Dear lord, its amazing. The mind does boggle.

I see Generative AI as a hot thing in photography, with replacement skies and other such tools. While I accept removing things from an image (heck, Ansel Adams did it!) , adding things to an image that were never there has always made me uneasy. Generative AI is an evolution of that concept, gone wild and mad.

Sorry to be that guy, but as a fan as a kid I have to do this: The book is "Charlie and the Chocolate Factory" NOT Willy Wonka and the Chocolate Factory, which is the name of the 1971 film with Gene Wilder, no idea why they changed it except Hollywood logic is sometimes weird.

please stop blaming the tool for the fool (and/or grifter) who uses it

ain't ChatGPT's fault an evil man (and/or stupid man) utilized it

this is as if I blamed the cabbage for Liz-the-Wiz (which I don't) or network television for T(he)Rump's high profile with American voters (which I do)

guns don't kill people, the store that sold the gun to they guy who shot the gun is more to blame for not upholding laws... with legislators being most to blame for being bought by Big Gun lobbyists...

{ yes I know it was a lettuce but cabbage sounds better when read aloud... cabbages are simply more suited to slapstick comedic sendups... I suspect its the higher fiber content and being less prone to wilting }

big adjustment by Hollywood, instead of black Oompa-Loompas were made orange... in later years the author was overt in his bigotry

as to the naming of movies... it's weird to recall this but the movie marquees were limited in number of letters and lots 'n lots of titles got trimmed down to fit... so... count the letters in both versions and also the spaces

"close encounters of the third kind"

IIRC... ended up in my childhood neighborhood as

"close encounter 3 kind"

I like the way they got the bot to apologise for itself and it was no better at that than it had been at the original task.

Nothing material is easy.

The ur-grift is the idea that somehow, that can be changed.

Before there is anything like language, there's expensive signals; if it costs to send it, it's a credible statement.

What we're seeing is a bunch of tech-priest grifting where the immense expense of running the grift is meant to convey that there's utility lurking in the device beyond its inevitable function as a staggeringly expensive space heater.

There's a lot to be said about how you can use computers to implement paradise, or at least that place where everything material is easy; it's a core trope about what computers do, and so far as I can tell, even knowing what computers actually do does not render one immune. (Vinge!) Anyone at all really ought to know better than to think you can materially implement adjectives, but here were are, again.

It ought to be funny, and maybe if it wasn't an enshittification accelerator it would be.

I saw a headline for this but didn't read past the first few sentences. I had no idea where it was located or that it was much worse than those few sentences I read implied.

Wow.

The purpose of the stunt might have been to make generative AI look bad.

even knowing what computers actually do does not render one immune.

It's a repeating cognitive bias in humans. Early 19th century had it bad with steam power (hell, the popularity of steampunk as a genre today is a throwback to that mindset of just under two centuries ago). It was superseded by electricity, and then by "air mindedness" in the 1920s-1940s, although aviation hit the thermodynamic wall in the 1960s, and space -- with today's propulsion systems -- is a specialized fork of aviation tech with higher energies but exponentially greater distances.

It's no accident that as the exponential increase in transportation speed/decrease in price of transport plateaued out by the early 1970s -- its last gasp wasn't Concorde, it was the rapid spread of multimodal containerized freight that cut the cost of bulk shipping by 98% -- the exponentially increasing curve of computing power took over and exhibited the same cognitive bias towards unlimited progress.

I wonder if it could be reenacted for Glasgow WorldCon. Karen Gillan has said she wants to be in any film version.

Yeah, I try to avoid the AI thing, because the expectation of results like this goes with the package.

At the very least one would need to put as many billions of dollars into developing a 'self-critiquing' function as was put into the 'generative function' for AI to be useful, but it really goes beyond that. We've previously discussed how to build true AI here, and what's being created is barely an idiot-savant. (In fact, an idiot-savant would rightly be insulted if compared to an AI...)

Essentially what's built is a job-stealing machine without any useful function, and the errors it can make with even the best, most expensive programming make it useless even for writing advertising copy. And it steals from everyone and everything, probably even from other AI.

We've previously discussed how to build true AI here, and what's being created is barely an idiot-savant.

What's being touted here as "AI" is not AI. It's a marketing label that's been slapped on top of large-scale statistical models trained on gigantic gobbets of input data (of dubious quality because it's hoovered up off the internet without adequate vetting).

It's a gold rush targeting VCs and PE investors with deep pockets now the shine was worn off cryptocurrency and apps and the dot-com boom. It's particularly inflamed right now because interest rates have risen and the startups need to bring in shitloads of revenue to pay off their funders.

This coincides with the press being butchered by the same PE corporations that have bought up publishing houses to milk the titles they own for goodwill built up over generations, as the behavioural advertising market slowly tanks (it was always a long con, but google rode it to a trillion dollar market cap).

I repeat: AI is part of a multi-headed hydra of a bubble that has inflated in the wake of the 2008 global financial crisis caused by the previous bubble (housing, loads, credit default swaps) exploding. This is the next bubble. It won't take much to crash it: we've already hit the point beyond which improvements in GPT models requires unfeasibly huge amounts of GPU horsepower and stolen data, destroying the entire intellectual property and media markets and warping the semiconductor industry roadmap.

It's all going to end in tears or mushroom clouds. Personally I'm hoping for tears, but I'm open to a sweet dinosaur-killer comet instead.

it steals from everyone and everything, probably even from other AI.

You know how there's a lucrative market in lead recovered from the batteries of U-boats that were sunk prior to July 16th, 1945? Because quite a lot of sensors require lead shielding that is not contaminated with radioisotopes released from A-bomb fallout ...

We're already beginning to see a problem with LLMs being trained on data sucked in from websites contaminated with output of earlier generation LLMs. Which causes a marked deterioration in the quality of output the newer LLMs emit.

What's being touted here as "AI" is not AI.
I tend to refer to it as "Artificial Stupidity".

For some reason, people struggle with the idea that what may seem to be linear/logarithmic/etc. growth trends are actually just an S-curve. Thermodynamics do not allow for infinite growth in anything (I'm just hoping that collective foolishness/stupidity will start levelling off soon).

working title of the movie version...

"A Poverty Of Imagination"

or

"Warehouse Emptied of Dreams"

{ oh golly me I've invented a new meme game }

The „wedding“ thing in the apology is easily explained if you assume that the apology was written by ChatGPT…

If you haven't already seen this, an article titled "Cellular functions of spermatogonial stem cells in relation to JAK/STAT signaling pathway" has been somehow accepted for review, peer-reviewed by two different people and recently published online in a Frontiers journal, despite being obviously fabricated by a computer program, including the figures. (The poor rat in Figure 1. Ouch.) It took a lot of bad publicity for someone to finally "notice" the fabrication and retract the paper.

I strongly suspect but don't know that the movie name change stemmed from the source of funding. The candy is branded "Wonka" and since the candy company was stumping up the cash they probably made sure they got that name right up front.

"What's being touted here as "AI" is not AI. It's a marketing label that's been slapped on top of large-scale statistical models trained on gigantic gobbets of input data (of dubious quality because it's hoovered up off the internet without adequate vetting)."

Reverend Stross, you're preaching to the choir!

Looks & sounds like a tory Conference brainstorming, whilst on acid!

There was a cryptoexchange that went bust, and it turned out the CEO was a completely made-up person. Had no existence in reality, stock photos, completely fabricated LinkedIn profile, etc. I had assumed that the people behind this particular scam would be a similar fabrication, it's interesting that they've been identified.

The wedding mention at the end of the second apology is quite interesting, though.

Part of the problem in this case, I suspect, is that Cartchy tuns, exarserdray lollipops, a pasadise of sweet teats sounds almost like something Roald Dahl might have written.

But this seems to be a regular kind of scam. There have been a number of "winter wonderlands" that turned out to be muddy fields with a Santa in a tent and a couple of reindeer in a pen. And of course there was the Fyre Festival which did the whole thing on a much bigger scale.

The law generally says that if you take money with no intention of providing the promised service, that's fraud. If, however, you do your best but fail, that's merely a civil matter: you owe refunds, but that's all.

Add in the Limited Company. A company has a separate legal identity and creditors can't generally claim money back from employees and shareholders. There are exceptions to that, and wrongful trading might reasonably be alleged in a case like this. But its up to the courts and likely to take some time.

So the scam goes like this. Create a limited company. Use it to set up some tissue-thin excuse for an event at minimum cost. Pay advance deposits where unavoidable, but put as much as possible off until after the event. Sell lots of tickets. Pay yourself a huge wodge of the money as "wages", possibly via other cut-out companies. Then when it all collapses promise refunds and disappear into the mist. When the inevitable court actions roll around, claim that you did your best but due to inexperience and misfortune things went wrong. Hopefully the court never gets around to "lifting the veil". If they do, well, the money is long gone anyway.

"...Youtube video, I haven't had a chance to watch it all yet..."

Wikipedia has it: http://en.wikipedia.org/wiki/Willy's_Chocolate_Experience

I would like to note again at this point, as I observed on social media nearly a week ago, that:

The Willy Wonka event in Glasgow feels like a potent metaphor for Brexit, to be honest: how it was sold versus the actual experience.

The caption to the image isn't even written. It's generated to "look right" in a purely visual sense, imitating the appearance of a thing written under a picture but not actually realising it's supposed to be English text it's generating. I think it's actually done remarkably well to produce something that does look like text with no more than spelling errors, but there is a vestigial "r" budding from the first "t" in "teats" which gives it away.

I wonder if they were trying to imitate (in a marketing way) this:

https://vangoghexpo.com/?cp_landing_term=cta_hero&cp_landing=cta_hero

Lots of people who visited it have liked it.

A lot of the descriptions on the marketing material seems similar.

The poor rat in Figure 1. Ouch.

My wife told me that when she was a kid, she had a pet rat not unlike that one. His name was Scrot.

please stop blaming the tool for the fool (and/or grifter) who uses it

ain't ChatGPT's fault an evil man (and/or stupid man) utilized it

This is implicit in what Charlie said above, but I am going to repeat it:

Behind ChatGPT is a marketing department. Behind the marketing department is thirteen billion US dollars (at last count). Behind the US$13B is a bunch of Microsoft execs who have decided that the survival of their company over the next decade depends on this product. And the sales pitch for the product, all the way down, is "You can skip doing the hard part. AI will do it for you."

I can absolutely blame the tool; it's inextricably tied to a system.

...and pay inflated prices to 'vendors'

USD$17.99 per roll of toilet paper

and it turns out, too late to track down, there was a truckload of foodstuffs sent to another location but billed to the 'event'

foodstuffs which in turn are loaded onto another truck, delivered to a legit restaurant, sold off at 90% of list prices

This little scam is a good example of the problems Science Fiction writers have when trying to write about the future.

To riff on J.B.S. Haldane: The future turns out to be not only weirder than we imagine; it's weirder than we can imagine.

You reckon? Surely "promise the moon on a stick, deliver a dog turd with luminous paint on" is one of the standard ancient scams. I'm not sure there aren't even instances of wild animals doing it to each other. The details of this instance are new, but so were all the other variations once.

I could see that maybe he thought it really would work, but he had to start selling tickets to fund the hiring of those expensive projectors. By the time it became abundantly clear that the AI-generated code was no where near capable of driving the fancy light show, it was too late.

I both agree and disagree.

If you look at early generations of any new tech, it gets silly. Repeating guns. Typewriters. Airships. Bicycles. Automobiles. Airplanes. Computers. Personal computers. Social media. Electric cars.

The thing is that’s the adaptive radiation. Then selection kicks in and the silliness mostly fades. Then the fuckers are ubiquitous. Then people start rebuilding civilization around them and they become necessary. Then investors start sucking them dry and they become enshittified.

If more SF writers simply followed this pattern, their writing would be weirder and more realistic.

The thing I’m unhappy about with disembodied AI is that it seems to be really good at BS. I’m not looking forward to civilization rebuilding itself to lionize transhuman purveyors of ubiquitous bullshit. That would seem to take enshittification entirely too far.

"please stop blaming the tool for the fool (and/or grifter) who uses it... ain't ChatGPT's fault... guns don't kill people"

Actually, guns do kill people. And generative AI fools people. That's what they're built and designed to do. Pretending otherwise doesn't help.

If modern guns had just been invented and murders using guns were suddenly becoming common, saying "it's not the guns fault, it's just a tool" would not be a helpful response: prevalent guns make murders much easier and more common, people would need to learn about that and decide whether to change their behaviours or laws to handle that.

Generative AI is an ideal tool for fooling people. It is explicitly designed, tuned, and built to produce that look good, that appear accurate, that are persuasive. A truly vast amount of money and work has gone into training it to be that.

So I expect the killer apps for generative AI will be individualized spam, targetted advertising, and fraud. It might - might! - also have other uses. But mostly that is what you get from a tool that can produce convincing-looking results tuned to its audience.

So cheaply-produced grifts will look more like the real thing, and people will need to learn about that and decide whether to change their behaviours or laws to handle that.

whomever @ 2:

Sorry to be that guy, but as a fan as a kid I have to do this: The book is "Charlie and the Chocolate Factory" NOT Willy Wonka and the Chocolate Factory, which is the name of the 1971 film with Gene Wilder, no idea why they changed it except Hollywood logic is sometimes weird.

IIRC Willy Wonka was the secretive, eccentric owner of the chocolate factory & Charlie was one of the five kids who "found" a golden ticket & got a tour of the "factory" ... and Charlie turned out to be the only one who wasn't a complete asshat.

Trouble is I haven't really READ the ORIGINAL book, only flipped through a copy after the 1971 movie came out, so the copy I looked at may have already had names changed to protect the innocent ...

OTOH, I haven't really watched either of the films either.

Princejvstin @ 1:

Dear lord, its amazing. The mind does boggle.

I see Generative AI as a hot thing in photography, with replacement skies and other such tools. While I accept removing things from an image (heck, Ansel Adams did it!) , adding things to an image that were never there has always made me uneasy. Generative AI is an evolution of that concept, gone wild and mad.

I've had a long standing beef with a nature photography group I belong to over "What is a NATURE photograph?. They have what I feel are unrealistic expectations about "Hand of Man" in the images, e.g. an excellent photo of a couple of bear cubs playing near a tree stump disqualified because there was an OLD fence post with the remnants of the strands of barb wire where a fence once was in the image.

Yet obviously manipulated images (deer floating above grass in somebody's back lawn) are perfectly acceptable and recently there has been much discussion of (and praise for) PhotoShop's new GenerativeAI tools.

PS: What the hell is a grassy lawn if not the "Hand of Man".

That Fyre Festival was the first thing I thought of when the story showed up in my news last week.

Kevin Marks @ 10:

I wonder if it could be reenacted for Glasgow WorldCon. Karen Gillan has said she wants to be in any film version.

I wouldn't be interested in seeing it unless they can somehow include Rory.

aitap @ 18:

If you haven't already seen this, an article titled "Cellular functions of spermatogonial stem cells in relation to JAK/STAT signaling pathway" has been somehow accepted for review, peer-reviewed by two different people and recently published online in a Frontiers journal, despite being obviously fabricated by a computer program, including the figures. (The poor rat in Figure 1. Ouch.) It took a lot of bad publicity for someone to finally "notice" the fabrication and retract the paper.

I'll see your "Cellular functions ..." and raise you Transgressing the Boundaries: Towards a Transformative Hermeneutics of Quantum Gravity

🤣🙃😟😕😂

At least Sokal did all the work himself.

icehawk @ 35:

So I expect the killer apps for generative AI will be individualized spam, targetted advertising, and fraud. It might - might! - also have other uses. But mostly that is what you get from a tool that can produce convincing-looking results tuned to its audience.

So cheaply-produced grifts will look more like the real thing, and people will need to learn about that and decide whether to change their behaviours or laws to handle that.

I'm afraid it's already here - DEEP FAKE robocalls.

Early 19th century had it bad with steam power

Don't forget atmospherics, an offshoot of steam engines. Feorag probably has some knowledge of them.

Obituary: The Atmospheric Railway (from Punch, 1848, on the financial collapse of attempts to power a train with pneumatic pipes and pistons by the South Devon railway)

Died last week, the Atmospheric Railway. Its death is supposed to have been hastened by the want of breath. When the tube was opened, it was found quite gone. Its loss is deeply regretted by a large circle of India-rubber buffers. A stone will be erected to mark the melancholy fact, with the following epitaph : – “The earth hath bubbles, and this is one of them.”

Bad news Mr jvstin & Mr S: AI-like substitution is already infecting photos taken by the handy communication device in your pocket. It HAS to, to compensate for the tiny lens (lenses) and when users want "great" photos automatically. Your "photo" was put together out of multiple pieces & substitutions both before & after you pushed the button.

The article below on Computational Photography is coming up on 5 years old, and we've about reached the prediction all the way on the bottom where "your phone says: you know, your selfie is shit. I'll put a nice sharp picture of the tower in the background, fix your hair, and remove a pimple above your lip."

https://vas3k.com/blog/computational_photography/

Yup: here's a particularly infamous example taken by a recent iphone camera in a bridal photograph (look at her hand positions in the changing room mirrors).

(Although that's not AI hallucination. Rather, the photographer's phone was in panorama mode, in which it stitches together a bunch of separate frames captured in very rapid succession: she moved her arms while they were shooting, and instead of a blur the result is a model whose reflections in two angled mirrors show her in different positions.)

There was also the infamous Samsung moon inserting camera, which looked for blurry moon pictures and swapped in a previously taken crisp one. https://arstechnica.com/gadgets/2023/03/samsung-says-it-adds-fake-detail-to-moon-photos-via-reference-photos/

Hypothesis: the people who use "AI" generated text because they don't know how to write also don't know how to proofread, which is why so many embarrassing examples get published.

That's a pretty easy hypothesis to come up with, but a lot harder to prove! (More likely they don't even recognize that proofreading is necessary, as they've swallowed the line that AI will make writing and editing obsolete.)

You can use it to knock out your new LP, if you don't fancy all that business about bothering to learn how to play anything properly - https://www.currentaffairs.org/2024/02/will-ai-enable-new-musical-creativity-or-undermine-musicians. (Can't speak to the quality because my galloping adult onset hearing loss means all music now sounds ghastly and wrong for me.)

Add to that a couple other things, like the way language is constantly changing and is thus a moving target, or that the particular subset of the English vocabulary used to address science fiction fans is different than the subset used to address fans of romances or mysteries, or even ordinary people who start their day with a bit of news, and that humans frequently accomplish this through 'feel' rather than an algorithm, and that keeping up with language in the fashion of a real reporter/editor is a full-time job by itself... I'd expect even a 'perfected' AI editor/proofreader to be both behind the times and also trained on lots of substandard language via the Book of Face or whatever, such that it's decisions would be marginal, at best.

I believe they're scraping pirate ebook sites and pretty much anything ever uploaded to kindle unlimited, the global public amateur slushpile. So good luck with that!

That's exactly my point, and why I put quotes around 'perfected' in my previous post, plus using the word 'marginal.'

The very best AI editor you'd ever get wouldn't be as good an editor as I am right now (it might be a better proofreader, at least where spelling is concerned.)

comedy spins on this Wonka horror

https://www.youtube.com/shorts/TQO3GweMsJk?feature=share

JD @ 43:

Bad news Mr jvstin & Mr S: AI-like substitution is already infecting photos taken by the handy communication device in your pocket. It HAS to, to compensate for the tiny lens (lenses) and when users want "great" photos automatically. Your "photo" was put together out of multiple pieces & substitutions both before & after you pushed the button.

Nah! I don't care.

I rarely use the iPhone camera and when I do it doesn't require any of that stuff (in the waiting room at the doctor's office, hold the page flat & make sure the text is in focus so you can capture the recipe well enough to cook it at home and you don't have to tear any pages out of the magazine, leaving it intact for the next patient to read).

When I do photography I use my "real cameras".

The article below on Computational Photography is coming up on 5 years old, and we've about reached the prediction all the way on the bottom where "your phone says: you know, your selfie is shit. I'll put a nice sharp picture of the tower in the background, fix your hair, and remove a pimple above your lip."

https://vas3k.com/blog/computational_photography/

I've been doing some of that stuff for 20 years now (it was, in fact, 20 years ago this February just past that I got my first DSLR).

Before that, I was already using PhotoShop5 because before I got the DSLR, I was using a Nikon Coolscan IV ED for 35mm film. Non-destructive editing (adjustment layers) came with PhotoShop7

[HINT: If your software doesn't do non-destructive editing, make all your edits on a COPY. Save your "original" image to go back to later if you want to make changes.]

I don't use the current PhotoShop ransomware, because I paid significant money for my perpetual licenses that Adobe no longer honors. My argument with Adobe is over their business practices, not the software.

I have no problem with image manipulation. I've done it myself. I've done it with film in the darkroom.

I DO have a problem with misrepresentation in images ... if it's manipulated, say so, don't try to pass it off as UN-manipulated.

PS: I have an older iPhoneSE & I don't think it can even do the photo manipulation newer phones can do & IF I AM going to do photo manipulation, I've got a whole computer built just for that purpose.

PPS: IIRC PhotoShop7 was also the first version that could work with Camera Raw files via a RAW Processor addon (named appropriately enough "Adobe Camera RAW".

Charlie Stross @ 44:

Yup: here's a particularly infamous example taken by a recent iphone camera in a bridal photograph (look at her hand positions in the changing room mirrors).

(Although that's not AI hallucination. Rather, the photographer's phone was in panorama mode, in which it stitches together a bunch of separate frames captured in very rapid succession: she moved her arms while they were shooting, and instead of a blur the result is a model whose reflections in two angled mirrors show her in different positions.)

I don't know why that image would be "infamous".

I think it's cool the photographer managed to capture it "in camera", but I wonder how many times he/she had to repeat the process to capture the arms in position, not moving ...

I think it would be a lot easier (and quicker) to take three images & merge them in PhotoShop.

YMMV.

Troutwaxer @ 51:

That's exactly my point, and why I put quotes around 'perfected' in my previous post, plus using the word 'marginal.'

The very best AI editor you'd ever get wouldn't be as good an editor as I am right now (it might be a better proofreader, at least where spelling is concerned.)

I turn spell-check ON and leave auto-correct (auto-DEFECT?) turned OFF.

Spell-check should just highlight the words it "thinks" are wrong and offer suggested corrections ... you can accept the correction, ignore the the suggestions or add the word to the dictionary (I frequently add words).

USE the tools that work & skip the ones that don't.

I think it's cool the photographer managed to capture it "in camera", but I wonder how many times he/she had to repeat the process to capture the arms in position, not moving ...

They didn't repeat it, that's the whole point: this just happened because they photographed a subject who was moving.

Re: '... maybe he thought it really would work, but he had to start selling tickets to fund the hiring of those expensive projectors'

Based on the Wikipedia article - nope! The guy's a repeat scam artist.

My understanding is that corporations (Ltd's/Inc's) must name their CEO/Pres/chief exec/owner as part of getting their corporate seal/papers OK'd. This usually means that the gov't OKing that corporate seal/papers does some form of background check on the principals.

Did the UK pols (i.e., Rishi and/or last 15 years of Tory PMs) fire the civil servants doing these types of checks ... checks that were put in place to help protect UK citizens/consumers?

Hmmm... just did a search on this. Looks super easy to register a corp online in the UK. However ... anything as simple as this form should also be simple/easy to cross-check against records like whether the people exist, have a prison record, unpaid taxes, etc. Maybe some journalist will look into this because this could be the tip of the iceberg re: UK-based scams. (I'll let someone else check how the EU OKs corporations. My guess is that the EU's plugged such obvious holes. Could be another reason some UK pols/ultra-rich pushed for BrExit?)

https://www.gov.uk/set-up-limited-company

Moving slit cameras have an inherent and essentially unpatchable vulnerability... I think that's what they call them, for taking very long and wide panoramic photos: instead of a normal shutter arrangement, they pull a strip of film past a narrow vertical slit in the focal plane while the camera rotates around a vertical axis. The projected image inside the camera is thus stationary relative to the film, so the width of the slit and the speed of rotation determine the exposure; the exposure time for any infinitesimal vertical strip of the film is the usual small fraction of a second, so moving foliage and the like still comes out sharp, even though it takes several seconds for the rotational movement as a whole to be completed.

A typical application for one of these is taking a school photo of everyone at the school at once, all lined up in a long row. And perhaps some of them twice, if they have managed to run round the back from one end of the row to the other faster than the sweep of the camera.

I'm not saying it's not scammy. But the whole thing does have vague hints of someone being conned by MRAs and AI peddlers

What is this "film" you speak of, Strange Person From The 19th Century?

What's kind of interesting is the juxtaposition of machine-generated vs artisanal nonsense: Séamas O'Reilly* is a great writer and AFAIK what he puts out under his own name is true, but he has a history of entertaining pranks (e.g. Banjo the Riot Elephant, or his first published fiction - a series of slightly bizarre letters to Metro's "Rush Hour Crush" missed connections column).

*he's the "took ketamine and got called into work to serve the President of Ireland" guy.

Moving slit cameras

I think the phrase you're looking for is rolling shutter?

The final question that killed it off being the longitudinal flap valve falling to bits from the physical and chemical interaction between seawater, frost, iron, and the leather of the valve, so they needed to replace the entire length of the thing at some awful cost.

There were two entirely unrelated developments taking place over those same few years: the South Devon atmospheric railway experiment, and figuring out how to make vulcanisation of rubber into a consistent and reliable process. This ended up with the material the railway had needed all along to make a reliable flap valve finally becoming available at the same moment they finally decided they'd had enough. But they didn't change their minds.

I reckon it is probably a good thing it didn't come along any earlier. If they had started off with a reliable flap valve, then at least it would have been working properly most of the time, so the shiny aspects like speed and silence - which it did achieve - might have been impressive enough that they thought it worth putting up with all the less obvious aspects that were a bit shit.

It was only 10 or 15 years after they got rid of it that the trains they were hauling over Dainton with locomotives began to get heavy enough that the train weight times the cosine of the gradient would have exceeded the piston area times the maximum possible pressure differential. So if they had got the atmospheric system working, it would now be leaving them screwed, and with a worse task to sort it out, just at the point when their traffic was properly beginning to pick up, and they would regret it much more than they did.

Brunel built some groovy structures but he did seem to lose the plot a bit when it came to anything dynamic. His locomotives also sucked, hence bringing in Daniel Gooch to take over the locomotive engineering side.

Atmospherics of course did become useful when people were doing it on a more sensible scale. The pipe and piston of an atmospheric railway, suitably shrunk, becomes a pneumatic message tube system, which we still use, if less than we used to. Perhaps more relevantly to the idea of "atmopunk", little vacuum-operated bellows thingies made of wood and paper/fabric used to be used quite a lot as well-controllable analogue actuators, for things like piano-playing machines that could do loud and soft.

Of course an atmopunk control system for something more complicated than playing a recording would have to use fluidic computational elements, which also had their burst of hype a few decades ago, but electronics is nearly always better. They do still show up in peculiar things they are naturally good for, like "chemical laboratory on a chip" systems.

What your school photos would have been taken on...

It's much the same principle as the mechanism of distortion that article describes, but applied deliberately in a controlled manner (and much more slowly) so the "distortion" it produces is an undistorted mapping of the inside of a slice of a cylinder onto a flat strip.

The concept of the Brunel/Gooch Broad Gauge (7' track gauge) makes perfect engineering sense and indeed can still be seen to some extent on parts of Western Region (the one I remember best being Bristol Temple Meads Station).
Where Brunel fell down more was in designs like his trans-Atlantic liners, where he was on, or possibly slightly beyond, the technology capabilities of the day.

More likely they don't even recognize that proofreading is necessary

That's for people who have heard of this proofreading thing and not seen the point, as opposed to people who have never heard of it at all.

There is a concept analogous with not trusting trust, per Ken Thompson. Successive of generations of students do not learn the processes that were mainstays in their area of expertise until the previous generation, when they were superseded. That is a sort of analogy for iterations of clean room rebuilds, I guess, while people actually learning "the old ways" do so through deeply hidden code. I may have the explanation a bit muddled, still have covid brain I think.

We should get a re-enactment of this event done by "Craig, the Scottish Tour guide" (the lovely Eleanor Morton)

paws
Not even wrong, I'm afraid ...
It's LOADING gauge that matters - look at Germany or the USA f'rinsatnce.
Alternatively, use two gauges, like India ( Actually 3 ) 5'6"" for main lines in non-mountainous areas & metre for almost all of the rest.

... a consumer fraud perpetrated by a company ("House of Illuminati") based in London, doing everything by remote control over the internet to fleece those gullible provincials of their wallet contents.

"House of Illuminati" does seem to have a London address. However, it's pretty clear the guy behind it is a Glaswegian, with a history of running a dubious local charity. (This bit is particularly amusing: "Mr Coull lists himself as a doctor, but this qualification is attributed to the University of Sedona, which claims to provide 'Metaphysical Degrees'.") He also attempted to run for Glasgow city council in 2022, but never followed through, so maybe Glasgow dodged a bullet there,

Ah, we did. Would have been too hard for her to resist, I am sure.

https://www.youtube.com/watch?v=Hb6T_iW97do

I'm not sure that the use of LLMs is the real issue here. Somebody who screws up like this using an LLM will screw things up regardless. The root problem looks to be that either they don't know they're screwing up or they don't care.

Both mental states are horribly common in computer-oriented folk (I am one). Everything in software is a cheap, cardboard-cut-out copy of something in the physical world. It's too easy to see the simplified, rubbish version as the thing that matters (Platonic archetype?) and the rest as the annoying, messy, over-detailed copy. "Why isn't this good enough?" is a standard whine of developers. And off we go to reduce the physical world to our standards.

Interestingly, the LLM has the same problem of agency with no taste or judgement. It truly doesn't know shit from Shinola. And in that, it's a very good simulacrum of an ignorant, incompetent and very vocal kind of human; you know, the kind who increasingly rule our lives. Artificial Intelligence is not that bad a name for it. Artificial Stupidity is better.

...ooooooh shit

takes me a while but eventually the rusty wheels doth turn...

that one photograph will be leveraged by defense attorneys of many nations as basis of court filings to have evidence withheld from the jury or outright redacted (lawyerverse jargon for "suppressed")...

photographs, videos, audio... maybe even text, e-mail...

not so much the unblinking eye of justice applied evenly as e might have been hoping for, when it would apply to the rich 'n amoral (0.1%) when they're hunting us (99.9%) for sport... you know, involuntary sex, contractual scams, misrepresentation, bait 'n switch...

[L]arge-scale statistical models trained on gigantic gobbets of input data (of dubious quality because it's hoovered up off the internet without adequate vetting)

Which is a big reason why every 'it will get better in the future' feels a bit off to me. You scraped the internet already, are you going to go for a smaller dataset? How do you sell it? How do you even know which parts to scrape now? (How did you earlier? Wait, you didn't so you just scraped it all...)

The vetting problem becomes harder all the time because there just is more and more crap generated to be scraped.

I admit that for smaller datasets and limited applications LLMs can be somewhat useful. Like I've said elsewhere, in my software development job automatic code generation is nice but typing in code is only a very small part of the job. I haven't yet seen any systems which would be so good as predicting what I want to type that I'd consider the effort in building them to be worthwhile.

Somebody who screws up like this using an LLM will screw things up regardless.

Still, there's a lot of stuff which the LLMs make easier to do. Generating posters, scripts and all that kind of stuff takes effort even if it is minimal effort. Even typing out something vaguely sensible-looking as a script is a time-consuming task, so using an LLM to vomit such things makes things easier. And making even bad posters takes some time and skill.

So this makes things easier and faster.

well... well... once again the nerds ara step ahead of the media

Last Week Tonight With John Oliver shreds Boeing and the FAA and reveals what is obvious: no hostile audits, no reliable safety

https://www.youtube.com/watch?v=ivLjlMBjQCU

More likely they don't even recognize that proofreading is necessary, as they've swallowed the line that AI will make writing and editing obsolete.

Oh, that line was around back when AI was mostly talked about in reference to prize bulls (as opposed to prize bullshit)…

Back around the turn of the millennium I was told with a straight face that children don't need to waste time learning spelling and grammar because spellcheck and grammar-check are now built into word processors. (By a high school principal, no less, which makes it even more terrifying.)

And I'm certain that you have tales from the coalface of publishing about the outsourcing/dumbing-down of copyediting.

I put this story in the same category as the recurring Rubbish Christmas Wonderland Experience was Rubbish - Children Disappointed event. Although to be fair, the 1995 Glasgow Worldcon on did draw an Outraged Mums Want Their Money Back local headline.

So this makes things easier and faster.

Agreed. So perhaps it should be IA not AI: "Idiot Amplifier".

The irony for something like this is that in the hands of people who have a clue, a show like this could actually be good and worth at least £35 a ticket - if it was done properly.

A few years ago, my ex-gf saw an ad for an immersive Alice in Wonderland experience/show in London. Tickets were pretty pricey, but not unreasonable by comparison with West End shows. Against my natural northern reluctance to put my hand in my pocket, I signed us up for it.

And it was, genuinely, amazing. In a warehouse space under railway arches, they'd constructed a small wooden labyrinth, where actors led small groups of people through it. All the walls were decorated with scraps of book pages and cards. Some parts were animatronic (the Mock Turtle), some parts were done by actors (a fantastic Tea Party scene with a mesmerisingly mad female Hatter), some parts were simply experiences (a shrinking process through an Ames room), some parts were projections or sound recordings, and so on. And the ticket price was supplemented by the final destination being a beautifully creepy flowerbed-themed bar serving rather nice (but expensive!) themed cocktails.

In the same kind of vein, I took my son to "The World of Beatrix Potter" in Windermere. It's simply beautiful - it's exactly like stepping into the illustrations from the books. They had a little show for the kids as well, and they'd taken the Disney tip of doing meet-and-greets with the characters too, but the real prize was the detail in the artwork covering every inch of the place. Well worth it.

The really depressing part of this story isn't just that people were being ripped off - it's that it could have been that good if it wasn't just being run as a grift.

Back around the turn of the millennium I was told with a straight face that children don't need to waste time learning spelling and grammar because spellcheck and grammar-check are now built into word processors.

As an error prone typist I leave grammar and spell checking on. Then after done typing read back over what I have typed looking at all the "errors". I make a lot. But at least 1 in 10 of what is flagged is not wrong. Just not something the Microsoft checker cannot comprehend.

I usually make 2 or more passes over what I have typed (not as much here as I should) and try and fix those 200 word sentences and such that the checkers don't catch.

"Idiot Amplifier" ==> "Idiocy Amplifier"

the problem is less the user-as-person as how it is being used as a service

perhaps the product category ought be re-labeled:

"Idiocy as a Software Based Service" (IaaSBS)

Re: '... well... well... once again the nerds ara step ahead of the media'

Hopefully Boeing's corporate policy insists that their senior management and major shareholders must fly the MAX. Apart from good PR. it's probably the quickest way to bring in safer planes or - failing safer planes - new management.

Oliver was on Jimmy Kimmel's late night show [see below] a few days after picking up his latest Emmy and mentioned that the show's production team tells the legal team about the main segment a week ahead of the on-air date. (Sort of a heads-up on what research they'll probably have to do.)

https://www.youtube.com/watch?v=rBE4cCTj1qE

Below is a Harvard student mag article about one of LWT's episodes: SLAPP suits. Good to know about esp. since there's at least one in-the-news pol noted for using them.

https://www.thecrimson.com/article/2019/12/8/wthh-john-oliver-bob-murray-lawsuit/

Charlie Stross @ 60:

What is this "film" you speak of, Strange Person From The 19th Century?

Film? That's the stuff cluttering up half my refrigerator.** 😏

According to what I've read, very popular with young people today.

See also: Rolling shutter

... why I think they had to do it more than once to get it right in camera.

** Film is easy to find, it's reliable film processing that's gotten' scarce.

anonemouse @ 62:

"Moving slit cameras"

I think the phrase you're looking for is rolling shutter?

Nope. Rolling shutter is an effect from focal plane shutters. It's why SLR & DSLR cameras have a "sync speed", the highest shutter speed** at which the leading curtain is fully open before the trailing curtain starts to close. At higher shutter speeds the shutter creates a slit that moves across the film or image sensor.

The slit camera is a different thing. The slit is fixed; during exposure the camera moves in one direction while the film is pulled past the slit in the other direction.

Panoramic, Slit Scan and Strip photographs of various subjects

** When using "flash" you want the shutter full open when the flash fires or else only part of the image is illuminated by the flash.

Flash Sync Speed

Damian @ 67:

"More likely they don't even recognize that proofreading is necessary"

That's for people who have heard of this proofreading thing and not seen the point, as opposed to people who have never heard of it at all.

There is a concept analogous with not trusting trust, per Ken Thompson. Successive of generations of students do not learn the processes that were mainstays in their area of expertise until the previous generation, when they were superseded. That is a sort of analogy for iterations of clean room rebuilds, I guess, while people actually learning "the old ways" do so through deeply hidden code. I may have the explanation a bit muddled, still have covid brain I think.

That raises a question in my mind ...

Why do schools no longer teach kids to proof-read their work?

Guy Rixon @ 72:

I'm not sure that the use of LLMs is the real issue here. Somebody who screws up like this using an LLM will screw things up regardless. The root problem looks to be that either they don't know they're screwing up or they don't care.

Too often I think it's BOTH at the same time.

Both mental states are horribly common in computer-oriented folk (I am one). Everything in software is a cheap, cardboard-cut-out copy of something in the physical world. It's too easy to see the simplified, rubbish version as the thing that matters (Platonic archetype?) and the rest as the annoying, messy, over-detailed copy. "Why isn't this good enough?" is a standard whine of developers. And off we go to reduce the physical world to our standards.

Interestingly, the LLM has the same problem of agency with no taste or judgement. It truly doesn't know shit from Shinola. And in that, it's a very good simulacrum of an ignorant, incompetent and very vocal kind of human; you know, the kind who increasingly rule our lives. Artificial Intelligence is not that bad a name for it. Artificial Stupidity is better.

Real life^(TM) does not allow you to load a saved game or re-spawn if you get killed. I think too many people today have forgotten this ... or never learned.

Back around the turn of the millennium I was told with a straight face that children don't need to waste time learning spelling and grammar because spellcheck and grammar-check are now built into word processors.

On that subject, did you know that modern spellcheckers watch their humans type and learn from that? If someone repeatedly types a word the spellchecker doesn't know, and doesn't fix it to match the spellchecker's suggestions, the software will eventually figure that this is some new word it wasn't told about and add it to the dictionary of acceptable words. (Good if you typed "Feorag," not so good if you typed "beleif.") So over time careless humans will actually weaken the robots trying to care for them.

I think every writer has tales of embarrassing clunkers that got through editing, and they're the professionals.

Actually, as I read your post, I realized the perfect correction: Artificial Idiocy.

"Film? That's the stuff cluttering up half my refrigerator."

Ah, yours too? :)

Though in my case the reason is that what I want is not easy to find. There seems to be plenty of black and white, and various colour films whose colour rendering and permanence is either unknown or is known to be unsatisfactory. I want Fuji colour film - Velvia and relations - and not only has there been a lot of uncertainty in recent years over what they are and aren't still making or planning to, there's even been a scandal over Fuji putting their name and label on Kodak-based film, which is something I'm particularly keen to avoid.

Processing, agreed, that's been a problem for a lot longer. Everyone uses the all-in-one machines that look like a giant photocopier; it was a long time before they worked properly, and then they started cutting corners in other respects, so you got prints with visible raster lines on. Slide films have fewer steps to get wrong, but still aren't "guaranteed" in the way they were when processing by the manufacturer was included in the price.

"According to what I've read, very popular with young people today."

I have read similar things, but they have also gone on to say that what it's mostly popular for is some utterly bizarre fashion for taking photos that are deliberately shit, using cameras of Christmas cracker quality and rejoicing in their crappiness. I was given a creaky plastic camera when I was a kid, which I abandoned as unusable after one roll of film because it had light leaks up the wazoo; I've seen the very same model praised as something to look out for in junk sales for the exact same seitilauq.

This disturbs me, because if film becomes thought of as something people are wanting so they can produce photographic annals of Volusius, its manufacture and processing are also likely to adopt the Volusian method.

However, all that aside, what I was originally on about was that not only is the example posted not an example of an artificially stupid computer getting its knickers in a twist, it's not even an example for which computers are a necessary condition; it's an example of an unintended consequence of using that algorithm, which people were already exploiting as a prank when the algorithm was being implemented on purely analogue systems.

And disastrous. Someone posted to a faceplant group I'm in that's for advertising open calls for fiction. The theme of their anthology was songs by Jethro Tull.

Number 1: a medieval-style inn. With a campfire IN THE MIDDLE OF THE WOODEN FLOOR.
Number 2, my partner noted that he's holding the flute wrong (and issues with number of fingers).

No, it's I want to wave a wand, and it'll do everything, and I don't have to check.

And two, more general things: for one, all of that... why, exactly, isn't chocolate the entire theme? It barely mentions chocolate.

And ..."threadreader". So, can someone explain to me why anyone who could step away from it for a minute would consider a website that you can only post 128 chars (sorry, now 240 or is it 256?) chars at a time a serious site for conversation, or notices, or ANY bloody thing at all?

Other than, say, the media, who in the sixties had sound bites of 42 sec, and now are down to 9 sec?

Twitter was originally designed to operate over SMS (text messaging) hence the short length. Then the service mushroomed and got repurposed for unintended uses -- almost nobody uses the SMS features any more, message threading/conversation was added later because users were trying to do that anyway, and so on.

Then Dilbert Stark got high, made a ridiculous offer to buy it, the lawyers held his feet to the fire, forcing him to pay up, the owners had a fiduciary duty to take the roughly $20Bn over market price he'd offered, and he's riding it into the ground.

It would have been a godsend, and a guaranteed pass, for English Literature O-level. Write some deep sounding bullshit about this passage in the style of the teacher's blethering about other things by the same author. Doesn't have to make sense, doesn't have to understand the passage or appear logically connected to any actually discernible aspect of it, just has to include the required buzzwords and an appropriate number of quotes. It's hard to see how the machine could go wrong.

86 - Well, schools (at least in Scotland) pretty much never taught proff-reiding, at least not going back to 1940s primary schools.

89 - That never actually occurred to me until now. :-D

90 - I'll agree that Kodak is no substitute for Fuji (reasons of colour balance for one thing). Have you tried Konika (formerly Sakura) stock?

"my partner noted that he's holding the flute wrong"

He taught himself to play it without reference to any form of official guidance, so he just did what was most natural to produce the sounds he wanted. He held it "wrong" for years and years and played most of their music like that before anyone told him about it.

Oh, I know exactly what it was all about (ask me how many messages I got on my pager (or, for two months, two pagers) when I worked at Ameritech in the mid-nineties.

But the entire idea of a website based on SMS messages was mind-bogglingly dumb in the first place. All I can see such a use for is snarky one-liners.

Never even seen it on sale, but will remember to look out for it; thanks for the hint.

I stepped away, to get a shower and lunch, and a thought struck me: the chatbots do not have spellcheck or grammar check enabled. From the ad, alone, it's obvious.

All I can see such a use for is snarky one-liners.

You are missing stuff like:

Emergency broadcast notifications (eg. tsunami alerts)

Status reports from software (eg. "mail queue has stalled") to a support team

News headlines with a link to the news agency headline

"I'm going to be in the local Nando's at 8pm; anyone joining me for a pint in $PUB afterwards?"

... And a bunch more.

In other words, your lack of imagination is not a limitation on the medium.

you mean...

he's unscrewing bolts from the wings to reduce drag weight?

uhm... where have I heard of such things happening elsewhere?

too bad the death of TW will remove what was a semi-usable "media for media wonks"

pistols at dawn, dude

I live for snark these days and the sharper the jab of any one-liner the (momentary) higher spike of laughter induced endorphins as easement of my bones

The Stable Diffusion artwork does not "understand" that the text atop the promotional images is text; it's just creating the sort of text like blobs that are normally found in such digital patterns.

Sorry, but I can do that with emails, or links to the news sites. "Meet me at" fits my description, but then, emails would do as well.

We still use them for system alerts today. We even let paying customer subscribe to those alerts via the website!

But we have alerts that either "go to all" or failover down the chain until eventually the system is sending SMS's via an android phone hanging off a USB socket. Or more likely, sadly beeping away in a deserted server room as it ponders the life of the last surviving human being in a world now populated entirely by machines...

And speaking of crssing the streams, an Australian economist has opinions about one of our stable atrractors and also one of our other stable attractors... Laundry Files and nuclear power :)

https://aus.social/@johnquiggin/112035934670247450

Missing information ...

When I mentioned lost online info, I hadn't read this article yet. I'm a non-techie and thought I'd screwed up when I lost some online info, but nope - losing online info appears to be pretty widespread.

'More than 2 million research papers have disappeared from the Internet'

https://www.nature.com/articles/d41586-024-00616-5

How this applies to this blog's topic [AI vomit] ...

So not only do we have to deal with possible AI hallucinations, we also have to be alert for random holes in memory (senescence?) because documents that we assumed to be available for the AI to read no longer exist in the online universe.

Anyone here have any idea how/why these research papers might have been erased from the Internet?

SFReader @ 57

On company registration and fraud in the UK.

https://www.theguardian.com/business/2022/dec/03/crisis-uk-companies-house-fake-directors-fraud

Sorry, but I can do that with emails, or links to the news sites. "Meet me at" fits my description, but then, emails would do as well.

I don't see anyone telling you to not do what works for you. But you seem to be yelling at us for using tech that, to be honest, you don't seem to understand.

It is very cool when I can link up two people I respect. I'm probably unreasonably happy about seeing Charlie tell someone "well actually I do have a new Laundry Files book out" :)

lost pointers

especially embarrassing is expired host contract or URL domain ownership

fractured URL

cross linked URLs

folder renaming

sever renaming

router renaming

hardcoded IP addresses of devices which at some point decide to assign themselves dynamically generated IP addresses

someone kick out the plug (data or electric) of a device

Mercury went into retrograde

"How this applies to this blog's topic [AI vomit] ... Anyone here have any idea how/why these research papers might have been erased from the Internet?"

"Bgzs FOS, write me a paper about genetic signalling in the abnormal genital development of rats, which is so inhumanly difficult and tedious to read that anyone trying to review it will be too zonked on headache pills to notice the errors."

Then smaller archives find they can't muster the resources to keep up with checking for these, especially since those who produce them know this and so prefer to distribute the stuff using the smaller outfits. It's even harder to check when their own data has been scraped and found its way into the guff. So they go off-line in self-protection.

Perhaps, anyway.

I can't speak for Whitroth, but, if I were to yell at you for using Twatter it would be for "assuming I use tech I have never taken any interest in or had an account on/with" rather than "using tech I don't understand for want of ability to understand".

Pigeon @ 90:

"Film? That's the stuff cluttering up half my refrigerator."

Ah, yours too? :)

Though in my case the reason is that what I want is not easy to find. There seems to be plenty of black and white, and various colour films whose colour rendering and permanence is either unknown or is known to be unsatisfactory. I want Fuji colour film - Velvia and relations - and not only has there been a lot of uncertainty in recent years over what they are and aren't still making or planning to, there's even been a scandal over Fuji putting their name and label on Kodak-based film, which is something I'm particularly keen to avoid.

In my case it's a combination of things ... several photographer aquaintences decided to get rid of all their film & film equipment and the film ended up in my refrigerator because it was otherwise destined for the landfill & I just couldn't ...

The other part is Large Format ... I still have my 4x5 cameras & film holders and film to go IN the film holders including a box & a half of Polaroid Type 55 (which has probably suffered the film equivalent of bit rot since I stored it).

... and some stuff I have NO IDEA why I'm still hanging on to it, including 110 film cartridges (3 of them Kodachrome64 which is NEVER coming back).

Processing, agreed, that's been a problem for a lot longer. Everyone uses the all-in-one machines that look like a giant photocopier; it was a long time before they worked properly, and then they started cutting corners in other respects, so you got prints with visible raster lines on. Slide films have fewer steps to get wrong, but still aren't "guaranteed" in the way they were when processing by the manufacturer was included in the price.

For several years I had access to a REAL darkroom (B&W and Color) AND I was running one of those mini-labs (Noritsu QSS-3301 with Kodak chemistry).

The largest prints I could do with the mini-lab was an 8x10 (although if I had wanted to hack the system, I could have probably produced 8x12 and 8x24 panaramic prints on it) ... but I kept it running in tip-top condition, with the best up-time in my district.

Good enough that the local professional sports team regularly used my lab for printing publicity 8x10s for players to sign & give to fans (about once a month during the season I'd get an order for 100+ 8x10 prints). I also did printing for one of the "Miss North Carolinas" (again 100+ 8x10s at a time) ...

"According to what I've read, very popular with young people today."

I have read similar things, but they have also gone on to say that what it's mostly popular for is some utterly bizarre fashion for taking photos that are deliberately shit, using cameras of Christmas cracker quality and rejoicing in their crappiness. I was given a creaky plastic camera when I was a kid, which I abandoned as unusable after one roll of film because it had light leaks up the wazoo; I've seen the very same model praised as something to look out for in junk sales for the exact same seitilauq.

This disturbs me, because if film becomes thought of as something people are wanting so they can produce photographic annals of Volusius, its manufacture and processing are also likely to adopt the Volusian method.

RICOH Pentax is bringing out a NEW film camera for the market. It will be a 35mm half-frame (with an image format similar to smart phones)

PENTAX Film Project Story #03

paws4thot @ 95:

90 - I'll agree that Kodak is no substitute for Fuji (reasons of colour balance for one thing). Have you tried Konika (formerly Sakura) stock?

That's at least as valid a comparison as the IBM PC vs Apple arguments we all know & love so well.

A lot of academic research is about as well-funded and institutionalized as a sci-fi convention committee or a free and open source software package. The personalities are just as eccentric too.

So what very often happens is that four young academics create the journal of so-and-so, get issued DOIs, then one changes institutions, another gets a teaching job with no paid time for research, a third becomes a data scientist and a fourth becomes a train conductor. A few years later the original institution is updating its content management system and erases the page for the journal ("hey, we emailed the contact email on file warning that the page would be deleted unless they stepped in to fix it, and nobody told us not to do it - we got some 'this address is no longer in service' messages but that just proves that the website was not important"). Maybe the train conductor and the researcher are vaguely sorry that the website went down but nobody will pay them to create a new one and they have new interests now. That is how a certain percentage of online-only academic publications are lost every year.

Some of the simplest things in website maintenance, like making sure that Cool URIs Don't Change, are the hardest to actually do.

assuming I use tech I have never taken any interest in or had an account on/with

I've never assumed such with him. And he makes it clear he doesn't or has never used it.

Did you actually just imply that brand1 film isn't better balanced than brand2 for specific fields of photography!?

UPDATE: A villain devised for the catastrophic Willy’s Chocolate Experience, who makes sweets and lives in walls, is to become the subject of a new horror movie.

paws4thot @ 117:

Did you actually just imply that brand1 film isn't better balanced than brand2 for specific fields of photography!?

Which is better - a wheelbarrow or a hand-truck?

I've used both. Neither is "superior" in every way. Depending on the results I want to achieve, one may be more useful than the other in a specific circumstance. In OTHER circumstances ...

Arguing about which is better makes no more sense than arguing about which computer platform is "better".

135, 120/220 or 4x5 sheet?

Arguing about which is better makes no more sense than arguing about which computer platform is "better".

I'll just drop 'ECS era Amigas with 4 Mb or more memory and a hard disk' for the MC68000 series here.

(I had an OCS A500 at the end of the Nineties, long after it was relevant, and never really programmed it, but many of my friends had Amigas and it was a brilliant machine - for its time. Atari ST was the competitor, at least here, but kind of lost the game. The hardware might have been part of it; Amiga had IMO better helper chips, at least in the beginning and later ST upgrades were too little, too late.)

Now I have the mini-Amiga re-issue, but it runs on an emulated MC680X0 and not the real chip. Sadly I think it's impossible to get new MC68000 chips with MMUs, otherwise I'd be tempted to build my own computer out of those. I haven't done anything real on the mini-Amiga either, I got annoyed because it comes with a joypad instead of a joystick and my trusty USB-Competition Pro doesn't map the buttons properly, so it's unusable.

I had an 8086 XT clone and later a 386, so I was never personally part of the Amiga-ST wars, but for a (seemingly) long time the x86 computers were worse gaming machines.

The social media rumor to actual news article pipeline has gotten frighteningly fast

Sadly I think it's impossible to get new MC68000 chips with MMUs, otherwise I'd be tempted to build my own computer out of those.

You might be interested in this project: http://www.apollo-core.com/ (Apollo Core 68080).

It is a 68000 implementation on a FPGA, but with a bunch of stuff that came out since the 68060 came out, including SIMD and 64 bit instructions.

{ my snarky emphasis added just because }

A new movie from Kaledonia Pictures is being rushed into production to capitalise on the global infamy enjoyed by the story.

what would crack me skull open is if the resulting movie becomes in of itself a mega-uber-ultra-success akin to Harry Potter or Friday the 13th or Final Destination: GHCU

GCU ==> Glasgowian Horror Cinematic Universe

so... how about any local legends of...

evil(ish) undead? or maybe the postal worker who never ever delivers to the correct address? or the fish-n-chips joint serving up prophetic imaginations? or the lawyer who neve told a lie?

Fuck me, this obviously needs to include the Gorbals vampire incident!

Re: ' ...company registration and fraud in the UK'

Thanks for the article - published back in December 2022!

I'm guessing that this is still a problem in the UK and disappointed that it's also a problem in some of the EU countries. (Not surprised by which countries though.)

HowardNYC @ 110:

Re: 'Mercury in retrograde'

Yeah, why not!

Pigeon @ 111:

Re:' ... so inhumanly difficult and tedious to read ... to notice the errors ... smaller archives find they can't muster the resources to keep up ... even harder to check when their own data has been scraped ... So they go off-line in self-protection.'

Immediate reaction was - that sucks for small publishers, next reaction - so why not see whether there are any bored retirees who're obsessive enough to take on overseeing/babysitting this sort of stuff. (Like the obsessive guy mentioned in the Guardian article on UK co. registration & fraud.)

Pixodaros @ 115:

Re: '... So what very often happens is that four young academics create the journal of so-and-so, get issued DOIs, then one changes institutions, another gets a teaching job with no paid time for research, a third becomes a data scientist and a fourth becomes a train conductor. A few years later the original institution is updating its content management system and erases the page for the journal ("hey, we emailed the contact email on file warning that the page would be deleted unless they stepped in to fix it, and nobody told us not to do it - we got some 'this address is no longer in service' messages but that just proves that the website was not important").'

I'm hoping these content management system people don't have access to dead tree libraries! Are archives that difficult/expensive to operate? Also tanks my trust in institutions if they're blasé about maintaining their research/academic papers/journals.

Charlie @ 118: Chocolate factory villain

Makes me wonder whether this is some sort of contest to see who can come with the vilest so-called children's stories.

I saw the Gene Wilder version wa.a.y back - still don't understand how that movie could ever be considered a children's story: lots of smiling, toys, sweet treats and jokeyness --- and zero humanity. Dickens wrote about pretty grim (real-life) scenarios but he was also able to show that there is such a thing as humanity (compassion).

Some of the demos I've seen for the AI make me think that it would be a great tool (not a replacement) for copy editors.

In theory, you should be able to do something like: "Read the following text. Give me a list of characters and a description. "

The output could then be reviewed by the copy editor, and if some character shows up with brown/yellow/red hair they may want to chase it down (unless the character wears disguises, then it gets trickier).

The point is, this is what computers SHOULD be used for, to amplify human capabilities, not replace them.

In theory, you should be able to do something like: "Read the following text. Give me a list of characters and a description."

That's your problem, right there: LLMs lie creatively. They'll give you a list of characters and their descriptions and two thirds of the way through you'll discover you wrote Barney the Dinosaur and Captain James T. Kirk into your strictly realist-mode 19th century novel. Just because the LLM ran out of facts derivable from the input text and added a few.

And proofreading to detect that kind of lie -- not a simple error -- is hard.

Charlie Stross @ 124:

Fuck me, this obviously needs to include the Gorbals vampire incident!

Why? Everybody knows vampires don't exist.

And the actress dubbed "Meth Lab Oompa-Loompa" is now cashing in on the experience (after suffering several days of Internet hate):

https://gizmodo.com/viral-oompa-loompa-cameo-greetings-kirsty-paterson-1851307899

Barneiosaurus is now considered to be the same species as the previously-known Ecclesiosaurus, first described by a pioneering but obscure early Canadian bone hunter who was later imprisoned as a suspected terrorist, following the rediscovery of the only recorded specimen underneath a pile of vampire's toenail clippings inside a piece of monumental architecture in the Gorbals cemetery, formerly used for archive storage by a religious institution dedicated to securing his release, to which he had donated it as a token of gratitude.

Nice passive aggression there. Show me were I said "YOU should not use it".

I said I did not ever want to use it. I mentioned having been stuck with pagers... which might have suggested to you that I know more about it than you do. Maybe a B.Sc in computer and information science, and almost 40 years of experience as a programmer and Linux sysadmin, having worked on mainframes, minicomputers, and workstations, and that I created a database system, and assembled one, might give you more of the idea.

I know what I want to use, and if you want to use stuff I consider ridiculous, that's fine... that's your problem.

Make them into roaches who "create" the chocolate "treats", and you'll really gross them out.

You wrote: "I'm hoping these content management system people don't have access to dead tree libraries!"

Sorry, too late. Early 70's, I was working as a lab tech at the Franklin Inst. research labs. A competitor to Current Contents was brought into the labs. And then they got hold of the Institute's library. They sold off stuff (from the 1800s), and they'd get a subscription to a journal, and stop the subscription when they lost the contract, or the contract ended...

ADMINISTRATIVE NOTE

These are the comments to my blog posts, not an open forum for old men to wave their walking sticks at passing clouds and shout angrily at each other!

(Also, I am deeply uninterested in all the film neepery.)

...what I meant to type -- 4Q long covid brain fog -- was "lawyer who was cursed to never ever tell another a lie"

huh... 1954... pre-twitter... thus no teenagers texting 'n TikToking... nor a 3 Ghz laptop with video editing tools...

"Hysteria spread throughout Gorbals, and soon among the wider population"

so... update the tech from 1954 to 2032 when there's 6G or some other brain rotting higher speed of data to really vamp up TikTok... and just enough genuine competition to lower the cost of mobile phones to the point everyone over the age of three has one...

some foolish brats decide upon a Blairwitch Project all their very own as a combo school hand in and TikTok mega stir up to draw in eyeballs to see ads and their mockup of this creature as the click bait... now add in illegal firearms... improvised thermite-based flamethrowers in the feverish hands of comix-nerds and the usual lowbrow Molotov cocktails in sacks carried by drunken truck drivers (excuse me, would that be sodden lorry drivers?)...

WhatCoundPossiblyGoWrong(tm)?

oooooooooooooooh...

you win the net this day... or leastwise this site...

hmmm... how about... genetically spliced termites whose purpose is to gnaw on crumbling wood in abandoned houses to produce downmarket sweets for the poor folk...

just so happens there's a generous amount of benzodiazepine in the sweet treats...

how's that for dampening down upon unruly ghetto youth?

cheaper than providing decent education or allowing 'em to work at decent job

Sorry for that I created some of your irritation.

just so happens there's a generous amount of benzodiazepine in the sweet treats...

That's a terrible (read: depraved and murderous) idea. BZPs aren't safe, they just have a much wider therapeutic index than predecessors such as barbiturates. BZPs are also ferociously addictive, although the on-ramp and off-ramp are much slower than with opiates. Finally, tranquilized folk are extremely dangerous behind the wheel of a car, or on a bicycle. They're also a danger to themselves on public transport.

And sweets laced with tranquilizers are going to get into everyone's mouth sooner or later. (There's a reason jellies -- tempazepam capsules -- were made a controlled drug in the UK in the late 80s/early 90s; BZPs are downers and addicts in Glasgow were getting horrible abcesses from injecting the gel inside the temazepam capsules when they couldn't get hold of actual heroin.)

It is not the university's journal, its a private organization run by the said four academics which faded away. The university just provides web hosting, and as none of their staff or faculty is involved in the journal any more, who should pay for the necessary work to keep it up?

Did I mention that making sense of the details as you migrate the site to the new software or version may require proficiency in multiple languages or graduate-level education in an obscure topic?

If its an old-school article which you can represent as a PDF, in theory the DOI could point to a server with the PDF and you just have to update the DOI when it changes servers (although its possible that the one academic with the password died in a tragic yo-yo accident or moved to Nepal and took a vow of silence). But that is not very machine readable, and does not work if there is say a main article (the "executive summary"), some appendices (the actual article), some sets of data, and so on. So more often the DOI points to a webpage generated from the database with a list of authors and links to their other publication, marked-up bibliographic information, one or more files with the contents of the article, etc. And as soon as you have that page powered by a database you have something which requires manual adjustment and updating whenever the university's systems change or the underlying software changes.

If you want a professional infrastructure team for your journal you could partner with one of the giant academic publishing corps like Eisevier. They will keep it up forever because they use it to squeeze as much money out of universities as possible.

And as soon as you have that page powered by a database you have something which requires manual adjustment and updating whenever the university's systems change or the underlying software changes.

Over on Reddit people keep asking on the Wordpress forum why their "simple" web site request for their small store keeps being quoted by most developers in the $1000s. Or even over $10K. Why does it have to be so much and why is there maintenance? They keep thinking they are buying a hammer or circular saw at the local home center. Not something that has to be maintained over time. And in their mind things like credit cards are easy. After all they have this little terminal that Costco sold them for $300 that handles it now. Why is a web site different?

Oh, they want people to be able to BUY, have accounts, save credit cards, handle shipping charges and options, inventory with pricing and pictures, and so on.

{ takes a bow }

horrid being the point

the only difference between your Tories and our Republicans is that "U" missing from "coloured poor folk"

{ to be clear about my politics, this is a big loud 4Q to both versions of amoral 1%ers }

if they had thought of it soon enough -- maybe in the 1990s -- there'd be bags of benzo-laced sweet-secreting termites in every impoverished neighborhood on both sides of the Atlantic

as to addictive properties... why else do you think I selected benzo?

And as soon as you have that page powered by a database you have something which requires manual adjustment and updating whenever the university's systems change or the underlying software changes.

One day (in the late 90's), a suit from headquarters came to see us and asked the same kind of question. The N4 nuclear plant series were the first that were designed using 3D software instead of a drafting board. How could we garantee that plans and isometrics would be available for the lifetime of the plant (~80 years), as was mandated by law and also common sense?

I was tasked with answering that question (in a formal ISO9000 compliant way), and after a bit of research came to the conclusion that there were two ways of achieving that:
1- print everything on paper and store it at the site
2- create and maintain a team dedicated to updating the database software and data and other various programs.
Of course, solution 1 was chosen as being by far the cheapest and easiest to implement, but I had a feeling that the suits were disappointed I didn't come up with some magical solution.

When I was in college in the late '70s, Valium were a popular recreational drug. They had the interesting property that mixing them with alcohol resulted in memory loss. Folks mixing them would appear to be reasonably coherent (*), but the next day have no memory at all of the events. They'd wake up having no idea what had happened to their stash.

(*) In context that is. When everybody was drinking pretty heavily, the standards for coherence were pretty low.

create and maintain a team dedicated to updating the database software and data and other various programs.

Autodesk Revit has taken over much of the architectural and other design worlds. And is rev'd every year. And they stop issuing updates / fixes after 3 years. Not as long as a nuclear power plan but most commercial buildings take over 3 years from design start to finish. S firms get stuck with keeping the old software going or closing their eyes, hitting the IMPORT button, and seeing what happens.

For data and publications that are "just files" there is a solution. Some universities maintain a repository for academic results, where deposited files are kept indefinitely. The implied intention is to keep them for a long as the university library continues in existence. The University of Cambridge has such a repository, and the university library has been around ~800 years, so the persistence is probably quite good, looking forward.

There is still the problem with being able to read the files after some great time. The repository curates metadata, not the file contents. It will not re-serialise into next century's preferred format for publications.

Any data-set that is active and searchable like a database is not supported (we tried, on one project, and they were not keen). You could deposit an SQLite file and hope, I suppose.

So the obvious enhancement is "read this text; using only information from the text, give me...". You can say that now, but I don't know if the LLM would understand what you wanted, or how to comply, or whether it would actually try to comply if it did understand. I'm not sure that we could establish these things with a black-box test.

found it!

what happens when people cannot read content for its information

(not sexist there's dozens of guys doing this exact response)

https://www.youtube.com/shorts/JncsrRSy46c?feature=share

Update: house of illuminati claims "copywrite"

https://filmstories.co.uk/news/the-unknown-willys-chocolate-experience-event-organiser-unaware-of-film-plans-we-will-protect-our-interests/

This calls for a lawsuit with ChatGPT legal briefs.

"How did you do that?"

These people will keep tech support employed forever.

I wonder if there are any laws that prevent a judge from using ChatGPT to write the decision... Codes of ethics, sure, but would it be illegal?

You can't copyright names.

Also, in US case law there's a precedent that you can't copyright LLM output. Not sure how that'd translate to Scottish law, which is not the same as English law, which is not the same as US law (although English/US law forked in the late 18th century, and English/Scottish law was legislated by the same assembly from 1707 through 2000).

I suspect this is "House of Illuminati" trying a gimme for money to make them go away. If so, it's not going to work without lawyering up, and everything I've seen so far suggests that HoI is a low-capitalization grift.

stirner @ 143:

"And as soon as you have that page powered by a database you have something which requires manual adjustment and updating whenever the university's systems change or the underlying software changes."

One day (in the late 90's), a suit from headquarters came to see us and asked the same kind of question. The N4 nuclear plant series were the first that were designed using 3D software instead of a drafting board. How could we garantee that plans and isometrics would be available for the lifetime of the plant (~80 years), as was mandated by law and also common sense?

I was tasked with answering that question (in a formal ISO9000 compliant way), and after a bit of research came to the conclusion that there were two ways of achieving that:
1- print everything on paper and store it at the site
2- create and maintain a team dedicated to updating the database software and data and other various programs.
Of course, solution 1 was chosen as being by far the cheapest and easiest to implement, but I had a feeling that the suits were disappointed I didn't come up with some magical solution.

It seems obvious to me that solution 1 is the ONLY reliable solution ... although my OWN bias would tend towards making backup copies of all the documents on microfiche.

A friend of mine used to work in IT for a British government agency that dealt with infrastructure.

At one point he was charged with vetting the documentation filed by a construction firm who were building a major traffic-carrying bridge.

They were most irritated when he told them that no, providing a single copy of their proprietary database software running on a Windows laptop (Vista) with no other import/export option for the bridge schematics was not acceptable and did not meet the requirements for data retention and accessibility over a 75 year period ...

(It's a bridge! If it fails catastrophically, people probably die. If the conditions for which it was designed change drastically, eg. by climate change, coastal erosion, heavier usage, then it may fail. And if the maintenance crew don't have easy access to the design they may not be able to work out what's going wrong in time to save it. This nearly happened to the Forth Road Bridge a few years ago -- there's a reason it's only open to buses, bicycles, and pedestrians these days, and there's a shiny new bridge next to it that was build in a blinding hurry. Nobody wants to be responsible for the next Tay Bridge Disaster!)

A few years ago one of the trades chaps at my school asked if I had any data on water pressure in the science labs. He was trying to do repairs to one of the five(!) heating systems in the building and had no schematics or blueprints to refer to. Those things _used) to be stored in the head caretaker's office, but a decade previously the Board decided to centralize storage so they were shipped off to the main office to be digitized.

Apparently the clerk in charge of running the scanner didn't notice that the building had five different heating systems, so digitized the last one built (assuming that it was the latest revision to a single system).

There's something to be said for retaining physical copies. And having your filing clerks actually know something about what they're filing…

When I first started watching Stargate SG-1 I thought it was silly that the "ancients" left all their documentation on stone tablets. Another decade experiencing the ephemeral nature of digital storage gave me a different perspective....

Zoë on Sluggy (https://sluggy.com/#2024-03-05) gets it right - ChatPCP.

(Fully Immersive Actual Reality Tay Bridge Experience: you get a guided walk along the new bridge, to take in the spectacle of the remnants of the old piers alongside it. The tour guide plays wind and steam train sound effects on a ghetto blaster and when you get to the bit in the middle they push you off.)

Guy Rixon's post demonstrates the matter pretty well: much of the reason the library can succeed in being a useful repository for 800 years is that the data format and the equipment required to read it have stayed the same for 800 years. Modern methods go through multiple iterations of becoming increasingly obscure, for both those aspects in independent cycles, in less than a tenth of that time, and the difference is so stark and the signs of anything other than it getting worse so nonexistent that it's impossible to justify doing anything other than continuing to keep everything on paper.

Perhaps plain text files could provide a similarly long-lasting data format in the computer world, but I can't see anything comparable being possible with the equipment required to read it.

In either case you are imposing a kind of de facto "house style" in that anything the institution can accept for publication/archival must stick within limits little different from "what you can do on a typewriter" plus a few etchings. You naturally have to forget about including a database and a fancy custom client for it, and things like that. I don't really see this as much of a problem. As long as the data is there, and people can read it, the researchers en l'an 2800 can fettle up their own methods of getting it into whatever database package happens to be flavour du jour at the time, and still find that easier than trying to get some ancient Vista package to run. Plus they'll probably have invented so many more funky things to do with it by then that being limited to the ones we like doing now would if anything be even more awkward for them.

Adding PCP to semi-(sweet-)hallucinated fully immersive experiences might just be the missing link! Enhance your imagination!

*Offer varies by region and international criminal law treaties

This calls for a lawsuit with ChatGPT legal briefs.

How about legal arguments delivered in verse by a Greek chorus of lawyers in green wigs?

Wonky doodles, fiddle dee dee,
We have a lawsuit waiting for thee;
Wonky doodles, fuddle day day,
Mess with the estate and you will pay.

I once worked in a team whose (company internal) abbreviation was 'PCP'. I found it hilarious but I think not all of the team members were that delighted or even knew other meanings.

Absolutely. Dateline mid 1996CE - I had to read data recordings made on DEC PDPs, using ADA83 running on Sun SPARCStations under SunOS 4.1.3 (if I've remembered the OS revision correctly). This required knowing that the PDP used a 16 bit word length, and placed the LSB of the recorded value at bit 7 (indexing from 0) and the MSB adjacent to that at bit 8, so string slicing the PDP recordings by bytes, and swapping odd and even byres. Thankfully both systems were at least "same endian".

Charlie
And, as we all know the greatest monument to Victorian Engineering still stands as a lesson for future ages.
And, yes, the successive operating companies & owners have kept VERY CAREFUL records.

Perhaps plain text files could provide a similarly long-lasting data format in the computer world

You appear to be completely unfamiliar with SGML and the huge body of work on archival data formats that has been developing since the early 1970s, when mass storage sufficient to deal with the problem became available. (SGML -- standard generalized markup language -- is a semantic structured data representation intended for textual documents, developed by IBM and recognized as an ISO standard since 1986; XML and HTML are vastly cut-down subsets of it that very much lost the plot. Javascript was the original sin ...) Other computer documentation formats have been stable for longer periods of time, eg. the UNIX man page format (first released to the public in 1972, in use for over 50 years now).

Images are harder because the storage requirements are so much larger that compression algorithms are essential and so serialization is non-trivial.

But this stuff isn't rocket science and it's only the ineptitude and feckless indifference of commercial software vendors that has kept stable formats from becoming ubiquitous by now.

Guy Rixon's post demonstrates the matter pretty well: much of the reason the library can succeed in being a useful repository for 800 years is that the data format and the equipment required to read it have stayed the same for 800 years.

If I give you a piece of 800-year-old writing, do you think you could read it*?

*Offer not valid to people who've learned Middle English and Middle English orthography, since you're exactly my point: you've been taught to read the data format.

...or mayhap bikini briefs

though I'd opine for g-string, which are the briefest of briefs

Hell, if you were given a piece of 570 year old writing printed in a language you understand, but using the typographical and layout conventions of the 1650s, could you read it?

I've seen a Gutenberg Bible and even if the language was comprehensible the typeface, kerning, line spacing, lack of familiar paragraph markers, etc. would render it opaque!

I wouldn't say "opaque" (well not yet anyway) but certainly "damned hard work".

still having flashbacks to the nightmare of users complaining their keyboard was missing an "any" key

this, due to so much software in both MS-DOS (1980s) & WIN (1990s onwards) displaying the prompt "press any key to continue"

in one memorable case, a bunch of extremely high salary employees at a gigabuck bank had a contest -- fifteen guys each tossed in a thousand dollars cash into winner-take-all kitty -- with the victory condition defined as which of 'em could reduce a tech support nerds to tears first

in my case, I was dragged into doing coverage due to staffing shortages during flu season and overall attrition (low salaries, miserable circumstances, no hope of advancement)

I was determined to never be forced into coverage so I drew out calls as long as possible along with scheduling way too many in-person responses by technicians

so unknowingly my scheme and their contest dovetailed and the guy who got me in the random timing of things was dead last loser after support call last 93 minutes

...and no, I was never dragged from my own assigned tasks to provide coverage

No need to go back 500+ year. The US founding documents are hard to read in the original images for most English speakers. Even in the US.

All kinds of typography issues plus things like what we call "f" today being used for things where we use an "s" today.

Then how about cursive? Do they still teach it in the UK and other English speaking places? Apparently "kids today" are not being taught it in the US much anymore so they can't read grandma's letters.

Then there is a 5 generation genealogy done by my mother in law in her pre-teens or early teens in Germany. It was a thing to everyone to do such. Back when. Written in High German (Hochdeutsch) and using a very ornate script (I forget what it is call), well, finding someone to interpret it is hard. Even amongst Germans. Most of those who can read it were born before the 50s/60s.

this, due to so much software in both MS-DOS (1980s) & WIN (1990s onwards) displaying the prompt "press any key to continue"

In the 70s and 80s offices getting computerized for the first time, at least in the US, got very very very upset when it was explained that no the lower case "l" can not be typed when you need to enter the number "1". Many low to mid range typewriters didn't even have a separate key for the number "1" and thus all kinds of muscle memory generated errors non stop with people trying to enter numerical fields.

After all they'd been doing it with the letter "l" key on their typewriter since Moses. Or so you'd think from the dressing down you'd get when caught up in the complaining.

But this stuff isn't rocket science and it's only the ineptitude and feckless indifference of commercial software vendors that has kept stable formats from becoming ubiquitous by now.

How much is ineptitude and feckless indifference, and how much is a desire to lock customers into a proprietary software ecology where they must continually pay to access their own data?

One reason I keep my old computer running on outdated software is that I have a 12 TB Aperture photo library. Apple no longer supports Aperture, and I don't want to pay Adobe annually to access my photos. At one time Serif was planning on releasing a Lightroom competitor (like Affinity Photo is a Photoshop competitor) but that was years ago and it looks like it won't happen.

how much is a desire to lock customers into a proprietary software ecology where they must continually pay to access their own data?

This attitude predates pay-to-play or cloud storage services by decades. It's easier to roll your own proprietary database with support for only those indexing/metadata features you specifically need than to implement a full-blown standards-compliant system that supports everyone's necessary features. (SGML is very much a kitchen sink approach to semantic information storage: IIRC the spec for a generic DTD/SGML parser using DSSSL pretty much mandates a lisp or scheme interpreter.)

still having flashbacks to the nightmare of users complaining their keyboard was missing an "any" key

One of my acquaintances used to work tech support for the PMO. On Mondays many of the computers didn't work. He quickly learned that pointing out that the cleansers had unplugged them to vacuum the carpet got him in trouble, because the real issue was that the (self-)important people who worked there didn't want to crawl under the furniture themselves.

So he listed the solution as "rectified power supply discontinuity" on the support ticket, which sounded technical enough to avoid bruising egos, and collected his paycheque…

IIRC the spec for a generic DTD/SGML parser using DSSSL pretty much mandates a lisp or scheme interpreter

IIRC = If I Recall Correctly. (I assume?)

No idea what the other acronyms mean. Or why a speech impediment would be necessary :-)

I remember some Amiga game which had a screen which basically said

• press any key to see the instructions
• press a key to start

It took us some time to figure out how to start the game.

It's easier to roll your own proprietary database with support for only those indexing/metadata features you specifically need than to implement a full-blown standards-compliant system that supports everyone's necessary features.

In a reverse bit of this, PDF files. Which started out as a way to print paper like things to a file you could send to someone else to read. Now, OMG. Security, text fields for entry, layers, and all kinds of nonsense.

A while back someone at a Mac club user meeting was getting upset that we couldn't tell him how to make PDF reader A give the same view as PDF reader B when viewing PDF files generated by a local government iMaps/GIS system. (A black hole to infinity of conflicting standards in itself.) He kept saying, "BUT IT IS A PDF FILE!". We kept trying to explain that PDF was a standard and not everyone did it right. He left the meeting and never showed up again.

IIRC, SGML was "designed" by a lawyer, Charles Goldfarb, who became fascinated by computers and started selling IBM machines as a sideline. Standardized Goldfarb-Mosher-Lorie (the other two guys who helped him). The syntax and grammar are so complex they don't match any formal grammar families. So you need a full programming language to parse it, theoretically, because existing parser systems are all defined against formal grammars.

All kinds of typography issues plus things like what we call "f" today being used for things where we use an "s" today.

Oh, you mean the 'long s' ('ſ')? If you're calling it an 'f', you're confusing it with a different letter that is drawn differently. It's a variant form of the lower case 's', but not used as the last letter of a word. (c.f. Ancient Greek 'lower case' sigma which also has σ except at the end of a word, where it uses ς)

It's an easy mistake to make as few people these days are trained to distinguish them, and to that extent you make a valid point.

(I'll leave the thorny difference between 'Ye' and 'Þe' for now.)

you just had to open that can of worms... horrid, spiky smelly toxic-to-the-touch worms

one of the many reasons I tried to move from coding to design-n-documentation was the repeated and undying refusal of people to understand {0,1} NQ {O,l} and therefore typing upper “O” or lower “l” into zipcodes 'n phone numbers was wrong never mind MS Excel spreadsheets...

you got any idea how insane it gets to audit a workbook of fifteen tabs which when printed was sixty pages? and was composed of about 70,000 cells? consisting of a zillion formulas? now with 20/20 hindsight I know how to perform automated audits for many forms of corrupt data but back then I did by eyeballing 210,000 cells one-by-one... had to perform the examination 3 times due to my fear of missing something...

whereas purging any tainted databases is straightforward table lookups for valid zipcodes 'n phone numbers any of which are tainted can be resolved with about two hundred lines of code for parsing thru tedious, meticulous business rules... but due to unrepaired flaws in the application(s) in use, the data kept getting tainted over 'n over... so the repair thingie had to be run every morning...

and then there's all the other horrors of tainting 'n corruption 'n cluelessness I ran into over 40 years... it is going to take at least half a bottle of cheap arse vodka to wash out those memories

you opened up that can, now you get to stuff all those toxic-to-the-touch worms back into the can

{ cue: whimpering }

The script was very probably Sütterlin (the German printed font was Fraktur).

The long s lacks the crossbar (although bad print quality can make it hard to distinguish the two).

ITYM Standard Generalized Markup Language. It was an attempt to do for semantically tagged textual information what SQL and the Relational Algebra did for databases. Goldfarb went down that rabbit hole initially because he wanted to store, search, extract, and export complex legal documents in a meaningful way.

Basically any large ongoing technical publishing project that needs a huge catalog of text with metadata uses it. I seem to recall US DoD documentation and specifications require it. XML is a very cut-down minimalist subset of SGML, with web affordances like hyperlinks and (blech) Javascript. CSS in turn adds formatting hints to the XML DTD -- a DTD (document type definition) is the prologue to an SGML document and defines its structure. (In programming terms: an SGML document is a data structure, and the DTD is its type declaration.) DSSSL was a metalanguage for specifying formatters for DTDs; you feed DSSSL a DTD and some lisp-like code and it spits out a renderer, if I recall correctly.

It got picked up elsewhere: for example, in the late 1980s the Royal Pharmaceutical Society coded up the British National Formulary in an SGML repository on a VAX cluster. The BNF is basically a compendium of the medicines prescribable by the NHS in the UK (and other non-NHS drugs nevertheless legal to prescribe under the Medicines Act) -- some thousands of drugs, their dosage, contraindications, interactions, and a bunch of other notes, cross-referenced and indexed six ways from Sunday. In 1988 it ran to about 600 pages; a more recent copy I bought in 2018 is over 1300 pages long, with two columns of fine print per page. And the paperback is updated with a new edition every six months. (These days pharmacies all access it online: but they still sell print copies sized so they're just small enough for a medic to jam in a lab coat pocket.)

NB: Yes, SGML is the all-singing all-dancing document representation, but there's a reason I write novels using Markdown (or a tool that creates a hierarchy of simple RTF documents, namely Scrivener). I don't need the complex semantic metadata! Also, it melts the brain.

Update on XML: XML doesn't have any Javascript affordance (although Javascript has slithered into the PDF and ePub 3 ebook formats). And the proposed XLink hyperlink capability never got adopted. (Source: Tim Bray, who co-authored XML.)

Shudder. Why I'm not looking forward to going back to work on Monday (I've been off for 2 weeks). We got lumbered with a data dump from the NHS which we had to check against our database to see if (a) the person is still alive and (b) is or was an active social care service user. Before I went on leave I built a semi-automatic comparator to do the matching against a look-up of surnames. Out of about 6000 rows, I matched slightly under 50%.

The NHS data was dubious: Names (Forename and/or Surname) had hospital data included Dates of birth weren't valid date formats (days and months were missing the leading 0) or were in the futurel (01/01/2999???) Postcodes were incorrectly formatted or incomplete

The initial tranche was for people most likely to be on the books; I think I'm going to be asked to do the complete list...

Oh, no, we don't want to use your data format, ours is Better (we invented it, after all), and out financial department insists we lock our users into it... until other people use it, then we'll invent a new one...

Please, I don't ever want to need to read something typeset in Fraktur.

https://en.wikipedia.org/wiki/Lisp_(programming_language)
https://www.scheme.org/

Sheesh, you computer nerds want data that last for a thousand years?

DNA.

I'll leave it to you to figure out effective coding, reading, and data redundancy strategies, since long sequences do physically break into shorter sequences over time. But if researchers can pull partial Denisovan genomes out of the dirt in Denisova cave via shotgun sequencing, coding something more bigger and more ephemeral should be doable.

Here's a Popsci explainer: https://www.youtube.com/watch?v=c2ppreiB1PQ

There is a simple answer to that.

'91-'92, I was working on a project to do with the Lake Michigan Ozone Study. We were going to tell all these sources how to format the data they were going to send to us. 100% of them said, "sorry, we don't have the budget for that. We'll send them in whatever format we have it in," which ranged from flat files to Lotus. The latter were the worst: reboot DOS computer, getting rid of ramdisk, boot Lotus, export as comma-delimited file, get out of Lotus, reboot.

Meanwhile, I had written a loader for the database in C. With the format we were told to expect. Har-de-har-har.

What saved me was... awk (a Unix scripting language). I wound up with about 30 awk scripts, which read the text files, checked all fields for validity (is it in range, etc), and wrote them out in the correct format for my loader. No eyeballing needed.

It's an easy mistake to make as few people these days are trained to distinguish them, and to that extent you make a valid point.

Which was my point. [grin]

In Canada in the 80s, you could be a brand of mini computer terminal with the US keyboard, the Canadian keyboard, and if you wanted a French keyboard.

The US one was what we Mericans used. The French one was meant for the France market in Europe.

The Canadian one was the one Canadians of all stripes wanted as it had the French Cedilla letter where the Canadian's expected it to be located on the keyboard. The US keyboard didn't have it at all and the "French" keyboard had it in the "wrong" places. Lots of keyboard swaps as people kept ordering the wrong models.

anonemouse @ 166:

"Guy Rixon's post demonstrates the matter pretty well: much of the reason the library can succeed in being a useful repository for 800 years is that the data format and the equipment required to read it have stayed the same for 800 years."

If I give you a piece of 800-year-old writing, do you think you could read it*?

*Offer not valid to people who've learned Middle English and Middle English orthography, since you're exactly my point: you've been taught to read the data format.

Probably not.

OTOH, I don't have to figure out WHAT the data format is before I can extract it to hard copy, so half the struggle is already accomplished. Even if I can't read it myself, I'm pretty sure I'd be able to find someone who CAN read that "data format" & translate it into something I can read.

And if I DO know what "data format" it's written in I can probably find a recent (last 100 years) translation already available on Big River ...

Peter Watts mentioned that sort of thing as an aside in Echopraxia. Biologist type reminiscing about the days when you could discover novel gene sequences that didn't turn out to be plans for someone's municipal sewer system.

Oh dear, oh dearie me ...
David L
I have a massive 88-ytear old tome, lavishly illustrated for the period ...

"Hundert Jahre Duetsche Eisenbahnen" ... all printed in Fractur, of course.

Um, err ...

I still write in Italic cursive, as carefully taught by my then "art" teacher, back in 1957/8, though my "hand" is nowhere as really neat as it once was - the joints are stiffening ...

Howard NYC @ 170:

still having flashbacks to the nightmare of users complaining their keyboard was missing an "any" key

this, due to so much software in both MS-DOS (1980s) & WIN (1990s onwards) displaying the prompt "press any key to continue"

During my short stint on the phones in the Help Center for a major computer manufacturer, the quickest, easiest answer was the "any" key is the big long one in the middle of the keyboard's lower edge. My experience was even the WORST, most clueless callers were perfectly happy with that answer and didn't need to be told more than once.

helpful hint:

look for 02/029/xx in non-leap years

look for 09/31/xx in all years

as to last names, mine is ten letters long and has been misspelled in at least fifty-seven differing ways... I stopped tracking simply because it stopped me from being happy in the moment {g}

you should try for phone numbers... no surprise those are never assigned twice... not until someone surrenders a number and the telecom re-assigns...

David L @ 171:

No need to go back 500+ year. The US founding documents are hard to read in the original images for most English speakers. Even in the US.

All kinds of typography issues plus things like what we call "f" today being used for things where we use an "s" today.

I think y'all are missing the point. As hard as it may be to read old written texts - whether due to typography or language drift issues - you don't need to first figure out and "recreate" the hardware & operating environment before you can extract the text from storage ...

Then how about cursive? Do they still teach it in the UK and other English speaking places? Apparently "kids today" are not being taught it in the US much anymore so they can't read grandma's letters.

They weren't being taught "handwriting" even when I was in school. Cursive writing was still taught, but the techniques of letter formation that make cursive legible had already gone by the wayside ...

I think it's mostly because legible cursive writing takes LOTS of REPEAT practice and the schools just don't have enough time to devote to that practice nowadays. They certainly didn't have time when I was in school and as far as I know the teaching burden is orders of magnitude greater today.

I learned cursive in the second grade, and by the third or fourth, my handwriting had become an impenetrable scrawl (that I can't read even one day after I've written it) ... I reverted to printing everything I write ... especially after I got into High School & took drafting classes and learned to print neat, even ALL CAPS very quickly.

Teach kids to write in a way they're going to be able to read later ...

Then there is a 5 generation genealogy done by my mother in law in her pre-teens or early teens in Germany. It was a thing to everyone to do such. Back when. Written in High German (Hochdeutsch) and using a very ornate script (I forget what it is call), well, finding someone to interpret it is hard. Even amongst Germans. Most of those who can read it were born before the 50s/60s.

Still, if it is valuable enough to you, I bet you can find someone in the Languages Dept at a nearby university (maybe even NCSU, absolutely UNC) who can not only read it, but can translate it into modern German (or English) ...

029 ==> 29 & 30 & 31

David L @ 172:

"this, due to so much software in both MS-DOS (1980s) & WIN (1990s onwards) displaying the prompt "press any key to continue""

In the 70s and 80s offices getting computerized for the first time, at least in the US, got very very very upset when it was explained that no the lower case "l" can not be typed when you need to enter the number "1". Many low to mid range typewriters didn't even have a separate key for the number "1" and thus all kinds of muscle memory generated errors non stop with people trying to enter numerical fields.

After all they'd been doing it with the letter "l" key on their typewriter since Moses. Or so you'd think from the dressing down you'd get when caught up in the complaining.

Yet, anyone doing any real "number crunching" BEFORE the office PC came around was using a 10-key adding machine ... which is why the extended keyboard with a number pad became so popular.

The person using the computer didn't have to keep two separate machines (typewriter & adding machine) and then have to manually integrate two functions.

FWIW, my old IBM Selectric typewriter DID have separate keys for '1' & 'l' ...

I learned to type on an old Remington that didn't have any letters or numbers on the keys. Had to memorize where things were located without looking at the keys or at the output. I think having the Selectric with the additional keys eased my transition to computer keyboards.

I think computers have made me a lazy typist because nowadays I DO look at the output and "auto-correct" any errors as I go along (i.e. backspace & retype as I go along).

HowardNYC: MaddyE specified going back to work in the NHS. Which means your helpful date search advice will not work because they are almost certainly not dealing with American-formatted dates: it'll be either DD/MM/YYYY or ISO standard YYYY/MM/DD. Nobody else uses the American nonsense.

I've no idea what o/s Maddy's using at work, but if they're interested, I'd be willing to write an awk script to help (and of course it's available on Windows, as well as Android, and Linux, and...

OTOH, I don't have to figure out WHAT the data format is before I can extract it to hard copy, so half the struggle is already accomplished. Even if I can't read it myself, I'm pretty sure I'd be able to find someone who CAN read that "data format" & translate it into something I can read. And if I DO know what "data format" it's written in I can probably find a recent (last 100 years) translation already available on Big River ...

You have confidently and firmly missed my point entirely.

"Text on paper" is a storage technology; knowing the data format involves knowing the writing system and language it's written in, which - if you're not an expert - can be completely a mystery. It's not like manuscripts come with a universally understandable note about what language they're in!

If I give you a piece of 800-year-old writing, do you think you could read it*?

Several scripts of that approximate age exist that no living person can read. So the trivial answer is: no.

Even if we exclude the deliberate Spanish destruction of quipu, there's the Rapa Nui script to consider. Plus the even more recent invented probably-a-language Voynich manuscript.

Robert Prior @ 176:

"IIRC the spec for a generic DTD/SGML parser using DSSSL pretty much mandates a lisp or scheme interpreter"

IIRC = If I Recall Correctly. (I assume?)

No idea what the other acronyms mean. Or why a speech impediment would be necessary :-)

I didn't know either, but I did recognize it had something to do with "data standards"

"According to Google":

Document Type Definition/Standard Generalized Markup Language ... Document Style Semantics and Specification Language ... apparently ISO (International Organization for Standardization [and why isn't the TLA for that IOS?]) standards for document stored electronically.

Lisp should have been capitalized to differentiate the programming language (LISt Processor), or more correctly LISP ...

According to Wikipedia "Scheme is a dialect of the Lisp family of programming languages created during the 1970s at the MIT Computer Science and Artificial Intelligence Laboratory" ... 😏

198 - Maddy, do you have patient CHI number as a field in your table(s)? If so, that should be a correct short form DOB, for someone born today 070324, followed by a 4 digit unique number. As implied it's even already 0 (zero) stuffed for values 00 to 09, so that First January Two Thousand starts 010100.
And you can extract DOB with =LEFT({that column},6). I'll leave you to sort out SUBSTR if you want day, month and year in separate columns.

202 - Charlie, the UK NHS uses short form UK, that's DDMMYY, not long form or USian.

202 - Charlie, the UK NHS uses short form UK, that's DDMMYY, not long form or USian.

Ah yes, being unable to handle centenarians is so much less of a problem now that COVID19 has been encouraged to thin the herd ...

It's not like manuscripts come with a universally understandable note about what language they're in!

This is where it helps to have access to the NSA's bible archive! (Warning: google/internet search is useless for locating this, because "bible" is contaminated by far too many jeezemoid sources.)

Basically the NSA used to (or still does) grab a copy of the Bible in every known language, to use as a Rosetta Stone. And there are a lot of translations, because the first thing a Christian missionary does on discovering a hitherto unknown tribe somewhere is to learn enough of their language to translate the Bible, so they can evangelize the heathen.

Per Bamford's book "The Puzzle Palace" they had over a thousand Bibles in their archive ... and that was in the early 1990s.

Going back to the original topic, a friend claims that someone has proposed a "Dune Experience" ...

... a warehouse with a pile of sand containing a couple of earthworms.

Search for:

Falsehoods Programmers Believe About Names Falsehoods Programmers Believe About Addresses Falsehoods Programmers Believe About Time Falsehoods Programmers Believe About Time Zones Falsehoods Programmers Believe About Phone Numbers Falsehoods Programmers Believe About Gender Falsehoods Programmers Believe About Geography Falsehoods Programmers Believe About Email

(I'm not sure which are the definitive versions, so no links, I'm afraid.)

The way I heard it, the more commonly known acronym expansion (Generalized Markup Language) was produced as a back-formation after the Goldfarb-Mosher-Lorie project had already yielded GML. Not sure it's the truth, but a marginally more interesting story.

XML was an attempt to produce a variant for which single-pass parsers could be written, for the Web. No (indeed, bletcherous) Javascript involved.

I don't have to figure out WHAT the data format is

Well, no. We define "recorded history" such that if you don't immediately recognise it then it's not a record of history and thus the primitive savages using it can be exterminated to make way for superior races. Hooray for progress, glory to the empire!

Meanwhile we have such things as the stone currency of Yap, the many whistled "spoken" languages, and written forms from hieroglyphics to morse code suggesting that humans have an amazing ability to come up with diverse solutions to problems. Not to mention arguments about "what is agriculture" and "what is language" not to mention the joke about "I don't know what that's for, must be a ritual purpose" that complicate the question even further. If you don't realise that it's a language, how could you possibly look for a written representation of it?

So, for example, if orca use carefully arranged seal bones stuck in the sand as a form of vagrant sign, who would notice and how would they work out what those mean*.

https://www.bbc.com/travel/article/20180502-the-tiny-island-with-human-sized-money

(* a special shout-out to the geeks so desperate for new project names that almost any combination of search terms leads to a recent computer-based project rather than a historical explanation)

anonemouse @ 204:

" OTOH, I don't have to figure out WHAT the data format is before I can extract it to hard copy, so half the struggle is already accomplished. Even if I can't read it myself, I'm pretty sure I'd be able to find someone who CAN read that "data format" & translate it into something I can read. And if I DO know what "data format" it's written in I can probably find a recent (last 100 years) translation already available on Big River ..."

You have confidently and firmly missed my point entirely.

"Text on paper" is a storage technology; knowing the data format involves knowing the writing system and language it's written in, which - if you're not an expert - can be completely a mystery. It's not like manuscripts come with a universally understandable note about what language they're in!

And you might have missed my point as well.

"Text on paper" doesn't require me to figure out what the different "bits in storage medium" are [and how to extract them] before I can even produce "text on paper" ...

I can't even see the bits, but I can see text once it's on paper.

I recognize quite a few languages¹ I don't read, write or speak ... If I've got "text on paper" I've already got a head start figuring out the language (... and finding a translation I can read).

So no matter how obscure a text might be, it's still more accessible than "bits in storage medium" where I don't even know how the bits are assembled into data.

And maybe nobody knows because the machine is so obsolete no one who ever used it is still alive today? Text is just easier to decipher.

¹ "Middle English" being a bit more instantly recognizable than "Olde English" ... being that by "Middle English" a lot of modern words have crept into the language.

... a warehouse with a pile of sand containing a couple of earthworms.

They need to be what we call night crawlers. 20cm or so long. So big ones to similar the ones in the movie. Well to the standard of the Wonka experience.

But no refreshments allowed. Just a catheter from there to your mouth. You know, for the TRUE experience.

Charlie Stross @208:

"202 - Charlie, the UK NHS uses short form UK, that's DDMMYY, not long form or USian."

Ah yes, being unable to handle centenarians is so much less of a problem now that COVID19 has been encouraged to thin the herd ...

For personal use on the computer I find the format "YYYYMMDD" easiest, because even Windoze can sort the date correctly.

For writen dates, I prefer DD Mmm YYYY - for me, today is 07 Mar 2024**. If I get my camera out, any photos I take will be stored in the folder "20240307-whatwasItakingphotos_of"

** Or more precisely 16:54R07Mar2024

At the moment, we're on Win 10 until the new laptops arrive. As it happens, I work in local government in Adult Social Care, so we have a lot of dealings with badly formatted NHS data. What do you mean we can't put the ward name in one or both name fields - the delivery driver knows what it means? (At least there's no warnings about exotic pets in the name fields in this dataset.) Why is that hyphen necessary??? What's wrong with missing out 0 from days and months - it's still a legible date... The system being used for these Occupational Therapy equipment orders doesn't appear to have heard of data validation.

Basically, the data dump is in Excel, I clean it up, deal with multiple variants of multi-word surnames (with or without hyphens). Out of that, I pull a list of unique surnames. Those get input into Business Objects which then extracts a list of possible matches in our database and creates 7 different look-up values using combinations (at least 2 or) of Forename, Surname, Date of Birth & Postcode. I then match in Excel using the same combinations. The only thing that's a niggle is that serial dates are slightly different in BO and Excel, so I have to turn the DoB into text, split it into dd, mm, and yyyy (preserving the leading zeros) and create text look-ups.

At least now I have created a template that does it mostly automatically instead of querying the database manually. The main time sink is the initial data cleansing, but that's easy enough in Excel with column filters turned on - I drop the raw data into the template, and check the calculations. If I find something weird, I fix the raw data in the template.

Thinking about it, I think the data dump isn't from the NHS spine; it's from the company providing the equipment, which explains a lot about the poor data quality.

Sadly not, that would be far too easy. It's why I think the order system isn't NHS-built.

Clive Feather @ 210:

Going back to the original topic, a friend claims that someone has proposed a "Dune Experience" ...

... a warehouse with a pile of sand containing a couple of earthworms.

You can tell your friend that at least one other person thought the joke was funny. 😉

Oh, gag. And no one in upper management cares enough to say "it SHALL be done this way". Glad you have some means of checking that is not all eyeballs, all the way.

Since the headmaster of the primary school I attended (in Kent, circa mid-60s) was the author of one of the standard textbooks for teaching cursive (we knew it as ‘italic’) writing, yes, we learned it. Interestingly he was also left-handed, so I didn’t suffer the common torture of being forced to write wrong-handed, unlike my wife (damn those nuns forever). And for more amusement he & his teacher wife spoke Welsh in classrooms when they didn’t want the kids to know what was being said. I never did tell them I understood; not that I was fluent and certainly not anymore.

As for the cursive handwriting.. well I don’t exactly write much on paper these last 40 years, so I have trouble reading anything not done with great care and slowness.

... and just to screw things up further, I learned while standardization was still an ongoing process, so my personal way of doing things doesn't conform to modern practice

... it should technically be 215420240307 (Zulu time implied, no colon between hours & minutes).

There are multiple "falsehoods programmers believe about..." collections on Github. I found a collection there years ago with over 200 of these documents.

A quick search just now produced this "curated list" (first in the results): https://github.com/kdeldycke/awesome-falsehood

Here's a thread discussing that list: https://news.ycombinator.com/item?id=24405941

TL;DR Nightmare fuel for programmers.

so we have a lot of dealings with badly formatted NHS data

You sound like me with Excel.

5+ years ago we had an office convert from an ancient in house mail system to Microsoft 365. And they wanted the managements (5 people) contacts cleaned up and merged. Which meant 15 difference sets of lists. Address books in phones and in email and that Christmas list in Excel and ....

I started with about 30K entries. Lots of sorting by various fields and moving clumps of columns around. Phone 2 used by one person to hold their child's name and such. But matching on phone numbers and emails (left side first) and such.

When done I had it down to around 5K names that were unique and had unique, well mostly unique, emails.

But there was still a lot of crap floating around. .co or .cmo or whatever that should have been .com and on and on and on. Not to mention email address or phone number changes that only got put in 1 or some of the various collections.

Then a partner died and the company split in half. Big mountains of emails to various people that were bouncing telling them about the changes.

I first did this when the BBC sent us a list of licence holders for the digital switchover way back when. Later, I improved the 'automation' when I was asked to provide a list of people who had died and had been notified via the 'Tell us once' system to flag records for closure.

It's the later techniques I've recreated for the current use-case.

I agree that data quality these days is an afterthought and nobody cares enough to enter data correctly in the first place (or is in too much of a hurry to do so).

"If I give you a piece of 800-year-old writing, do you think you could read it*?"

In what script/medium/language?

Something written in a book by monks in 1224 would be a struggle, to say the least. Something written on a gravestone by the corpse marker production operative, not so much, as long as it hadn't eroded away. There is plenty of graffiti still around from ancient Romans writing on walls which is as legible as anything done by someone whipping out his pencil as the twilight falls (and says much the same things). On the other hand the kind of large-scale multicoloured graffiti that people execute at lightning speed with a sack of different spray cans, while it is nearly always a considerable improvement on the ungraffitied appearance of whatever it's on, nevertheless annoys me because despite its use of Roman alphabet letters it's still impossible to work out what it says. I can't read any variety or age of Chinese characters at all, but I have still once or twice found the Chinese side of an instruction sheet for some gadget more informative than the English side.

But that kind of problem is at a higher level than what I was talking about. The data format in all those cases, in the sense I meant, is the same: visually contrasting marks which can be recognised in relation to particular pattern archetypes by the human brain. This is also the fundamental form in which any archived data must be represented in order to be any use to start with. Existing archives can maintain their usefulness for 800 years because they only have to concern themselves with the physical preservation of the items, and as long as they can manage that the data they represent remains available.

Finding it hard to read old monky writing is just another instance of the individual need to be able to understand the kind of thing you're dealing with in general before you can understand a particular example of it. It's the same thing as not being able to understand a modern paper written in Chinese because you don't understand Chinese, or not being able to understand a paper about boosting the body's response to HIV because you don't understand advanced immunology.

Computerised archive storage brings in a whole new extra set of levels that have to be negotiated before you even get to this point, which are themselves just as much of a specialism as advanced immunology and are full of obscure wrinkles that only one or two people know about. Any particular behaviour, even if it looks simple, depends on an enormous number of intertwined strands once you come to dig into it, and they go right down into the abyssal levels where people swimming at lesser depths, if they are aware of them at all, take them so much for granted that it never occurs to them that they are not actually eternal truths. Like paws's example of an architecture that stores words inside out. It's at a low enough level that even assembler programmers won't notice it most of the time, and it will never cross the mind of anyone just using the system at an ordinary user level. Since everything you use on the system is written and compiled on that system, it all works, and continues to work as different bits are updated, so everything is fine. But when years later you try and read the data on a system that stores its words differently, everything is screwed, and it needs a specialist in unscrewing that particular kind of lesion to come in before the specialists in whatever the data is all about can start to use it. Which is like people new in a department of a paper office needing to read old documents and discovering that the thing the previous occupants thought was a long term storage box was actually a shredder.

Hah. I thought computer character sets were great, because finally I could actually type a 1. l never looks right even with serifs. I looks worse. And I could also type a 0. And a !.

In what script/medium/language?

Bingo. :-)

I seem to remember someone suggesting that digital data be archived as bitmaps printed on acid-free paper at some point, though I can't find it again now. Presumably this was for maximum inconvenience on every axis?

So are your comparing your zeroz and letter "ooohs".

Which are narrow and which are wide. Does one or the other have a slash?

I've seen all 4 combinations within an hour on paper.

And those stupid 30 to 50 character license keys that for security reasons you can't paste into the field. If possible I tend to copy them into a text editor, blow up the size, then make sure I'm on a font where zeros and ooohs and the letter ell and the number one are all easy to tell apart.

Something written in a book by monks in 1224 would be a struggle, to say the least

Someone on Reddit recently put up a picture of a very ornate letter. And asked if it was a "B" or an "E". The consensus seemed to be "E". I could have gone with either answer. But there were subtle clues to people who worked with such things.

ISO8601 starts from years and works its way down to milliseconds but it can be truncated at any interval. 2024 is a perfectly compliant ISO8601 date specifying a year, as is 2024-03-07 specifying a day. 2024-03-07 23:30 specifies a time to plus or minus a minute and so on. The punctuation can be omitted but it does make it easier for meatbags to read at a glance and provides a certain level of verification and error checking.

It's also useful for sorting by time. (Sorting with more general text can get complicated, however. Unicode makes a total mess of it.)

So I use ISO8601 dates whenever I'm given a choice.

"Then how about cursive? Do they still teach it in the UK and other English speaking places? Apparently "kids today" are not being taught it in the US much anymore so they can't read grandma's letters."

I learned that as a matter of course at my first school. I also learned to call it "joined-up".

At my second school other kids saw me doing it and went "Ummmmm. Look, he writes double. I'm telling." The teacher didn't really object, but still seemed to be a bit put out.

At my third school it was expected that we had already learnt it; writing "double" was normal, and people who wrote "single" must be thick. If you had really neat writing you were even allowed to write in ink.

At my fourth school nobody knew the expressions "single" and "double". If we called it anything, we said "joined-up", but it was so universal that there was rarely a need to distinguish. And everyone wrote in ink.

At my fifth school some people started to write in biro. Gradually other people would notice that they were doing it all the time and consistently getting away with it, until eventually nearly everyone did it. Some especially daring iconoclasts even started writing in felt pen.

These days there seem to be a disturbing number of adult people even in responsible positions who don't merely write single, they write baby writing - the simple, unformed, rounded script that people produce at the stage of having only learned to copy the maximally-simplified letter forms the teacher wrote on the board, before they have developed any individual style of their own. They haven't even learned not to mix up big and little letters yet. And even more than was the case with the class at that stage of learning, they all write the same. They're still all trying only to produce direct copies of the teacher's blackboard style, except that now they're grown up they can copy it more closely.

Not that they're much better using a keyboard. And nobody even stops them getting away with it. I've had letters and emails from customer service departments which, although the computer ensures that they use correct spelling, are still semantic soup. The management don't seem to realise what a bunch of incompetent shits it makes their company appear - but then how would they, when they do no better themselves.

underlying notion can be applied to:

YYYYMMDD

MM/DD/YY

DDMMYY

people will mistype values which any well written code ought trap before proceeding further

problem being so little code is as well written as we need it to be

would you like to guess how many entries of dates include howlers such as 31-SEP-2018?

my personal favorite was the developer on a Y2K remediation project at REDACTED BANK in late 1900s who refused to understand that 29-FEB-1900 being invalid but 29-FEB-2000 is valid and 29-FEB-2100 is also invalid

no actual bloodshed but my refusal to back down got me fired at 2:00PM and rehired at 3:00PM as I was being escorted off site

regarding your

why isn't the TLA for that IOS?

blame the French

I've made up my mind so stop trying to confuse me with additional facts {G}

besides... most such posts are hardly distilled down into pure wisdom...

do you know how it hurts my eyeballs when ever I see "ill" as the first word in a sentence on web pages using 'elegant fonts' that turn it into "lll" rather than "Ill"?

plenty of days I loathe those obsessive bastards who've created a zillion fonts for no better reason than "why not inflict blurriness upon the masses?"

One of the problems with sorting text is homoglyphs leading to ambiguity in sort order. If you don't know which language you're sorting it's hard to do it correctly (despite the pretentions of some here). There are likely multiple sort orders for a given language based on characteristics of the data (people's full names may sort differently from partial names, street addresses may sort sans numbers etc that precede them (it's the etc that will get you), there may be gender-based rules that rely on external data, and so on).

please note the following will be a rather shallow reason to get drunk

so plan ahead and stockpile low grade booze

2026-06-29 Mon ==> Day 100000 of the UK Gregorian Calendar

Also, with data putrification never underestimate human creativity. You may think that you can correct "Wahsington" to "Washington" but then you discover that wellactually the former is a town in Durango and now what do you do?

Having to manually fix an address database has traumatised me forever. As with Shakspeer, not everyone who roamed Australia bestowing their names and the names of their associates on everything they saw had a consistent way to spell their name. We also have a number of similarly named individuals doing that, so you get Smith's Creek, Smyth's Creek, Smythe's Creek and Schitt's Creek all in the same area. Or not, depending on which exact Smith's Creek you're referring to.

Not to mention the updates to new revised improved updated spelling v2.32.18b so that now "Wanganui" is spelled "Whanganui" and every paper record ever was wrong (or not, depending).

The US National Archives are looking for volunteer cursive readers to transcribe Revolutionary War pension documents:

https://www.archives.gov/citizen-archivist/missions/revolutionary-war-pension-files

In German there's always the fun question of whether you sort the letters with umlauts with the "base" letter (e.g., ä with a), or after the z in the alphabet. Ditto questions about where to sort the ß. And there are some Germans who take pride in not sorting those letters in the same way as is recommended for the particular application.

"At least there's no warnings about exotic pets in the name fields in this dataset."

If you encounter a name field containing only the single word Pigeon, it isn't an error, it's me.

You will also likely find other blank fields, eg. "title", "date of birth", "phone number". Or the numeric entries especially may be invalid ones that the software isn't clever enough to catch, or just plain random. It all depends firstly on whether, having bypassed the client side validation entirely, I found enough holes on the server side to be usable, and secondly whether having failed to find such holes, I cared enough to apply for whatever it was by post instead of by internet so I could leave actual blanks, or was content just to fill in rubbish.

The point being that ad-hoc database fettling hacks will encounter not only genuine errors/carelessness, but also the consequences of people disagreeing with you over what counts as a correct entry and/or whether particular entries should exist at all. "You" does not mean you personally, since I consider my medical records to be one of the very, very few cases that actually matter, so I allow slack to the NHS over this. But it still requires a conscious effort to remember I'm talking to a medical person and still give them actual answers if the question itself isn't directly medical, instead of automatically spouting the first random rubbish that comes into my head as the modern explosion of computerised nosey parkers has trained me to do.

Which brings us back somewhere adjacent to the original topic, since the automated and indiscriminate scarfing of data followed by its unsupervised and unchecked barfing under a veil of apparent plausibility creates more new possibilities for bad things to result than I care to think about.

"the updates to new revised improved updated spelling v2.32.18b"

And new revised improved updated transliteration conventions. Which add extra confusion by seemingly always being revised away from looking like they ought to sound at least vaguely like what the word actually does sound like, and towards some abstract notion of "correctness" that has nothing to do with pronunciation - an idea that gets dafter and dafter as the source language's script gets further and further away from the English model of assembling words from an alphabet of letters with vowels and consonants.

800 years ago, I wouldn't know..

80 years ago, I could maybe work it out with some contemporary documentation

8 years ago, bitmaps could be recreated with some paints and i's and o's

8 days ago, I can't even tell who's speaking the same language

https://brill.com/view/journals/nun/39/1/article-p163_9.xml

And another version: in the late 1970's I was involved on the periphery of a data conversion project using paper based (card) records from a couple of decades earlier. They wanted the information loaded onto "mag tape" as the boxes of cards (and contents) got damaged over time when processed too regularly...

Our local NZ "Ministry of XXXX's National XXXX Statistics Centre" had the results of a number of decades of statistical data of a technical nature stored on punched cards that we were converting to mag tape to make them more "usable" for ongoing statistical analysis purposes. Literally millions of them.

Yes, the project did have a card reader that worked quite well. Problems encountered included the format of the cards (and column positions) changing from year to year, no characters written on the tops of the cards, the sticky labels on the boxes had frequently fallen off as the glue dried and failed, and things like disease codes (ICD?) changed on a regular basis. And some boxes the cards were apparently infected with a variety of wood worm that apparently liked chewing on cardboard. And if the cards got bent/folded/spindled/mutilated (or wet) - frequently by the box containing them getting damaged - the card reader couldn't/wouldn't reliably read them.

And they were converted/written to half inch 1,600 BPI 12" tape reels using a DECsystem-10 computer's tape drives. Which almost certainly no-longer exist and hence not more readable (or correct) today?

Nojay @ 231:

ISO8601 starts from years and works its way down to milliseconds but it can be truncated at any interval. 2024 is a perfectly compliant ISO8601 date specifying a year, as is 2024-03-07 specifying a day. 2024-03-07 23:30 specifies a time to plus or minus a minute and so on. The punctuation can be omitted but it does make it easier for meatbags to read at a glance and provides a certain level of verification and error checking.

Yeah, I don't know if my way of doing it meets ISO standards, but it allows the computer to sort the folders in date order so I can find the photos I'm looking for.

The other bit was just how I had to hand-write time & date (and other stuff) in message traffic when I was in the Army and it was my turn to monitor the radio ... format the message so anyone else who had to read what I copied won't have to waste time trying to figure out if that's the number '2' or the letter 'Z', number zero or letter 'O' ... that sort of stuff.

And were times & dates given in the message Local Time (Romeo or Charlie) or UTC (Zulu)?

As noted I think the NATO standard changed¹ since I retired, but I'm no longer in the Army so I don't have to conform to current standards when I'm writing notes for myself ... I just have to use a format I'll be able to understand when I run across the note again at some later date.

¹ I believe ALL times/dates are currently given in UTC (Zulu) and you have to calculate the offset to local time locally ...

And they were converted/written to half inch 1,600 BPI 12" tape reels using a DECsystem-10 computer's tape drives. Which almost certainly no-longer exist and hence not more readable (or correct) today?

NASA might be able to help them.

They started on dealing with such things over 20 years ago. Someone or ones starting noticing they had all kinds of sensor data from various missions that was on a pile of mag tape formats. And a non trivial number on tape formats created by wizards for the mission. They started a project to move the data off mag tap to actual disk storage.

Many times it involved finding old computers in various surplus places or in non first world countries in Africa, the former Soviet Block or similar where they found them still in use. Or recent use.

The hardest were the custom built for a mission or few tape drives where they couldn't figure out from the extant engineering drawings just how they worked and so went door knocking to see who was still alive and had a memory of the details of a project they were on a few decades earlier.

The US National Archives are looking for volunteer cursive readers to transcribe Revolutionary War pension documents:

I can read most cursive written by people in this or the previous century. Prior to that many examples I've seen get, well, interesting. Think of those monks and the illustrative ways they did things at times. And the ff/ss things discussed up thread.

I pulled up some microfiche of census records (USA) hand written about my family. This was around 1970 reading records 100+ years old. Penmanship was apparently not one of the most important qualifications of the folks who rode horseback to farms and small communities in the rural US taking down information.

AIUI, the Mormons have converted most of those into a data base you can search. But there are warnings about possible errors. And I've bumped into a few when searching their records.

There are likely multiple sort orders for a given language based on characteristics of the data (people's full names may sort differently from partial names, street addresses may sort sans numbers etc that precede them (it's the etc that will get you), there may be gender-based rules that rely on external data, and so on).

Then we get into Chinese, Japanese, and similar.

Anyone her know about the current debate about naming newborns in Japan? Apparently there are cultural rules. And for a while now some new parents aren't following them. And as the kids grow up not all places that require a written name are going along with what is on the birth records.

Or something like this.

218 - "Not NHS built"; In that case I feel your pain and sympathise, but can't help further. :-(

233 - I feel your pain also. In some cases I've responded to this sort of unpunctuated stream of conciousness with "Could you write that again, but in English this time please?"
A likely response to me will be (as they type it) "you fucking understood me".
A third party with then reply with something like "No we didn't actually."

243 - Similarly since both "Paws4Thot" and my birth certificate name with the last 2 characters substituted with "11" (eleven) are usually accepted by systems that demand that "you have a number in your username".

Aotearoa has the fun law that your legal name is the name you are known by, but there are rules for legal names. So you get situations where, say, someone's police record has to be printed out and umlaut's manually drawn onto the "alias" field entries before it is accurate, because "the computer says no umlauts" but Geörge has an umlaut and no two ways about it. Some computers are smart enough to say "the preferred name field is not subject to the Law Of Official Names" and others notsomuch.

This become an issue of racial tension when the macron is not permitted and those who would use it decline to use double vowels. The new gubbermunt has solved this problem by decreeing that official documents shall not use marry words. I don't know what their solution is where there's no commonly used engrish word. Like Taupo or Rotorua.

There was a joke of a book in the early 80s in the US called something like "The Yuppie's Handbook" or similar.

It had a chapter on names. Many young hip couples would name their new child something like Tiffany Amber Smith and get legal docs about it. Then a week later at the Christening prep the priest refuse to allow a name where at least one of the bits wasn't from the Bible.

Oops. Relatives in best clothes and such on the way in from far away for the service tomorrow. And many of the older ones if total agreement with the priest.

Surely the solution is to give the child a christian name that is different from their legal name? Much as you can be religiously married without bothering the government about it (or vice versa if you prefer vice)

I'm not in these faiths. But I think there is/was a strong tradition that your name in both realms match. No pretending to be of the faith and such.

Legally the name was fine. But to the "family" not being sprinkled soon after brith was a very very big deal. And for many still would be in terms of family relationships.

But to the "family" not being sprinkled soon after brith was a very very big deal. And for many still would be in terms of family relationships.

This could also be a source of (in my opinion) unnecessary strife when the person decides to change their name later in life. Either the real, daily name, or the official one. It seems there are a lot of people for whom the name of a person is basically set in stone when first given and it cannot be changed later, especially not by the person in question.

Which seems to me to be a bit... inflexible. Nobody has greater authority of what their name should be except the person in question (well, states often have some restrictions), and not trying to use the name people want to be used of them seems quite impolite, to say the least. Sometimes it feels like maliciousness when people use a name which the person in question does not want to be used.

Also people often do have many names, depending on the situation and circles, and time. I haven't changed my official name as I'm quite fond of it and comfortable with it, but I have had many nicknames and obviously I'm not called by my full name basically ever. 'Mikko' is quite a common first name in my age cohort so most of us have been called by surname or various nicknames.

I used to be called 'Jay' in some contexts in real life, too. This was the nickname of an MMORPG character I played for years and our player collection was very international. In live meetings it was easier to just use the character names and I still know many people only by their in-game name. The name has no connection to my 'real life' but I still got occasionally confused when some other people had that as their nickname.

Data point from England: out of myself, spouse, and six children, NONE of us have the same name on birth certificate and passport.

Another source of confusion: if you buy a railway ticket for use today in Britain, it will say "MCH" and not "MAR".

You'd be amazed how many forgeries that catches.

(Only five months - JNR, FBY, MCH, JLY, and DMR - have non-standard abbreviations. Wikipedia claims it's to increase the Hamming distance.)

I am not sure how they solved that problem. I got a different job in mid 1980's and lost track with that organisation's data processing division. But being a government organisation, I am sure they did try and solve it.

Certainly by late 1980's/early 1990's, those "reel to reel" tapes (and associated tape drives) were becoming increasing rare as they no longer had the capacity to backup the increasingly large disk drives being used. There was a massive move to "cartridge" style tapes and "robot" style tape silos. And while I am no longer up to date with he stuff, I understand the technology is continuing to evolve (or tending to get replaced by cloud based storage).

Going back to the original topic, a friend claims that someone has proposed a "Dune Experience" ...

... a warehouse with a pile of sand containing a couple of earthworms.

There's a photo meme going around to this effect, of a pile of sand in a warehouse.

Ah, Google finds it on Twitter now. Picture alone here, not embedded because it's big and the html scaling commands aren't getting parsed.

would you like to guess how many entries of dates include howlers such as 31-SEP-2018?

That's a perfectly valid date order outside the USA.

It's not in ISO order so it's not trivially sortable but it's absolutely no worse than, say, SEP-31-2018.

I think there might be a larger problem you're missing:
"30 days has September,
April, June, and November..."

It helps a lot if the text is in the language preferred for research publication across the whole planet. Currently English or American, previous Latin or Greek: all remain readable. Also if translations from sample works in that language are everywhere.

To be more time-proof, a repository could include sample translations from its language of choice on non-digital media. They could engrave them on the foundation stones of the building housing the repository.

More seriously, a digital repository is never likely to work when read by archeologists, not for any kind of storage with useful density. Such a repository is only supposed to work while its host institution continues. Translation aids for the languages (and scripts) of choice can go inside the repository.

You may think that you can correct "Wahsington" to "Washington" but then you discover that wellactually the former is a town in Durango and now what do you do?

There's a town in California named Eureka - and another one named Yreka. What happens when some dolt sends the system a package for Urika, and how should we plan for it?

(BTW, speaking of California, in the San Francisco Bay Area there's a neighborhood called Burbank, about two miles from where the San Jose WorldCon happened. It is not the same Burbank that's in Los Angeles. How does one know the difference? Context, hopefully.)

In Japan, there're apparently two ski centres, named 'Hakuba' and 'Shirouma'. The fun thing is that they are both written '白馬', meaning 'white horse'. One of them is read with on yomi and the other with kun yomi, basically with either 'Chinese' or 'Japanese' pronunciation.

This of course creates occasional problems.

Yeahbut...one can sequence old DNA but not recover enough of the detail to clone the organism (IIUC). Very lossy.

A while back, I met a researcher who was trying to do spectroscopy on freshly-harvested DNA. Apparently it's chemically extremely fragile when taken out of a cell and degrades if you look at it harshly. The housing for the sample to get it into the spectrograph (analogous to the slide in microscopy) had to be made of gold to slow down the degradation.

I'm now imagining a repository as a zoo of critters with the work encoded in their DNA...aaaand then they reproduce and all the data gets jumbled up together. Not unlike the output of an IA.

If you encounter a name field containing only the single word Pigeon, it isn't an error, it's me.

You don't go far enough.

Back when Thatcher's government was bringing in the odious Poll Tax, an acquaintance of mine -- today he'd be termed a white-hat hacker -- took exception to it and changed his name by deed poll to address the matter.

Due to a peculiarity of English law, a Magistrate's Court summons for non-payment of the poll tax had to be addressed to the person concerned using their full legal name.

After filing his deed poll, his name was mathew.

Note: no surname/first name, just "mathew". Not "Mathew". Nor "Matthew" (the normal spelling). Not "mathew M. mathew" (database demanded a middle initial or middle name in the field).

He got summonsed on multiple occasions and had it chucked out of court every time.

Basically he found a name that choked all the input validators.

These days we are wise to falsehoods programmers believe about names, or at least wise enough to have no excuse. But this was back in the 1980s and nobody liked the poll tax much (including the developers tasked with writing the software to issue the writs).

Dates in Excel are just a number. If a column is formatted as date the display and entry format depends on the individual users PC settings. I once had a complaint from an endocrinology professor that the clinical trial data I sent him used US dates. I showed him that the dates on my PC were the correct UK format and how to use the international' settings to his chosen format. And date entry can use either- or / as a separator. My problem with Excel dates was that some people enter them as text.

Just need a warehouse filled with sand, an arc light for the sun and a large supply of DMT. People will see sandworms soon enough.

Among my various hats is data custodian for a highly respected research group in Norway.

In addition to all of the problems noted above, there's one no-one noted - in much of Europe, it is common to denote decimals with a comma, not a dot ("decimal comma"). So the value one-half is written as "0,5". If thousand separators are used, those use a point. So the value "one million and one half" is written 1.000.000,5

Excel and the like will handle this kind of thing based only on the host computer's localisation. So good luck if you happen to prefer the normal way of doing things, because suddenly Excel sheets will interpret those fields as strings, not numerics. (In fact, good luck if you use Excel at all: if you're not doing accountancy, you bloody well shouldn't, says I. But a lowly no-longer-a-researcher like myself doesn't get to enforce standards on people).

Plus, of course, people have discovered the wonder of CSV for exchanging tables (it's, well, better than .mat files. Shudder.). Guess how well that handles decimal commas?

Then, of course, you wind up with collaborations between researchers. And you get data files that randomly mix decimal commas and decimal points and, occasionally, thousand separators too.

Fortunately, it's a relatively rare phenomenon, because if you're using any tool other than Excel, it rapidly becomes obvious that the extra logic to handle decimal-comma-numerics and convert them to an actual numeric type is just not worth the faff, even if you prefer to use your own local sstandard. But it's not sufficiently rare that it can be ignored - everything needs to be checked and cleaned.

The other problem, which continues to vex me every few weeks, is that all too often, one encounters the more fundamental problem that one can't even find the relevant data - I get to answer all of the requests for "we'd like to see the data for paper 'X', please", and since the group has published a lot of high impact papers over the years, many of these requests are for data from long ago (I've had at least one request for data generated while I was still at primary school).

While I have archival access to the complete network spaces of long-past users, it is often a challenge to even find where the data files are, never mind what format they are stored in, or identifying the relevant experimental sessions - while researchers are trained in many skills, data management is not, and never has been, amongst them.

While I generally don't include these sort of details in the data responses, a few reasons for incomplete or empty responses are below

• Here is all the data, and a copy of the program you will need to interpret them. You will need a virtual machine running Windows 3.11 or earlier to install it.
• That data appears to have been lost in the great data-corruption disaster of 2015, sorry
• I'm pretty sure that the data archive is complete. I know the paper referenced cell 26, but the data doesn't. It might be a typo. Maybe it was cell 16 instead?
• We can find no data associated with subject 'X', and there is no evidence it ever existed. The name may be a typo.
• As far as we can tell, the reference to "male subject X" is an error in the paper. Our best estimate is that that data came from "female subject Y" instead.
• The lab-books for that experiment were last seen before we moved out of the bomb shelter in 1998, and are most likely still in the sub-basement there. I do not have the relevant safety qualifications to continue the search.

here in the USA the tricks available are to leverage the "zipcode" (AKA postal code) and back before mobile phones (and flying cars and moonbases and other technological marvels of the 21 C) the "area code" portion of phone numbers as geo-location-validation

I've lost track (previously memorized in the dawn of time) of how many states have a urban center named: Springfield, Portland, Washington, Lincoln, Jefferson, Isaac, Abraham, Peter, Paul, et al

FUNFACT: most common street designator (AKA street name) is "Second Street" not Main Street or First Street

kid today with their rude slang {G}

"chasing kids off my lawn since I turned 75(tm)"

I heard from someone who stumbled over a cache of old magnetic tapes in a closet of [REDACTED UNIVERSITY] and she wanted to find out what was on 'em... contacted a high school friend who went career track US Navy and worked at ONI to ask if trainees in the ONI (Office of Naval Intelligence) needed something challenging

the cartons of 60 (70?) tapes was thusly dropshipped alongside a bottle of very good whiskey and without warning left upon their desks along with a note mentioning the team would be awarded the bottle upon completion of analysis... the ONI officer's intent was not fermenting silliness of competition but encouraging cooperation as a team building exercise...

took 'em nineteen days to find the necessary devices, slowly feed thru neglected/misstored tape, extract bits 'n bytes, ans then tear out their hair puzzling out whatever-it-was

...and it turned out to be student class rosters at [REDACTED] which had gone missing for more than a decade

remind me... how many days hath September?

Doesn't matter.

Remember, the data entry format is not responsible for bounds checking -- those are different levels of input sanitization!

It seems there are a lot of people for whom the name of a person is basically set in stone when first given and it cannot be changed later, especially not by the person in question.

Small data point here. My father had only a first and last name. Ditto his brothers and father. In basic training (WWII 1943) his drill instructor kept demanding he tell everyone his MIDDLE name. And kept accusing my father of having one and not wanting to give it out. So my father got extra duty (latrine cleaning and such) for the entire basic training. 2 to 3 months. As soon as he was done with basic he went to a local judge and got his third name bit added. TO THE FRONT. We never did think to ask him why he added to the front.

But he was known by his now middle name for the rest of his life.

While DMT would work better and probably be safer... It seems like a missed opportunity for "Spice" https://www.dea.gov/factsheets/spice-k2-synthetic-marijuana

There was a massive move to "cartridge" style tapes and "robot" style tape silos. And while I am no longer up to date with he stuff, I understand the technology is continuing to evolve (or tending to get replaced by cloud based storage).

There were multiple robots with various cartridge formats starting in the 70s.

LTO came out in 2000 and I think the entire planet has mostly moved to that format. I had a client using LTO-3 and later LTO-5 till we went to onsite dup with a cloud backup. In raw capacity LTO-3 was 400GB and LTO-5 1500GB. Currently it is up to LTO-9 at 18000GB. And the drives are designed to try and compress data on top of this capacity. The physics of the media is a bit of a boggle to my mind.

These cartridges are about 10cmX10cmX2cm.

You can get single drives or auto loaders. And I'm sure there are some robots based around the format.

NONE of us have the same name on birth certificate and passport.

That can be an irritation.

Mismatch your names on your passport or other ID and an airline ticket in the US some time. You'd better allow an extra hour or few at each end to be able to explain why. And if flying internationally, just go home and try again.

That's a perfectly valid date order outside the USA.

Also inside.

As a one time deep dive programmer, I HATED dates.

Monthly things that start on the 31st. Annual things that start on Feb 29. Ugh.

We also have a number of similarly named individuals doing that, so you get Smith's Creek, Smyth's Creek, Smythe's Creek and Schitt's Creek all in the same area.

In the US we can have up to 50 official duplicates, plus additional areas with historical names.

And there are a few airports near cities with duplicate names so a few times per year someone flies to the wrong side of the country and gets upset that the airline DID THIS TO THEM.

When I lived in Lexington, KY I was within an hour's drive of Paris and Versailles.

"good luck if you use Excel at all: if you're not doing accountancy, you bloody well shouldn't, says I"

IMHO if you're doing accountancy on my money, you shouldn't be using Excel either. You should be using commit/rollback operations on a proper database structured in one of those normal forms devised by Edgar Codd specifically to protect data integrity.

Excel is the Birmingham screwdriver of IT.

Anyone her know about the current debate about naming newborns in Japan? Apparently there are cultural rules. And for a while now some new parents aren't following them. And as the kids grow up not all places that require a written name are going along with what is on the birth records.

eh, i vaguely remember hearing about these names, sounds like there are limits on the kanji permitted but not on their pronunciation, so some people have been irresponsibly creative and the government have decided to call time on their activities

https://timesofindia.indiatimes.com/world/rest-of-world/unusual-names-can-complicate-life-in-japan-now-parents-are-being-reined-in/articleshow/105739790.cms (nyt version is paywalled)

they have a huge capacity for fuss here at times

»IMHO if you're doing accountancy on my money, you shouldn't be using Excel either. You should be using commit/rollback operations on a proper database structured in one of those normal forms devised by Edgar Codd specifically to protect data integrity.«

Those are not important parts of accountancy, those are mere implementation details.

The important part is the "double-entry-keeping" method, invented in Venice 700 years ago.

As inventions go, it is pretty damn important, enabling amongst other things inter-generational debt and practically eliminating trivial embezzling of money.

anonemouse @ 262:

I think there might be a larger problem you're missing:
"30 days has September,
April, June, and November..."

Oooo! I just had a thought ... leap year as a movable feast. After 29 Feb 2024 you have 32 Mar 2028 ... 31 Apr 2032 ... until it rolls around to 29 Mar again ...

Use the same formula for calculating leap years, just move the additional day around 🙃

I have a much simpler answer, the one from my youth: "Ø". (Oh, good, the cut and paste worked.) That's a slash through it, what line printers always used for zero.

Rolls eyes.
My signature is, in fact, actually readable, as opposed to many people. Those... how can you prove you signed it, when all there is to see is what looks like a seismograph trace with no earthquake?

When I worked for the Scummy Mortgage Co (a Fortune 500 company), I went to my boss, the VP of DP one day, and showed him what the guy they'd fired had done for an "algorithm" for leap year: if it's 78 or 80 or 84 or 88 or 92, it's a leap year. No, I'm not making this up, nor the response: if it ain't broke, don't fix it. When it breaks, we'll fix it.

In a mortgage co, with most being the usual 30 year mortgages.

I have a friend who, I understand, bragged in the late 80's that he had 700 fonts. Another friend's response was that he needed another 300, so he could be the man of a 1000 fonts.

Have another friend, and we were in an APA together for a bunch of years. It is my opinion that 60% of all fonts look almost identical to the commonly used ones, and the rest were used once, and NO ONE ever wanted to read anything in those fonts ever again.

Check computer labs/machine rooms. Look in cabinets, or on shelves under desks. We had one drive, bought in 1990, that I understand was still workable... that is, at least until the disaster of 2018, I think it was. (Machine room. Steam pipe blowout.).

Jay - the way some people refer to me as whitroth. It's my username/email account, and not a "real" name, completely artificial. I was the only one from '94, when my late wife and I got our first non-work 'Net account, until '16, I think, when some teenaged girl used it (but hers wound up with a number after it.) If you wonder why, my website's https://mwr.5-cent.us

"The important part is the "double-entry-keeping" method, invented in Venice 700 years ago."

True. But how do you guarantee that both entries are consistently made? In a database transaction you do it by wrapping them in START TRANSACTION; and COMMIT;(or ROLLBACK;) statements. Like double-entry, it's a simple idea, but someone had to think of it.

I'd say that the concept of atomic, consistent, isolated, durable all-or-nothing transactions, in accountancy and elsewhere, is more than a mere implementation detail.

Note that there is at least one Springfield in every single US state.

How odd. My father went in (drafted?) in 1942, and his dogtag read [first name] NMI [last name]. The NMI, of course, was "No Middle Initial".

For those not in the US, I'll note that the last of the towns he mentioned are (mis)pronounced "ver-sales".

There is a SF short story called "MS Fnd in a Lbry" about multiple generations of data compression used to store data on denser and denser substrates. It does not end well.

That can be an irritation. Mismatch your names on your passport or other ID and an airline ticket in the US some time.

Why? All of us use the names on our passports and I'm fairly sure my youngest doesn't even remember the name on her birth certificate. I can't remember the last time I even used my birth certificate for anything official. Nobody would ever ask me for it.

The most excitement I've had with US Immigration was the officer at MSP who thought I was Irish (having seen the word "Ireland" on the photo page) and got worked up about the glossy cover to that page.

Sorry, that's number two. The scariest was the JFK immigration officer who was convinced that I had overstayed my last entry permit and was illegally in the USA. He eventually accepted that, given I was standing in Immigration among lots of other people arriving on the same flight, carrying the stub of my boarding pass and a paper ticket (remember them?) that I probably hadn't been hiding out in Queens for the last four months.

(Scariest overall was, back in the 1970s, being stopped by the KGB in departures at Leningrad Airport for them to find Russian army movement orders in our bags.)

Monthly things that start on the 31st. Annual things that start on Feb 29. Ugh.

There's UK case law on this.

One month from 31st March is 30th April. One year from Feb 29 is Feb 28.

Move the correct number of months or years forward or back. Then, if the resulting month doesn't have a day with that number, go to the last day of the month.

Mismatch your names on your passport or other ID and an airline ticket in the US some time

Name mangling does that admirably. My ex is a "Thi something something Le", combining the 'first name' used by approximately half the women in Vietnam with the most common last name in that country. Obviously such women are known by one of their 'middle names', with no great redgularity as to which. Her and her four sisters have between them five 'first names' so, for example, the older two at 'Thi A B Le' and 'Thi B A Le' because they're twins and that just makes sense.

Australia has a very formal naming system. For each government department or computer system. There's no rule says they have to use the same naming system, so they don't.

Thus my ex at one stage had a NSW driving license with the name "Some Thing Le", an Australian passport in the name "Thi Some Thing Le", a Medicare card in the name "Le, T Some T" and so on. No two alike!

Her solution was to take all the documents when she needed any of them. From birth certificate right down to school photo ID card. Pile them on the counter and let the PTB work out what to do.

She got out of Australia then into Vietname very easily, and equally easily out of Vietnam into Laos and back easily, then out of Vietnam and back.... ah, yeah, so "into Australia" took several hours. She's a citizen, she used the citizen "speed aisle" and she didn't get drug searched, just quizzed about her name and left to think about her choice of name/parents while people fussed about how to process her entry.

The scariest was the JFK immigration officer who was convinced that I had overstayed my last entry permit and was illegally in the USA. He eventually accepted that, given I was standing in Immigration among lots of other people arriving on the same flight, carrying the stub of my boarding pass and a paper ticket (remember them?) that I probably hadn't been hiding out in Queens for the last four months.

I had that experience in Toronto once! Back in the days of paper/cardboard entry permits, at a point when the airport was a building site so everything was temporary, which I guess is where the confusion happened -- we surrendered our cards when leaving after a trip, then got hauled aside when we returned a couple of years later.

In Canada there was no shouting, just forty minutes in the seating area of a waiting room while they confirmed that no, we couldn't have magically overstayed our permission.

Re: ' ... de facto "house style" in that anything the institution can accept for publication/archival must stick within limits little different ... As long as the data is there, and people can read it, the researchers en l'an 2800 can fettle up their own methods of getting it into whatever database package happens to be flavour du jour at the time ...'

Sorta like Latin but for 'publishing' code? Latin was the preferred language for official/serious publications/communications until at least Galileo's time. Didn't stop regional languages from developing and made it much easier for people from different countries/languages to communicate. (Pretty sure there are commenters here who could probably give a semester's worth of lectures/write a few books on this.)

About maintaining libraries ...

Reading that published scientific articles are being allowed to be lost forever makes me ask: Haven't these people ever heard of the Library of Alexandria (one of the seven wonders of the ancient world) which when destroyed resulted in a huge setback (hundreds to a thousand years' worth) in ancient as well as then-contemporary knowledge.

How ticked off would these publishers/institutions be if the reason for the loss of these papers was a cyber/unfriendly attack? Would they take more care in protecting these articles?

I am truly, royally ticked off ... the indifference to loss of knowledge, total idiocy!

Haven't read the rest of the comments yet (could take a while) so apologies if this has already been mentioned.

My family at the time -- my sister and I, our mother, our step-father and his two children -- checked into a motel once and confused the management by having three surnames between us.

Not helped by having two dark haired adults, two dark haired children and two blondes.

We have Gove at one end of the country and Grove at the other. Then there are the problems caused by laziness. Our service is in South Australia and we have a community called Nyapari. There's a community in the Northern Territory called Nyrippi -- both of the these communities are serviced by the same hospital. We get a LOT of things that are supposed to go to Nyrippi because doctors type in the first two letters and go "It must be that one".

On a probably trivial level, in the USA the usual English-speaking name convention is GivenName [optional other given names] Patronymic.

But a ton of USAians are of the Hispanic persuasion, where it's GivenName [optional other given names] Patronymic Matronymic. Of course, given patriarchy, the patronym has dominance so Maria Dolores Garcia Gomez would give her one-last-name as Garcia. Which really confuses a lot of Anglophone USAian officials.

Those... how can you prove you signed it, when all there is to see is what looks like a seismograph trace with no earthquake?

When it is an exact match to all my other IDs.

How odd. My father went in (drafted?) in 1942, and his dogtag read [first name] NMI [last name]. The NMI, of course, was "No Middle Initial".

You father didn't have a jerk for a DI. And in those days I'm better there was a lot of local windage for exactly how forms got filled out.

Why? All of us use the names on our passports and I'm fairly sure my youngest doesn't even remember the name on her birth certificate.

Spelling or transcription errors.

My wife worked for a major US airline for years and we did a LOT of travel. We got to see a lot of interesting discussions at ticket counters and gates. With a side order of "please step over here" at TSA.

Lots of South American countries have laws about citizens taking money out of the country and bringing high in country taxed items in. And some with dual passports would try and play the games of one passport in one direction and the other in the other direction. When she was doing international phone calls people would call up and ask how to get out of "home" with their money or through "home" customs with their stuff. And she got to politely tell them "not my problem".

My grandmother had some difficulty getting out of Hungary in the 80s. She was a Canadian citizen with a British passport and a Hungarian last name. The Hungarian officials just could not wrap their heads around it.

My last name is Hungarian and somehow manages to confuse a lot of people trying to pronounce it. Back in the days of word processor mail I used to get government correspondence to various horribly misspelled versions of my last name, presumably because whoever was typing it ran into a couple of consonants, panicked and just threw a bunch more in rather than simply transcribe the 9 letters in order.

My birth sirname is shared with a town in England. It's very bloody English indeed. All three common spellings of it. Which get applied at random by many native English speakers. It's not unusually long, and doesn't do any of the more fun things the English do with names (no "ough pronounced ch" or anything). But it does lead to people, including officials, utterly failing the "type what you see on the card" test.

Viz, you don't have to have a "foreign" name for the English to have trouble.

Sincerely, Bob Lundon. Or Londan. Landoun? Something like that.

A Shaun Micallef character used the term "tricksy" a lot but I can't find a relevant video.

Anyway, using aboriginal names for places is a very dirty tricksy thing to do to naive (native?) English speakers.

On a couple of Oz blogs I've been signing myself "Moz of Yarramulla" for a few years because it drives a certain sort of misbegotten pedant completely insane (some search engines autocorrect it). Yaralumla is some important government site, Yarramulla is a shed in the bush in Queensland. But they're both real places, and given the choice I'd visit the latter over the former any day. I kind of want to visit Yarramulla just so I can say I've been there :)

https://winsart.com/prints/sanctuary/ random painting of a site near Yarramulla Ranger Station in Undara National Park, Queensland is one of the few matches...

I had that experience in Toronto once! Back in the days of paper/cardboard entry permits, at a point when the airport was a building site so everything was temporary, which I guess is where the confusion happened

Point of interest: it's still a building site. Last time I went through I had to ask for directions, because scaffolding was covering the signs and no one had thought to post temporary signs…

Most airports are always building sites, but that one time International Departures at Toronto was particularly bad (temporary desks in front of plastic sheeting to wall off the concrete dust and jackhammers) ...

About maintaining libraries ...

Never heard of labs referred t as paper mills? I’ve seen it in action. Sturgeon’s Law definitely applies to science and every other Publish Or Perish field. Probably it applied to the Library of Alexandria too, to be honest about it. So does the Suck Fairy. And if not burning the Library meant that humanity created social media in 1300 and avoided the Little Ice through Anthropogenic Climate Change, that’s better than now because…?

I’m sorry to be nasty about this, but I’m currently downsizing the belongings of someone who never threw anything out, while that person slowly dies in a nursing facility. I’m plagued equally by those who want me to hire a service to throw everything in the landfill, under the idea that the house is worth more money than the contents, and those who don’t want me to throw anything out, because it might be useful someday. And I’m getting an entirely unwanted education into how dysfunctional their and my relationship with stuff is.

About the only thing I’ve learned is that I want to have some control over how my junk is discarded, because I have no desire to do unto others what’s been done unto me. That means I’m throwing away my junk too, while I still can.

And so it goes.

we've gotten tangled in the weeds

never mind how the dates are formatted, my examples were of gaffs which have all too oft led to nightmares when calculations were performed (loans, expirations, adulthood verification, etc) and appointments were scheduled for medical treatment (31-SEP-2025 will never happen so the patient died awaiting a phone call where to go)

every app I had any contribution on design of the interface, allowed customization by individuals: which date format, time format, greytone versus color graphics such as pie charts, etc

at most sites I got shot down and most project managers would pull me aside for STFU feedback

what's really funny, about 9 outta 10 managers would call me 2 (or 3 or 5) years later requesting those specifications, as if I had violated My NDA and took 'em offset... because? ...guess what the app was going international or there was a VVV-VIP who was color blind, mobility challenged or simply with old eyes and all my notes (and associated business rules) were now critical to the revised deliverables...

in all cases I pointed out there was supposed to be centralized backups so just request restoration... tee hee... turned out the tapes were unreadable or the fees of the offsite storage vendor were ruinous...

some managers hired me to rewrite those missing chunks and it was nice chunk of change for weekend moonlighting

as far as I can determine no major corporation in the (USA) financial sector has bothered to establish enterprise-wide standards for data validation, interface design, compliance with special needs of disabled, and various other * DUH * things too insignificant to concern the CXOs but keep slamming technical support

BTW: anyone else bought a new laptop recently? any of you finding WIN 11 a step backwards in usability and customizability?... fracking scroll bars are no longer adjustable and never mind the misery of button colorization/contrast

About the only thing I’ve learned is that I want to have some control over how my junk is discarded, because I have no desire to do unto others what’s been done unto me. That means I’m throwing away my junk too, while I still can.

The Swedish term for that is döstädning.

...and 'slashing the seven' also useful

problem being not everyone likes it

FUNFACT: the world was first tripped up by Y2K not in the 1990s but in 1968

yup... those 30Y loans... because back then it could take up to two years to wrap up all the lingering bits 'n pieces if ever a loan went to full maturity rather than being paid off early

...and then again in 1978 for 20Y loans

serious suggestion: contact your nation's version of CIA or ONI or NSA to inquire if they have intel trainees in need of a challenge

you get free expertise and they get a low stress (no nukes hidden in Boston basements ala The Merchant Trade) real world exercise in 'hostile' data retrieval

that one time International Departures at Toronto was particularly bad (temporary desks in front of plastic sheeting to wall off the concrete dust and jackhammers)

My worst construction experience at Pearson was trying to find my way between terminals at 1 AM with no signs and almost no one around, when the corridor I needed was an unmarked passageway that looks like an entrance to the construction site itself. (Some bright spark had posted the "No admittance" sign right at the entrance, apparently meaning "don't go behind the sheeting" rather than "don't go down this corridor".) Fortunately for me another passenger knew the right way to go.

a thousand blessings upon you... during the quarantine I tried to find that story to send to a buddy who needed to explain why disaster recovery planning was necessary and its documentation had to be in a form readable during a crisis...

my friend is now hunting down "17 X Infinity" so she can, in future miseries akin to that 2020-centered SNAFU have it at the ready

from Wikipedia:

"The short story first appeared in the December 1961 issue of The Magazine of Fantasy and Science Fiction.[2] It was anthologized in the collections 17 X Infinity (1963) edited by Groff Conklin and Laughing Space (1982), edited by Isaac and Janet Asimov.[3] It was also published as "Ms Fnd in a Lbry: or, the Day Civilization Collapsed".

but that one time International Departures at Toronto was particularly bad (temporary desks in front of plastic sheeting to wall off the concrete dust and jackhammers)

Sounds like a typical Thursday at LGA. (I understand they are between projects just now and it's a decent place to wander about.)

Our local airport will likely start this for 3 years in a few years. Add a new runway then expand the terminal out onto the existing taxiway for the old runway. Old runway become taxiway. But the main terminal building expands along its entire length.

Currently it is a very nice and easy to deal with mid sized airport. But running out of capacity and in desperate need to replace the long runway so overseas flights can continue.

RDU

Some bright spark had posted the "No admittance" sign right at the entrance, apparently meaning "don't go behind the sheeting" rather than "don't go down this corridor".) Fortunately for me another passenger knew the right way to go.

I ran into this at Miami 5+ years ago.

At the terminal with my gate I was something like gate 41. The terminal is sort of a Y shape. I was starting at the tail. Signs overhead said things like gate 8-44, then 13-44, and so on. At the Y split there was one sign for gates 20-30 and another for gates 31-37 or similar. Where are the rest? I doubled back, looked around a bit then found a map. Apparently they had extended the branch of the Y with the 30s gates but never changed that sign at the "fork".

I was going to the gate early so I still made the flight.

My wife had just taken over the digital airport signage product and so I took some pictures and sent them to her. Big sigh. She only had control over TV signage from the airline, not hard signage. And if you ever wonder why the TV displays next to each other in an airport at times display different and maybe conflicting information? Well about 1/2 of those TVs are controlled by the airport and 1/2 by the airlines. Go figure.

Many modern ideas about the Library of Alexandria come from Carl Sagan who had not learned anything about the history of science since his undergrad days. Probably it suffered a series of disasters, a long trend of defunding and milking for money ("hello scholars! welcome my drinking buddy, your new boss!"), and was not as important as the general decline in the ability and willingness of people in the Roman world to pay to have secular books copied. A good series of blog posts are here https://kiwihellenist.blogspot.com/2022/11/alexandria.html

Post-300, but I thought it might interest some here:

https://www.pulpmags.org/index.htm

The Pulp Magazines Project is an open-access archive and digital research initiative for the study and preservation of one of the twentieth century's most influential print culture forms: the all-fiction pulpwood magazine. The Project also provides information and resources on publishing history, multiple search and discovery platforms, and an expanding library of high-quality, cover-to-cover digital facsimiles.

Kardashev @ 305:

On a probably trivial level, in the USA the usual English-speaking name convention is GivenName [optional other given names] Patronymic.

But a ton of USAians are of the Hispanic persuasion, where it's GivenName [optional other given names] Patronymic Matronymic. Of course, given patriarchy, the patronym has dominance so Maria Dolores Garcia Gomez would give her one-last-name as Garcia. Which really confuses a lot of Anglophone USAian officials.

All you gotta' do is put a hyphen between Garcia & Gomez and your last name becomes Garcia-Gomez. I don't think I ever encountered a system that didn't handle hyphenated last names.¹

Personally, all of my encounters with "official" names required me to give [LASTNAME] [FIRSTNAME] [MIDDLEINITIAL] ... and by the time I came along THEY had figured out to put [NMI] in that field when there was no middle name.

A couple of oddities I've encountered in my lifetime - a pair of twin boys named "John James Smith" and "James John Smith" and a kid who was the FIRST in his family to be born in a hospital. His parents didn't understand the initial certificate of live birth used "Baby Boy" as a placeholder ... so when I knew him he was already in his 30s & still going by his middle name "Boy" ... 🙃

¹ The guy who sat behind me in Sophomore year High School English class was Tremont Aloysius Yates-Smith THE THIRD, and you know damn well the schools DID NOT DARE fuck up HIS last name in school records ...

Heteromeles @ 314:

I’m sorry to be nasty about this, but I’m currently downsizing the belongings of someone who never threw anything out, while that person slowly dies in a nursing facility. I’m plagued equally by those who want me to hire a service to throw everything in the landfill, under the idea that the house is worth more money than the contents, and those who don’t want me to throw anything out, because it might be useful someday. And I’m getting an entirely unwanted education into how dysfunctional their and my relationship with stuff is.

I expect most people are like me, a lifetime accumulation of some things that have obvious value and a lot of things that can only be described as Why in the hell would you want to keep THAT? ... all I can say is I inherited a dominant pack-rat gene.

Plus paranoia so that NOTHING that comes here in the mail with my real name & address printed on it can be discarded into the recycling without going through the shredder.

Shredding that shit is tedious (not to mention the recycling system around here DOES NOT want to deal with shredded paper).

Best that can be said about the situation is that someday I'll be dead and then it will be someone else's problem.

Howard NYC @ 315:

we've gotten tangled in the weeds

never mind how the dates are formatted, my examples were of gaffs which have all too oft led to nightmares when calculations were performed (loans, expirations, adulthood verification, etc) and appointments were scheduled for medical treatment (31-SEP-2025 will never happen so the patient died awaiting a phone call where to go)

every app I had any contribution on design of the interface, allowed customization by individuals: which date format, time format, greytone versus color graphics such as pie charts, etc

About the only place I ever have to give a date (Date of Birth) to any more is using the automated refill system at the VA and they are very explicit guiding you through how to enter it ...

"four digit year, two digit month and two digit day. FOR EXAMPLE, IF you were born January First Nineteen-Sixty you would enter One, Nine, Six, Zero, Zero, One, Zero, One followed by the pound sign ..."

[...]

BTW: anyone else bought a new laptop recently? any of you finding WIN 11 a step backwards in usability and customizability?... fracking scroll bars are no longer adjustable and never mind the misery of button colorization/contrast

I kind of want a new laptop because my existing laptop screen is too small, but I'm still pissed off at Windoze10 and damn sure am resisting Windoze'leven. And anyway no NEW laptop is going to allow me to install PhotoShop CS6 (my PhotoShop system still runs Windows7).

I might think about resurrecting my OLD laptop that did have a 17" diagonal screen, Windows 7 & Photoshop CS5 ... I don't really do any image processing on the laptop anyway, but it's nice to be able to review my images while I'm traveling and the larger screen helps with that.

My favourite batshit database-breaking name has got to be Richard Grosvenor Plunkett-Ernle-Erle-Drax. Who is a Tory MP and a complete and utter arsewipe, but you can torture-test your input validation routine for names using that surname alone! (It was granted to his family by King George V to his grandpa, Sir Reginald Aylmer Ranfurly Plunkett-Ernle-Erle-Drax, younger son of the 17th Baron of Dunsany and younger brother of the famous author Lord Dunsany.)

On the subject of "wonky" experiences ... I've found a new way to freak out my little dog.

The "exhaust" fan built into the underside of the microwave (mounted above the stove) doesn't really do a very good job.

I've lately been trying my hand at doing roasts in the oven and for some reason whenever I open the oven to check the temperature of the meat it sets off one of the smoke alarms (at the other end of the house).

Good idea - and maybe they did that (as one government organisation working with another). But I left that organisation in 1980's and have no idea how they solved that problem.

Certainly re-processing 'zillions' of punched cards looking for wood worm holes (aka "borer" in NZ) wasn't considered practical when that problem was discovered - retrospectively as - far as I am aware before I left. I suspect the cards were dumped/pulped after being originally copied to tape to reclaim the storage space occupied by the large quantities of punched card storage boxes.

All you gotta' do is put a hyphen between Garcia & Gomez and your last name becomes Garcia-Gomez

Except for the minor point that that's not her name, right Seán?

That's a lovely painting. Do you actually have a connection to Yarramulla or do you just like doing people's heads in? ;-)

Ahh yes, the character Female in Cat People -- love the Malcolm McDowell version, haven't seen the original

Three of the most common and important things we keep on computer systems are names, addresses and date/times. And what are some of the most difficult to handle things? Yes...

I have a colleague originally from Morocco who explained that it is common for an address to get seriously convoluted as time passes. Roads get renamed or kinda disappear or extended, and new buildings get stuffed in between old ones. You can end up with addresses akin to “3rd building past number 372 on el thingy avenue, behind the shops of bin-Dibbler, upstairs number 19” . A student intern from Turkey claimed it could be even more complex there. It’s bad enough in UK - one of our addresses was ‘house name’ ‘street name’ ‘village name’ county postcode. When the street is 5+ miles long it can get annoying to find the right house.

A colleague in California had a double first name with no hyphen; loads of systems fail that test, too many assume a space means ‘end’. She got regularly annoyed by it.

And as date/time, well I have colleagues that have spent pretty much their entire careers trying to sort them out.

We could do with name and date handling that works like the mapping website https://what3words.com/

Re: 'Sturgeon’s Law'

Yeah, except it sometimes takes generations to figure out what's crap and what isn't. Also - yeah, I've heard of paper-mill publishing and am also aware that over tens of thousands of academic articles have been pulled/retracted.

My sister and I had to sort through our parents' belongings - not an easy job. I've moved more often than my parents therefore have already had to sort and toss stuff every move. Last time I moved, I mentioned to my progeny that I was editing down. Immediate reaction was: don't throw out any of your books! Also found out that there's some furniture and artwork of interest.

Re: Library of Alexandria -

I think I've read this on more than just Wikipedia: according to Galen, Ptolemy II required that every ship entering the port hand over their 'books' to scribes to be copied for the Library. Guess that Sagan was referring to Hipparchus, "the father of astronomy". The only topic areas I've heard specific reference to as lost in the destruction of the Library were math and Greek plays. (Only 18 out of 92 Euripides plays survived, considerably more than the number of plays by Aeschylus or Sophocles. Imagine if we lost Shakespeare!)

All you gotta' do is put a hyphen between Garcia & Gomez and your last name becomes Garcia-Gomez. I don't think I ever encountered a system that didn't handle hyphenated last names.

I knew someone who went to the high school I ended my career at. When I found her picture on the wall in the "Class of 19xx", her name was wrong. I asked her, and apparently that's what she was called all through school because the computer system (and the staff) couldn't handle her actual last name because it was too long for the last name field (and 'hard to say').

So instead of Vinayagamurthy, they changed her last name to Vina — including on her diploma.

That wouldn't happen nowadays in the school system. Although some First Nations people have trouble because their names include digits and accents that government computer systems (and bureaucrats) don't recognize.

https://www.cbc.ca/news/canada/british-columbia/indigenous-names-vital-stats-1.6426239

Which is a long-winded way of saying that forcing someone to change their name to fit the database is a really bad solution.

Reading that published scientific articles are being allowed to be lost forever makes me ask: Haven't these people ever heard of the Library of Alexandria

And this popped up Friday.

https://arstechnica.com/science/2024/03/study-finds-that-we-could-lose-science-if-publishers-go-bankrupt/

or do you just like doing people's heads in

By now I have no idea. It's entirely possible that 15 years ago I met someone who worked or had worked there and the name stuck, but it's equally possible that I randomly heard it somewhere or indeed misspelled a search and found it.

I'm taking the question on notice as rhetorical :)

I mentioned to my progeny that I was editing down

Anyone after "heir and spare" is superfluous once they reach adulthood?

Re: '... popped up'

Thanks! I read the article and looked up the referenced author, Martin Eve - very interesting combination of arts, science, tech and accessibility.

https://en.wikipedia.org/wiki/Martin_Paul_Eve#External_links

Did a double take when I read he was expert on David Mitchell (author/TV writer/screenwriter). The only David Mitchell I was aware of is a comedian/actor/writer and frequent guest panelist on a bunch of BBC comedy panel shows (QI - Quite Interesting, Would I Lie To You, etc.) Good example of the importance of a middle name/initial*.

https://en.wikipedia.org/wiki/David_Mitchell_(comedian)

*Okay, there are a bunch of fairly common European surnames but I'm guessing that this is nothing in comparison with some Asian (esp. Chinese and Korean) surnames. Wonder how they keep track of who's who - they must have come up with some sort of system.

Moz @ 340:

re: 'heir & spare'

I think that distinction only matters for royalty. Biggest surprise was that there was any interest in my stuff: different generations, different tastes.

Yes, lots of people who are reliable on natural science or engineering repeat myths about the ancient world and the history of science from Carl Sagan or 19th century anti-Catholic pamphlets. They often got them via an intermediary who was just as ignorant about those topics. There is a lesson in that: someone can be skeptical and scientific in one area, but repeat whatever someone told them to repeat about another topic.

OTOH, we have an actual letter of Galen listing things which were lost in a specific library fire at Rome in 192 CE. But as long as the Roman empire was rich, papyrus was plentiful, and secular learning was fashionable, no one disaster could do too much damage. The Library of Alexandria gets used as a standin for the long anonymous processes that reduced the resources available for copying Greek learning. https://www.jstor.org/stable/41724875?mag=library-fires-have-always-been-tragedies-just-ask-galen

I'd gently suggest that our problem as a civilization isn't just that we're using paper, but that we're committing our records to electronic media. And I think most of us are justifiably cynical about how long those will last. We moved fast, and one of the things we seem to have broken is our intellectual links to the deep future of humanity. Does this matter, or not? Any interest in building it back better?

And I think most of us are justifiably cynical about how long those will last.

i think someone here mentioned microsoft's project silica, which sounds promising, though there may be room for cynicism about how long our ability to read examples of it will last

it could end up forever tantalizingly out of reach, like the 100 million or whatever in bitcoin that guy forgot the password to

'You can end up with addresses akin to “3rd building past number 372 on el thingy avenue, behind the shops of bin-Dibbler, upstairs number 19”.'

In Japan, roads mostly don't have names (why would you name an empty space?) so addressing is by city, district, block number, building number, ... and those numbers may not be assigned consecutively. To find an address, consult the collective cultural memory, i.e. go to the nearest police box (there are many) and ask.

not enough cow bells {G}

for sure, lacks suffixes sufficient to break the code

"Jr" is so pedestrian as to be shrugged off

I've had situation where people with PhD insisted their fracking mailing address include that suffix

...and prefix of "Dr." with mandatory period... she made a concession in not insisting upon "Doctor" spelled out

and anticipating everyone's next snarky inquiry, why yes, she was a graduate of Yale, with her doc in Foreign Affairs

her insistence was based upon others of equal social stature who would be miffed if ever any of [REDACTED BANK]'s statements failed to show 'minimal' respect

there's grease accumulated on the blade surfaces from vaporized fat throwing off the balance as well heavier working load than designers planned for

it is one of those moments when it is clear engineers were never expecting their microwave to be used to cook food that would reach vaporization threshold

Scotland, Isle of Skye. On one of the townships on the B883 (just off the A87 at the North end of Glen Varigill) there is the postal address "Quarter of Five", $Township. This is because the croft was split between 4 male siblings.

hmmm... not inflicting Romeo and Juliet on a room filled with hormonal ticking bombs seems now in retrospect a good idea...

high school was bad enough without inflicting tales of interclan gang wars and teenage mass suicides

I'd gently suggest that our problem as a civilization isn't just that we're using paper, but that we're committing our records to electronic media. And I think most of us are justifiably cynical about how long those will last. We moved fast, and one of the things we seem to have broken is our intellectual links to the deep future of humanity

One of the things examined in Benford's Deep Time (which, if you haven't read, I recommend).

One of the lessons I learned from reading it years ago is that engineers, especially electronics engineers, apparently have no concept of time. Or maybe it's scale that they struggle with — both the amount of information we've generated and the amount of change that can happen over long periods of time. ("Long" apparently being anything longer than their professional career to date.)

addressing is by city, district, block number, building number, ... and those numbers may not be assigned consecutively

I was under the impression that they were assigned sequentially, but by order of construction rather than location on a road. So ten houses built at the same time may have numbers in sequence as you walk up the road, except of the house in the middle that came later so has a higher number.

I could be totally wrong about this.

palate cleansing...

"The Color of Magic"

https://www.cwtv.com/shows/the-color-of-magic/

not inflicting Romeo and Juliet on a room filled with hormonal ticking bombs seems now in retrospect a good idea

My students were rather bored with Romeo and Juliet until I pointed out to them that all the characters were their age. Their English teacher hadn't mentioned that, and the actors they had seen were all adults, so between that and the costumes they'd thought of it as "boring inexplicable adult stuff".

Once they realized that it was a story of teenagers rebelling against their parents it made much more sense to them, and they got more into it.

We moved fast, and one of the things we seem to have broken is our intellectual links to the deep future of humanity. Does this matter, or not? Any interest in building it back better?

Who is "we", Kemo Sabe?

"Deep future of humanity" is something very few people ever cared about, let alone had intellectual links to. Vast majority of people through history were just too busy surviving to think about such matters. Those he had such luxury rarely did, and those few who did usually assumed that things will remain the way they are forever.

The idea that future will be significantly different from the present is only about 200 years old. And while many many people during these 200 years did think about it, almost all of them were concerned with the scale of decades or centuries. Hardly ever millennia, let alone longer.

I daresay this blog has an extremely high concentration of the people with "intellectual links to the deep future of humanity", and it distorts the perception. (I freely admit I am not one of them)

Way off topic: fans of Penric & Desdemona should not miss this interview.

In the mid-90s, I worked for a software company in the UK that specialised in name and address management. We had a product for validating addresses and adding postcodes if missing. About 10% of the code did about 99% of address validation and the other 90% of the code dealt with the 1% of weird exceptions that were complex to manage.

That said, the UK address system is pretty well organised, even with the vagaries of house names, apartments, etc. For probably 90% of addresses in the UK, you can get away with as little as a house number and a postcode, and the mail will find its way to your door (little-known detail: every address in the UK has a unique identifier made up of the postcode and an additional 2-digit code known as the Delivery Point Suffix - this limits the number of delivery points/mailboxes in a postcode to ~80). Many software and cloud service providers out there take the Royal Mail's basic Postcode Address File (PAF) https://www.poweredbypaf.com/ and enhance it with all sorts of other data. A common enhancement is grid references to varying levels of accuracy.

Of course for personal use, Google Maps or one of its competitors does most of this now anyway.

TL:DR, it's exceedingly rare in the UK for a business to be unable to deliver an item to an address, and Royal Mail is very good at it (for all their other faults).

"I could be totally wrong about this."

No, I think you're right. I could have been clearer; what I meant was that they are chronologically but not geographically sequential.

Yesterday I was UTC-5. Today it's UTC-4. I'll spend the rest of the week resetting all of my clocks. I wish they'd just pick one or the other & stick with it.

anonemouse @ 332:

"All you gotta' do is put a hyphen between Garcia & Gomez and your last name becomes Garcia-Gomez"

Except for the minor point that that's not her name, right Seán?

It's closer to it than either Garcia or Gomez alone. How do database administrators in Spanish speaking countries handle double barrel last names?

PS: ... that's Jean to you ... or Juan.

PPS: You have no idea how many different ways I've seen my name mangled, so you're going to have to do better than that 😏

Pixodaros @ 343:

OTOH, we have an actual letter of Galen listing things which were lost in a specific library fire at Rome in 192 CE. But as long as the Roman empire was rich, papyrus was plentiful, and secular learning was fashionable, no one disaster could do too much damage. The Library of Alexandria gets used as a standin for the long anonymous processes that reduced the resources available for copying Greek learning. https://www.jstor.org/stable/41724875?mag=library-fires-have-always-been-tragedies-just-ask-galen

Two thoughts - don't get caught "stealing" articles from "jstor"

My current "bedtime" reading is From Hittite to Homer The Anatolian Background of Ancient Greek Epic

Comparative study based on translations of the cuneiform tablets excavated at Hattusa. Page & a half and "boom boom - out go the lights!" ... more effective than "Ambien" and fewer side effects. 🙃

Howard NYC @ 348:

there's grease accumulated on the blade surfaces from vaporized fat throwing off the balance as well heavier working load than designers planned for

it is one of those moments when it is clear engineers were never expecting their microwave to be used to cook food that would reach vaporization threshold

Could be.

I'm going with the theory the [EXPLETIVE!! DELETED!! "builders"] fucked up the microwave installation (AND the smoke detector installations) just like they fucked up so many other things ... incompetence rather than malice, but I won't rule out MALICIOUS INCOMPETENCE!

Plus the oven hasn't been used enough yet for that much grease to accumulate ...

I don't how it plays out elsewhere

my high school (NY, 1970s) had a streak of suicides and attempts... 4000 students of whom at least a dozen who simply were not in class the next day and teachers mumbled about the family suddenly moving away

as to violence... we had lots of medium range aggression that was more bully-on-nerd which made for sweaty moments for many of us and somehow reading R&J encouraged some knuckleheads in trying to form up a posse

feh on Shakesppppppear

Howard NYC @ 347:

her insistence was based upon others of equal social stature who would be miffed if ever any of [REDACTED BANK]'s statements failed to show 'minimal' respect

A few years back I used to regularly get another person's bank statements (BIG name NYC bank) - we were in the same ZIP code, but lived several blocks apart. My address was 506 Frank and his was 506 Franklin.

I finally figured out the postal service wasn't reading the address & zip code, they were using an Intelligent Mail Barcode printed on the front of the envelope below the Name & Address for sorting ... and the bank had screwed that up.

The bank was printing the Intelligent Mail Barcode for MY address on HIS bank statements.

dang... yet again cold facts getting in the way of one of my whimsical pet notions {G}