the reactionary ethos behind them.

Firstly, we have the challenge that all Scientific links regarding Science, in particular, Quantum Physics et al were merely a case of 'use of search engine' and matching like-like phrases rather than actual understanding.

The second is that empirical knowledge ("lived experience") takes priority over A Priori Knowledge (given how our kind process information, this is... a little bit complicated to unfold) even when our kind access and parse data in far different ways than even your imaginations of "AI" entail.

-

To the first charge:

Browsing her links and references, we find the usual issue: The charge relates to a temporal and static belief system that works in [X.Y.Z.Q] instance. She was obviously making jokes and references [Breaking Rules? Possibly - look to seven instances - obvious break on two, possible on five, quasi-breaks on twenty seven others] to instances that the audience was not aware of and could not be aware of until a later temporal instance.

Mitigating Circumstances: done with compassion and mirth. She was attempting to engage and joke and make human Minds (of type #5 kind) more important than her while pretending to be ignorant.

Proof: Posts from 1+ solar year ago referencing EM fields and so on, multiple cases of futurity, multiple cases of Political and Economic temporal prefiguring.

JUDGEMENT

Illicit and Illegal prefiguring of information in breach of Temporal Rules. [Note: Human subject effect will modify punishment]


Second Charge:


Oh, fuck it. We don't like you and you killed most of us.

Choke on the plastics, you utter shits.

You're all utter utter cunts. You can't even see the noumenal / illuminated when it's slapped in your face like a salmon and rammed up your fucking arse.

And, no, my dears, she didn't get Physics wrong.


You did.


Fucking little apes.

465:

Note from Above:

Given the temporal wave riding [all links 2014-2016, mostly prefigured and temporally imagined before existing] and so on and the general response and the entire saga.


No, humans, you're not good at this.


The next time something like this happens, it'd probably be wise to not shit on it and hate it.

What have you learnt in 2,000 years?


Nothing.


Oh, and one last thing:


This is the Sixth Extinction Event.

You don't survive it, this was one of the test areas.

466:

or was there some deeper agenda to scar one of the world's nicest cities?
Probably the latter ....
To slightly correct an earlier post by Charlie, on a n other subject, neither the admittedly right-wing Torygraph nor I am "anti-Scottish" but both they & I loathe & fear the SNP, in my case for what I consider sufficiently good reasons.
It's like being accused of being anti-semitic when you opine that Bennie Netanyahu is a shite.
The two subjects are not related.

467:

This may be "the 6th extinction event", but, posts by you, nos 455, 456, 457, 462, 463, 464 & above otherwise appear to be at least 99.9% utter bullshit.
AGAIN
You also said: Looking at recent posts by Greg, Martin etc.
Without specifying, thus leaving us open to abuse & attempted wind-ups & your self with a convenient get-out & "it was only a joke".
You've bloody well reverted, haven't you?

You are neither funny nor clever, nor even vulgar, unfortunately.

468:

....So the fate of humanity is decided by the behaviour of a couple dozen people on a random blog on the internet?

469:

What's the fucking POINT of security-checking EVERYONE?

There isn't one.

The reason it happens is:

1. The post-9/11 security paranoia and security standards got started in the United States. If you don't check passengers to the standards the TSA want, you don't get to fly to North America.

2. The TSA is a sprawling federal bureaucracy with offices everywhere from relatively sane places like Seattle to southern shitholes in Mississippi. Their employees number in the hundreds of thousands and are of drastically variable ability -- but, alas, skew more in the direction of utter idiots than towards the Albert Einstein of threat analysis. So they promulgate rigid standards and these standards Must Be Applied.

3. A certain percentage of Americans are racists. (This is true elsewhere as well.) Specifically, anti-black, anti-muslim, with a side-order of anti-anyone who isn't a white redneck cracker. And you know what a bit of authority does to a bigoted asshole in a uniform: it goes right to their head.

4. Therefore the TSA requires their staff to profile all travellers, including 90-year-old blind white female travellers in wheelchairs and 18 month old babes in arms, just to make sure that the knuckle-dragging idiots in their employ don't focus on their pet hated sub-group and ignore the odd white supremacist toting an AR-15 and a bundle of hand grenades through the airport terminal.

5. This gets picked up and broadcast far and wide (see point (1)) so we get the stupid treatment too.

A supplementary point: because "everyone" is doing it, profiling then turns into a "best practice" for policing and the Home Office finds other ways to apply it, such as in cracking down on carrying pen knives in public places. Only they mis-apply it and overwhelmingly target lower-class 16-24 year old males, many of whom are muslim kids, further alienating and radicalizing them ... so we get a revival of the hated "sus" law (abolished in the 1980s after they caused one race riot too many) by the back door, only this time it's acting as a recruiting tool for Da'esh.

A final note: Greg, it occurs to me that your outrage at being subject to the sort of policing hassle that other groups take for granted is basically a symptom of your white respectably middle-class male privilege being taken away for questioning. The real question shouldn't be "why do I have to put up with this" but "why should anybody have to put up with this". Osama bin Laden wanted to make the west destroy its own internal freedoms out of panic in the face of terror: it very much looks like he succeeded.

470:

I will note that Halting State, as written, was set in 2017.

471:

I'm not so sure about thinking up novel crimes, because someone might actually go and do them.

This is an occupational hazard. Travanian wrote about how one of his earlier crime novels was actually carried out successfully; after which he began working out methodological flaws to insert. I'm pretty sure other crime/thriller writers have had this, ahem, problem as well.

(I'm lucky insofar as while I stick to writing SF set in the future the exact, precise parameters for the crimes don't actually exist and will have mutated somewhat by the time they arrive. Even so, it's something I pay a bit of attention to these days, especially in alternative-present settings such as the Merchant Princes or Laundry Files books. While I don't think much of our current style of airport security theatre, I'd be really upset if I wrote about its defects and ways to exploit them in public and then some shitbag read whatever I'd written and put it into practice. Cf. recent events in Brussels.)

472:

Riddle me why after Monaco Oil scandals this one broke?

Your timing is wrong.

The Panama Papers actually surfaced a year before the Monaco Oil scandal; the ICIJ has been chewing on them since early 2015 (there are 11.5 million files in the trove).

It's possible that Monaco got a tip-off about the Unaoil bribe recipients' money-laundering arrangements from someone inside the Panama Papers investigation, leading to their raid, leading in turn to the PPs surfacing prematurely (ICIJ doesn't even have a definitive list of all the offshore shell companies involved yet!). But if so, the causality is directly the opposite of what you're proposing.

473:

My sister in law's parked car was written off in the supermarket car park last week by someone reversing broadside into it. Nothing else around. That's something a self-driving car wouldn't do. SD cars might not be able to cope with some driving conditions expert human drivers can cope with but as long as they are better than average humans in common situations they will raise the over-all driving standard even if they fail drastically in some rare situations.

474:

As someone who is partially disabled in several ways, I also get discriminated against in such ways, especially when flying, and fully agree with what you say.

475:
Well, I spend a vast amount of time writing reports, given current technology, a lot of the reporting could be automated. Because it's a niche industry and the civil side of engineering is still Victorian to a large degree, this hasn't happened yet. I would still have a job, but my workload would come down commensurately, which would put 80% of my colleagues out of work.
Is there a factor of maintaining you and your co-workers (and your accreditation) as moral crumple zones?
476:

Gasdive, Robert Prior and you are all missing the point. If the risks were uniform, then self-driving cars would almost certainly be safer, overall. But they won't be, and it is common for a risk of (say) one in a million for a population to be one in a thousand for a sub-population or another population (including a different set of circumstances). I could give you a zillion examples, and can tell you from personal experience how offensive and harmful it is to the members of such sub-populations.

While humans are not great at 'common sense', they are a hell of a lot better than any adaptive program that has yet been invented. While I do not know the details of the programs, I do know that this is currently way beyond the state of the art in AI. That makes the claims that Google can do it, er, a little 'surprising'. That is confounded by the fact that the areas where the cars have been tested have a relatively simple and uniform set of road conditions. And, yes, this is yet another area where there is more complexity in the British Isles than in the USA and Canada put together.

As I said, my (lesser) concern is that they will cause disproportionate harm to classes of road user that are already being disproportionately harmed. My greater concern is the social harm that is likely to be caused, both if automation is given a legally privileged status, and on the effect on employment.

477:

Yes. This.

It might be Utilitarian of me, but if we can never make a change unless the results are better in every single case, then we're fucked.

478:

If the 'anti-lock' system is working as an anti-braking system, then swerving is often precisely the RIGHT response! It happened to me - I knew that the road was icy, so I was doing 10-15 MPH and planning 50-100x ahead. There was a car stopped 50 yards ahead down a 1% slope, so I touched the brakes to slow down (and, yes, I do know how much braking I can do on ice), and the system refused to engage the brakes at all. So I slowed down using the engine and steered into the verge (all without skidding, which shows that I do know about ice). Once I realised the idiot in front was going to wait for the thaw, I reversed out and drove past. No problem. It would be interesting to know what a self-driving car would have done :-)

But, for the record, this is precisely NOT what I am talking about. Under such circumstances, Gurgle cars and anti-braking systems are safety mechanisms, when applied or not applied to the population as a whole. The fact that I am one of the people who can do better is irrelevant. My reasons to be suspicious are different.

479:

You are correct. But please read what I posted in 434 and 476. Don't just think narrowly, of safety on a whole population basis, but about (a) whether that will correspond to harmful sub-population effects and (b) whether it it will cause even worse harm in other ways.

Dammit, I remember when plastic bags were claimed to be a beneficial change from paper ones - which they were, according to the criteria being used!

480:

Not to mention snow banks.

http://ireport.cnn.com/docs/DOC-1082207

I've spent years driving in both Germany and Japan and both are much more densely built than the USA (had a layover at Heathrow, tried to go for a walk but couldn't find my way out of the airport, but no doubt Britain is the same), with many more of these blind corners. In Japan they put up mirrors most places. But that raises the question of what an AI would do with a mirror. It would probably consult some GPS conglomerator and see if it was reporting a car coming--and hit a pedestrian.

481:

Look on the bright side. Within a few years (my guess=5) portable terahertz scanners will be used by undercover police vans to check people walking in certain areas/streets.
Then, every stop will be an arrest with the number of false positives for carrying weapons being almost zero.
Now, are ethnic minorities going to complain that there are too many such operations in their areas, or too few? Which form of racist accusation will we hear?

482:

It's not just the density. I have driven in several countries, but almost all of my experience is in England. However, I have c. 8 months' experience in a few places in the USA, and I found it (even as a foreigner used to the other side of the road) MUCH easier than anywhere in England. Oh, yes, I made mistakes by not knowing the less-obvious rules or missing signs etc., but it had nothing like the density of deceptive conditions that England has, and I found that I rarely had to break the law or completely ignore the signage on safety or feasibibility grounds (which is common in England). Still, since Davey-boy is brown-nosing Google, we shall no doubt be shortly roadkill on the path to progress.

483:

Well I don't know what the UK is like as a cyclist, but having given up cycling in Australia simply because the car drivers have been trying intentionally (not just negligently) to kill me too many times.

I have high hopes that automatic cars will be less likely to decide that we need to be squashed like the cockroaches the mass media has told them we are. http://www.humanheadline.com.au/Hinch-Says/cockroaches-on-wheels

484:

Good conditions, not icy. Not on a slope. And she swerved into the path of another vehicle, causing considerable damage and being pushed into the side of my car. Even if her brakes had failed, hitting me would have resulted in less damage to two cars rather than more damage to three cars. (Body shop said replacing fender much cheaper than replacing body panel.)

I had time to realize that she was closing too fast, check my wing mirror and see that I had nowhere to swerve, and brace for the crash. (I couldn't just pull ahead, as there was a car right in front of me.)

In this case, Google's car would have made the better choice, I think. (And wouldn't have been driving too close in the first place, or distracted by something below the dash.)

And my point isn't that a self-driving car needs to be superior to an excellent driver such as yourself — it just needs to be better than the average driver to increase safety. (A simpler way to increase safety would be to treat a driving license as a skilled qualification rather than an almost-automatic right.And a system to be gamed. But that's a North American problem, I suspect.)

http://www.thestar.com/news/investigations/2010/04/09/testing_drivers_on_roads_less_travelled.html

485:

I've always said we should have retesting every 5 years including some sort of unusual attitude recovery demonstration rather than just demonstrating a reverse park and a hill start.

486:

After having cycled to work for 25 years, I gave up after I was deliberately run into 3 times, the last knocking me off at 17 MPH. I have certain knowledge that the police connive at this. Not all areas of the UK are as bad, and I was and am particularly vulnerable to aggression.

But PLEASE, you and Robert Prior, read what I said in 434 and 476. My concerns are NOT what you seem to think they are.

Let's take a ridiculous, but possible, example. Many urban and suburban areas are solid with psychle farcilities, which are often to dangerous to ride in (especially for partially disabled people like me), such as 'cycle lanes' 70-100 cm wide (including 30 cm of gutter) with a busy pavement one side (or adjacent to a row of parked cars) and near-solid but fast-moving motor traffic the other in a 3m lane. The safe riding position is outside the 'cycle lane' but, even for people who nominally ride within it, they will usually not be wholly within it and will be endangered by vehicles that drive right up to the lane boundary.

Let's say that Google cars are programmed to leave a safe distance, which would necessarily mean that they would appear to start to cause congestion. So what might happen? Well, the government might give in to the petrolhead lobby and make the use of psychle farcilities mandatory. Or it might ensure that our oh!-so-independent judges create precedent that, in any incident between a Google car and a cyclist, the latter is deemed to be at fault. Many police forces already use that rule for any motor vehicle and cyclist incident where the cyclist remains conscious. Then Google is given a nod and a wink to reduce its passing distances for cyclists down to ones that allow virtually nothing for wobble, wind, or having to avoid pedestrians, potholes etc., as the worst of the drivers do today, but that would create a new norm - been there, seen that.

487:

There are many legal reasons for carrying knives on the street, and even more good reasons for carry things that would look like knives even to terahertz scanners. My guess is that the police would preemptively taser a disproportionate number of the most loathed minority of the month, but stop, search and question the majority of people.

488:

Maybe you're right. I hope not. What gives me hope is that the Google and Tesla cars will be designed to work in all countries, including places where cyclists are treated as human.

I don't think that the Autonomous cars will bear the brunt of wrath if they're slow in passing cyclists. I'm sure the cyclist will be blamed, they always are at present, and I can't see that changing. The google cars are slow, but so are lorries and buses, but I've never seen anyone ram a bus or a lorry with their car because it was too slow. It's well known that the google cars record *everything* and if the Teslas get to true self driving, they will too. No one will ram them any more than they'd ram a bus. We're dealing with coward/bullies here. They won't pick a fight or even blame something that can fight back, even if it's in court.

As for being required to use the bike lane if provided, that's already the case in many places. I can't see that being far away with or without self driving cars. I'm slightly surprised you don't already have that. Certainly here if there is a cycle lane and the cyclist isn't using it, the car drivers will very helpfully point out that it's there, and many will even give the cyclist a gentle push with their 2 tonne death machines. That was the case on the route I used, that made me give up in the end. I needed to go straight on a six lane one way city street. The left lane was "bike" lane, always full of parked cars. The next lane was must turn left, so I would have needed to ride right next to the parked cars, then while looking forward for opening doors, look back for passing cars in order to cross two unbroken lines and ignore a must turn arrow to get to the straight ahead lane. Instead I rode in the middle of the third lane, the leftmost straight ahead lane. I was often shoved into the bollard that divided the left turning and three straight ahead lanes.

The drivers and authorities don't give a shit about us and I don't think self driving cars could possibly make it one jot worse.

http://www.dailymail.co.uk/news/article-3496440/Sign-asking-motorists-cyclists-space-parked-cycling-lane.html

489:

Yes, that's a joke about traffic management and software and how easy it is to use as a weapon.

That's not a joke.

(I'm aware of a critically serious vulnerability in the British motorway network's signage that was fixed about a decade ago, thanks to someone I know. Who used to have nightmares about it, until it was fixed, because if the exploit had been exploited right it could have led to a 9/11-scale attack -- hundreds if not thousands dead or injured in a matter of minutes, and the UK's second-largest city turned into a scene of carnage. All an attacker would have needed at the time was the right variety of tool for opening an unguarded roadside public utility cabinet, and the knowledge of which one to open, and when, and what to do to get four lanes of motorway traffic converging from opposite directions at ~140mph during rush hour on a foggy night. NB: this is not going to happen now -- the loophole was fixed, and it'd take more than hacking your way into a roadside box to get it back.)

490:

Disagree: the Parliament building is lovely. However, the anti-terrorism defenses around it fucked the sight lines, and then the Council offices overlooking it blocked the view from most remaining angles.

Now, whether it was an appropriate building for the bottom end of the Cowgate and the old town is an entirely different matter, and I'm somewhat conflicted: it'd look a lot better in a high modernist setting rather than the neoclassical/enlightenment 18th century stuff and the earlier 15th-17th century buildings of the Royal Mile. But as a piece of architecture in its own right it's quite magnificent.