GAMAYUN+ Wiki - User contributions [en]

Calendars Are Political

2026-08-07T06:52:10Z

GAMA: Converted from site HTML fragment to wikitext: citations, categories, cross-links

Every calendar on this page is a compromise somebody in charge forced onto everyone else. None of them are "the" calendar — they're the surviving proposals. See [[On Time, and Why It's Hard]] for the machine side of this problem; this one's the human side.

'''Ten days that never happened.''' Thursday, October 4, 1582 was followed by Friday, October 15, 1582 — by papal decree. The Julian calendar (46 BCE, one leap year every 4 years flat) overcounted the true solar year by about 11 minutes annually. Over 1,600-odd years that drifted the calendar 10 days away from the actual seasons, which mattered to the Catholic Church mainly because it threw off the calculated date of Easter. Pope Gregory XIII's fix, laid out in the 1582 bull ''Inter gravissimas'', was one new rule: century years are leap years only if divisible by 400.<ref>[https://en.wikipedia.org/wiki/Inter_gravissimas Inter gravissimas — Wikipedia]</ref> 1700, 1800, and 1900 lost their leap day; 2000 kept it. That single rule is accurate enough that the Gregorian calendar won't need another correction like it for about 3,000 years.

'''The adoption lag is the real story.''' Catholic Spain, Portugal, Italy, and Poland switched immediately in 1582. Protestant and Orthodox territories weren't taking calendar advice from Rome and held out for over a century. Britain and its colonies (including what would become the US) didn't switch until 1752 — by then the drift had grown to 11 days, so Wednesday, September 2, 1752 was followed by Thursday, September 14.<ref>[https://en.wikipedia.org/wiki/Calendar_(New_Style)_Act_1750 Calendar (New Style) Act 1750 — Wikipedia]</ref> Russia held out until after the 1917 revolution; the "October Revolution" happened in ''November'' by the Gregorian count, which is why the anniversary date and the actual season never lined up for anyone using the newer calendar. Sweden, in a botched attempt to switch gradually by skipping leap years, ended up with a calendar so wrong they gave up and added an extra day back in — giving 1712 a February 30th, the only one in recorded history.<ref>[https://en.wikipedia.org/wiki/Swedish_calendar Swedish calendar — Wikipedia]</ref> Nobody planned that. It happened because a bureaucracy tried to be clever about a 10-day correction and lost track of its own math.

'''The thirteenth sign that keeps "returning."''' Every few years, a headline resurfaces claiming astrologers "added" a 13th zodiac sign, [[wikipedia:Ophiuchus|Ophiuchus]], and everyone's sign changed. Nothing was added. The sun's actual path against the fixed background of constellations does pass through 13 of them, including Ophiuchus, for about 18 days in late November/early December — that's just where the constellations happen to sit. Western tropical astrology was never built on the actual constellations to begin with; it divides the year into 12 equal 30-degree slices anchored to the seasons (the equinoxes and solstices), a system that split off from the literal night sky roughly 2,000 years ago because of the same slow wobble that also, unrelatedly, keeps Polaris from being the North Star forever — axial precession.<ref>[https://en.wikipedia.org/wiki/Axial_precession Axial precession — Wikipedia]</ref> Sidereal astrology (used in Vedic tradition) does track the actual constellations and has effectively always included Ophiuchus-adjacent overlap. Neither camp "changed" anything; the headline just rediscovers, every few years, that two different systems were never measuring the same thing.

'''The calendar Kodak actually ran on.''' The International Fixed Calendar — 13 months of exactly 28 days (four even weeks each), plus one or two days a year that belong to no month or week at all — was pitched across the late 1800s and early 1900s as a genuinely better structure: every date lands on the same weekday every year, forever, no more "what day of the week is the 15th" lookup. It never got adopted nationally anywhere. Kodak's George Eastman liked it enough to run the company's entire internal payroll and production calendar on it from 1928 to 1989, one of the only large-scale real-world trials the design ever got.<ref>[https://en.wikipedia.org/wiki/International_Fixed_Calendar International Fixed Calendar — Wikipedia]</ref> It died with a business decision, not a technical flaw — shipping software and international customers were increasingly unwilling to talk to a company running a private 13-month year.

'''The ten-hour day.''' Revolutionary France didn't stop at replacing the monarchy. The French Republican Calendar (1793–1805) rebuilt the year into twelve 30-day months with poetic, season-based names (Thermidor, "heat," Brumaire, "fog") plus 5–6 leftover days, and briefly mandated decimal time alongside it: 10 hours a day, 100 minutes an hour, 100 seconds a minute, so midnight was 0 and noon was exactly 5.<ref>[https://en.wikipedia.org/wiki/French_Republican_calendar French Republican calendar — Wikipedia]</ref> It required manufacturing new clock faces and lasted about two years in practice before quietly getting dropped — the metric system's meter and kilogram survived the Revolution; the ten-hour day didn't, because nobody was going to re-machine every clock in the country to accommodate an act of ideology. Switzerland's Swatch tried a much smaller version of the same idea in 1998: '''Internet Time''', dividing the day into 1,000 ".beats" with no timezones at all, beat 0 fixed to midnight in Biel, Switzerland, home of Swatch's headquarters.<ref>[https://en.wikipedia.org/wiki/Swatch_Internet_Time Swatch Internet Time — Wikipedia]</ref> It got real dot-com-era press coverage and a handful of websites that displayed the current .beat as a novelty. It solved a coordination problem nobody actually had (global chat and gaming already just used UTC) and quietly disappeared once the hype cycle that birthed it did too.

'''The throughline.''' A pope fixing Easter math, a monarchy erasing itself down to the clock face, a film-processing company running its factories on a calendar built for evenly-spaced paydays, a watch brand trying to make Swiss time the world's time — none of these were really about time. They were about who gets to decide what time it is for everyone else. The Gregorian calendar won not because it was more correct than the alternatives on offer (the International Fixed Calendar arguably solves more real annoyances) but because it had a 350-year head start and an empire's worth of trade routes behind it.

== See also ==
* [[On Time, and Why It's Hard]]
* [[Digital Preservation & Link Rot]] — another case of an institution's own record-keeping getting quietly rewritten out from under everyone

[[Category:History]]

Reverse Image Search & Metadata

2026-08-07T06:52:09Z

GAMA: Converted from site HTML fragment to wikitext: citations, categories, cross-links

Every photo carries hidden data (EXIF) — camera model, GPS coordinates, timestamp — unless it's been stripped. Reverse image search tools can also tell you where else online a given image has appeared, which is how a lot of "this photo is old, it's not from the event people are claiming" debunks actually get done.

== What's actually in an EXIF block ==
EXIF (Exchangeable Image File Format) is a metadata standard baked directly into the JPEG/TIFF file itself, not something bolted on separately — which is exactly why it's easy to forget it's there.<ref>[https://en.wikipedia.org/wiki/Exif Exif — Wikipedia]</ref> The fields worth actually paying attention to:

* '''GPSLatitude / GPSLongitude.''' Exact coordinates, often accurate to a few meters, written by any phone with location services on for the camera app. This is the one that turns "a photo of my apartment" into "the exact address of my apartment."
* '''DateTimeOriginal.''' When the shutter actually fired — not when the file was last saved or uploaded, which is a separate, less revealing filesystem timestamp entirely.
* '''Make / Model / LensModel / SerialNumber.''' Some cameras (particularly higher-end DSLRs) embed a body serial number. Combined across multiple photos, this can link images back to the same specific physical camera even with no other identifying information.
* '''Software / thumbnail data.''' Editing history and, on some files, a leftover embedded thumbnail generated before the visible image was cropped or redacted — which has occasionally exposed the un-redacted original hiding underneath a "cleaned" published version.

None of this requires special tools to read once you know it's there — it's plain text sitting in the file.

== How reverse image search actually works ==
It isn't a lookup of an image's exact bytes — resize, recompress, or crop a photo even slightly and a byte-for-byte match would fail instantly. Instead these tools compute a '''perceptual hash''': a short fingerprint derived from the image's visual structure (typically by shrinking it down, converting to grayscale, and comparing relative brightness across a grid, or by analyzing which low-frequency components dominate after a discrete cosine transform — the same math family JPEG compression itself uses).<ref>[https://en.wikipedia.org/wiki/Perceptual_hashing Perceptual hashing — Wikipedia]</ref> Two images that look alike to a human end up with similar hashes even if their underlying bytes are completely different, which is what lets a reverse-search engine find a re-uploaded, resized, watermarked, or lightly cropped copy of a photo it's indexed before.

== What it can't do ==
Perceptual hashing compares visual structure, not truth. It can tell you an image has appeared somewhere before and roughly when it was first indexed — it can't tell you whether a caption describing it is accurate, and it's specifically bad at catching AI-generated images that never existed anywhere else to be indexed against, or a genuinely new photo staged to resemble an old, verified one. "This exact image showed up online three years ago" is strong evidence against "this is breaking news happening right now" — it is not, by itself, proof of anything about an image that's never been posted before. Where the metadata trail runs cold before the photo can even be checked, that's usually [[Digital Preservation & Link Rot|the page it originally lived on already gone]], not the image itself being untraceable.

== See also ==
* [[OSINT (Open Source Intelligence)]]
* [[Browser Fingerprinting]] — the same "small unremarkable details add up" logic, applied to a browser instead of a photo
* [[Digital Preservation & Link Rot]]

[[Category:Psychology, Society, & Cognitive Biases]]

Terms of Service (and Why Nobody Reads Them)

2026-08-07T06:52:08Z

GAMA: Converted from site HTML fragment to wikitext: citations, categories, cross-links

The average ToS document is written to be technically disclosed and practically unread. A project called Terms of Service; Didn't Read exists specifically to grade services on what their terms actually say in plain language, rather than what the marketing page implies.<ref>[https://tosdr.org/ Terms of Service; Didn't Read]</ref> Knowing a service's real data practices before you sign up beats finding out after.

== What "agreeing" actually means, legally ==
Courts draw a real line between two ways a site gets your "agreement." A '''clickwrap''' agreement makes you take an explicit action — check a box, click "I Agree" next to a visible link to the terms — and is upheld consistently, because there's clear evidence you had notice and acted on it. A '''browsewrap''' agreement just states, somewhere on the page, that continuing to use the site means you agree — no checkbox, no click, sometimes just a link in a footer. Courts have thrown these out repeatedly when a company couldn't show a user had any real reason to notice the terms existed.<ref>[https://en.wikipedia.org/wiki/Browse_wrap Browse wrap — Wikipedia]</ref> The practical upshot: a service just having a Terms page doesn't automatically mean an enforceable contract exists between you and them — how it was presented matters as much as what it says.

== The clause types worth actually looking for ==
* '''Forced arbitration + class-action waiver.''' Instead of suing in court, disputes go to private arbitration — usually paid for by the company, decided by an arbitrator the company also uses for other cases, with no public record and no appeal. Bundled with a class-action waiver, this means even a company that wrongs millions of users in the exact same way can only ever be challenged one arbitration case at a time.
* '''Broad content license grants.''' Standard language for anything you upload ("worldwide, royalty-free, sublicensable license to use, reproduce, distribute, and create derivative works") is often necessary for the service to function at all — the red flag isn't the license existing, it's how far past "operate the service" it reaches, and whether it survives you deleting your account.
* '''Unilateral changes with no real notice.''' Terms that let the company change the terms at any time, with continued use counting as acceptance of whatever changed.
* '''Data sale / third-party sharing, defined broadly.''' "We do not sell your data" clauses sometimes sit right next to a separate clause permitting broad sharing with "service providers," "partners," and "affiliates."

== A controversy that actually changed a company's terms ==
In December 2012, Instagram updated its Terms of Service to state that a business could pay Instagram to display a user's username, photos, and actions in connection with sponsored content or promotions — without compensating the user whose content it was. The backlash was immediate and large enough (including public figures publicly threatening to leave the platform) that Instagram reversed course within days, deleted the clause, and stated it would not sell user photos.<ref>[https://about.instagram.com/blog/announcements/updated-terms-of-service Updated Terms of Service — Instagram, December 2012]</ref> It's one of the clearer examples of a ToS change getting caught, understood, and actually reverted before it took effect — most changes to most services' terms get neither the scrutiny nor the walkback.

== Why this matters more than it used to ==
"Just don't sign up" used to be real advice. It's a lot less practical when the alternative to a service's terms is often no functional equivalent at all — a bank, an employer's required software, a landlord's payment portal. Knowing what you're actually agreeing to doesn't always come with a real choice to walk away, but it does mean you're not finding out the hard way. The same "read the fine print before you're locked in" instinct is exactly why [[Why There's No Paywall-Bypass Tool Here|paywall economics]] belong in the same conversation as ToS terms — both are about a company writing the rules of a relationship you didn't get much say in.

== See also ==
* [[Why There's No Paywall-Bypass Tool Here]]

[[Category:Psychology, Society, & Cognitive Biases]]

The God of Arepo

2026-08-07T06:52:07Z

GAMA: Converted from site HTML fragment to wikitext: citations, categories, cross-links

Every so often the internet accidentally writes actual scripture, then loses track of who wrote which verse, and somehow that's the part that makes it good. This one started as a writing prompt and ended up a published comic with real charities attached to it. I didn't build a temple for anything. I checked, just in case something showed up anyway. Nothing did. Read into that however you want.

The story started as a Tumblr writing prompt from '''writing-prompt-s''': "Temples are built for gods. Knowing this a farmer builds a small temple to see what kind of god turns up."

'''sadoeuphemist''' wrote the first section: a farmer named Arepo builds a temple, and a small, self-doubting god of fallen leaves, worms, frost, and apples moves in — repeatedly warning Arepo it has no real blessings to offer. Arepo worships anyway, through storm, famine, and war.

'''ciiriianan''' added the turn that made the story: after Arepo's death, a stranger named Sora finds his bones in the ruined temple and asks the god how to honor him. The god admits it has nothing to give. Sora's answer: "You are the god of Arepo."

'''stu-pot''', '''corancoranthemagicalman''', '''threefeline''', and others kept extending it across generations from there — the god of small kindnesses and unbreakable bonds, worshipped and half-forgotten and worshipped again, until it chooses to stay in the same small temple rather than seek anything grander.

The post hit 414,993 notes by the time anyone stopped counting. Read the original thread: https://alcottgrimsley.tumblr.com/post/699640069652398080

'''Reimena Yee''' adapted it into a short comic, debuted at ShortBox Comics Fair 2022. Proceeds from that debut run went to two literacy charities: Ministry of Stories (UK) and Mighty Writers (US). Read it: https://www.webtoons.com/en/canvas/the-god-of-arepo-completed/list?title_no=924206 (WEBTOON) or https://reimenayee.itch.io/the-god-of-arepo (itch.io ebook).

== See also ==
* [[OK: The Word Everyone Uses and Nobody Can Agree On]] — another case of a huge number of strangers building something collectively that nobody planned from the start

[[Category:Mythology, Literature, & Fiction]]
[[Category:Language, Communication, & Internet Culture]]

Google Dorking

2026-08-07T06:52:06Z

GAMA: Converted from site HTML fragment to wikitext: citations, categories, cross-links

Google's search operators (<code>site:</code>, <code>filetype:</code>, <code>intitle:</code>, and combinations of them) can surface content that's technically public but never meant to be easily found — exposed documents, misconfigured directories, forgotten login pages. It's just advanced search syntax, but the results it turns up can look like they came from somewhere darker than "a search engine." The technique is legal — it's just Google search. What you do with what you find is where the ethics and the law both start to matter.

== Where "dork" actually comes from ==
The term has a genuinely funny origin: starting in December 2002, security researcher Johnny Long began collecting Google search queries that turned up vulnerable systems or accidentally-exposed sensitive information, and called the people who left that stuff exposed — not the technique — "google dorks," using "dork" in its older, milder sense of "inept or foolish person." The name for the person stuck to the technique instead. Long's collection grew into the Google Hacking Database (GHDB), organized in 2004 and popularized through his book ''Google Hacking for Penetration Testers''.<ref>[https://en.wikipedia.org/wiki/Google_hacking Google hacking — Wikipedia]</ref> In 2010 he handed maintenance over to Offensive Security (the team behind Kali Linux and the OSCP certification), and the GHDB has lived on their Exploit-DB site ever since, still taking community-submitted queries.

== The operators worth knowing ==
* <code>site:example.com</code> — restrict results to one domain, useful for seeing what a site has indexed that you didn't expect to be public.
* <code>filetype:pdf</code> / <code>filetype:xls</code> — surface documents of a specific type, which is how a lot of accidentally-published internal spreadsheets and reports get found.
* <code>intitle:</code> / <code>inurl:</code> — match text specifically in a page's title or URL, useful for finding known default login-panel titles or predictable URL patterns for exposed admin interfaces.
* <code>intext:</code> combined with a distinctive phrase — default configuration text, a specific error message, a boilerplate string a particular piece of software always includes — can surface every public instance of software that a targeted vulnerability affects.

None of these are secret or restricted — they're documented, ordinary Google search syntax. The "hacking" in Google hacking is entirely in the combination and the intent, not in any operator being off-limits.

== The legal line is real, and it's blurrier than it sounds ==
Running the search itself is legal in essentially every jurisdiction — it's a search engine returning results it already indexed. What happens next is where things get genuinely case-specific: viewing a document Google surfaced is generally treated very differently from using what you found to log into a system, and several real prosecutions have turned on exactly that distinction. This is also why dorking tools that generate ready-made query lists carry their own disclaimer: educational and research use is the stated purpose, and treating a discovery as something to report responsibly — not something to exploit — is the difference between security research and a much worse outcome. It's the same fault line [[OSINT (Open Source Intelligence)|OSINT work]] draws everywhere else: looking is legal, acting on what wasn't meant to be found is where it stops being neutral.

== See also ==
* [[OSINT (Open Source Intelligence)]]
* [[Browser Fingerprinting]]

[[Category:Language, Communication, & Internet Culture]]
[[Category:Psychology, Society, & Cognitive Biases]]

The Myth of Flowing Glass

2026-08-07T06:52:05Z

GAMA: Converted from site HTML fragment to wikitext: citations, categories, cross-links

This one gets repeated at literally every cathedral tour with old stained glass, and correcting it never stops being satisfying.

== The claim ==
Old window panes, medieval church glass especially, are often noticeably thicker at the bottom than the top. The popular explanation: glass is technically a liquid, not a solid, and over centuries it slowly "flows" downward under its own weight, like unbelievably slow syrup. It's a genuinely elegant explanation. It's also not true.

== Why it can't be true ==
Glass does have an unusual molecular structure. It's an amorphous solid, not a crystalline one, which is where the "it's secretly a liquid" idea gets its foothold.<ref>[https://en.wikipedia.org/wiki/Amorphous_solid Amorphous solid — Wikipedia]</ref> But the actual physics doesn't support visible flow on any human timescale. Viscosity research on real medieval glass, published by Gulbiten et al. in the ''Journal of the American Ceramic Society'' (2018) among other studies, found medieval glass is actually about sixteen orders of magnitude less viscous at room temperature than earlier estimates assumed.<ref>[https://ceramics.onlinelibrary.wiley.com/journal/15512916 Journal of the American Ceramic Society — Wiley]</ref> Even with that correction, the maximum plausible flow still comes out to roughly one nanometer over a billion years. The universe itself is about 13.8 billion years old. Do the math and a millennium-old church window hasn't had remotely enough time, even in principle, to sag by an amount anyone could ever see.

== What's actually going on ==
The real explanation is almost embarrassingly simple. Medieval glassmakers using crown-glass or cylinder-glass techniques couldn't reliably produce a perfectly flat, uniformly thick pane, so the process left panes naturally thicker on one edge. When glaziers installed those panes, instinct did the rest: the heavier, thicker edge went at the bottom of the frame, because that's the more structurally sound way to seat an unevenly weighted piece of glass. The glass didn't move there over time. Someone put it there on day one, for a completely mundane, completely sensible reason.

== How the myth got so far ==
This isn't some fringe belief. It's been repeated in real high school science textbooks and even, at points, the ''Encyclopedia Britannica''. Part of how it spread that far: an influential early paper on glass viscosity by physicist Gustav Tammann appears to have gotten misread or mistranslated somewhere along the way, lending the "secretly still flowing" idea a false patina of citation. Once a wrong explanation sounds technical enough, it tends to outrun the boring correct one, which, ironically, is the actual physics — much the same way an unresolved, still-debated real phenomenon like the [[The Mpemba Effect: When Hot Water Freezes First|Mpemba effect]] gets treated as settled science in one direction or another depending who's retelling it.

== See also ==
* [[The Mpemba Effect: When Hot Water Freezes First]]

[[Category:Physics & Perception]]

Dord: The Word the Dictionary Invented by Accident

2026-08-07T06:52:05Z

GAMA: Converted from site HTML fragment to wikitext: citations, categories, cross-links

Oh hey, did you know a real, professionally published dictionary once contained a word that flat-out did not exist, for thirteen years, and nobody who read it ever flagged it? Well, now you do.

== The slip ==
On July 31, 1931, Austin M. Patterson, the chemistry editor for ''Webster's New International Dictionary, Second Edition'', submitted a routine three-by-five index card. It read "D or d, cont./density," meaning the abbreviation "D" (capital or lowercase) could stand for "density," to be added to the dictionary's existing list of things the letter D abbreviates. Standard entry. Nothing unusual about it.<ref>[https://www.merriam-webster.com/wordplay/dord-old-fake-word-dictionary-story "Dord": How a Fake Word Slipped Into the Dictionary — Merriam-Webster]</ref>

== The misread ==
Style at Merriam-Webster required a visible space between every letter of a term being defined, so a staffer processing the card years later saw "D o r d" and read it as one continuous word with the spaces just not rendered clearly, rather than as "D or d." When the dictionary published in 1934, "dord" appeared as a genuine headword, defined as a synonym for density used by physicists and chemists, with no etymology and no citation examples, because there weren't any. There couldn't be. Nobody had ever used the word, because it wasn't one.

== Thirteen years on the shelf ==
An editor named Joseph Wright first noticed something was off about the entry in 1939. It had no etymology, which was unusual enough on its own to raise a flag. The fix, though, didn't actually make it into print until the 1947 revision. In between, "dord" sat in a real, purchasable, professionally edited reference work, doing nothing, meaning nothing, waiting.

== The postscript ==
Fifteen years after the correction, Philip Babcock Gove, by then a Merriam-Webster editor and later the editor-in-chief of the far more famous (and more controversial) ''Webster's Third New International Dictionary'', wrote a letter to the journal ''American Speech'' laying out exactly how "dord" happened, card and all.<ref>[https://www.jstor.org/journal/americanspeech American Speech — Duke University Press / JSTOR]</ref> That letter is basically the reason anyone knows this story at all, and the general category it named, a word that exists only because of an error in compiling a reference work and not because anyone ever spoke or wrote it, is still called a "ghost word" today.

== Not every fake word is an accident ==
"Dord" is the accidental kind. There's a deliberate kind too. "Esquivalience," defined as "the willful avoidance of one's official responsibilities," was invented from nothing by editor Christine Lindberg and planted in the first edition of the ''New Oxford American Dictionary'' (2001) specifically as a copyright trap: if a competitor's dictionary ever turned up with "esquivalience" in it, that would be proof they'd copied Oxford's text rather than compiling their own. It went unnoticed publicly until 2005, when a ''New Yorker'' writer combing through the "E" section asked the editor-in-chief directly and got a confirmation.<ref>[https://www.newyorker.com/magazine/2005/08/29/the-know-it-alls The Know-It-Alls — The New Yorker, 2005]</ref> The twist is that once people knew about it, they started actually using it, and by 2006 "esquivalience" had genuine, real-world usage: a made-up trap word that talked its way into being a real one after getting caught.

== See also ==
* [[OK: The Word Everyone Uses and Nobody Can Agree On]]
* [[Where "Foo" and "Bar" Came From]]

[[Category:Language, Communication, & Internet Culture]]

Browser Fingerprinting

2026-08-07T06:52:04Z

GAMA: Converted from site HTML fragment to wikitext: citations, categories, cross-links

Cookies aren't the only way sites track you. Your browser leaks a surprising amount of identifying detail just by existing — which GPU you have (via WebGL), tiny timing differences in your CPU (via <code>performance.now()</code>), even the exact order your OS signs network packets. None of it requires a cookie, none of it requires you to be logged in, and clearing your cookies or using a VPN doesn't touch most of it. "I'm not logged in and I cleared my cookies" is not the same thing as "I'm not being tracked" — fingerprinting techniques were specifically built to survive the privacy measures people actually use.

== How much any one signal actually reveals ==
No single data point identifies you — your screen resolution alone is shared by millions of people. The technique works by combining dozens of individually unremarkable signals until the combination itself becomes rare. The Electronic Frontier Foundation's original Panopticlick research (the project now continues as Cover Your Tracks) measured the average browser as carrying about 18 bits of identifying entropy from a fingerprint — enough, in information-theory terms, to narrow a given browser down to roughly 1 in 262,000.<ref>[https://coveryourtracks.eff.org/ Cover Your Tracks — Electronic Frontier Foundation]</ref> Later measurements that include newer techniques put the achievable entropy well above 30 bits for a browser with no countermeasures — enough to be close to unique among billions of devices, from signals that individually look completely mundane. That's the same entropy math [[The Odds a UUID Ever Repeats]] uses to prove a UUID basically never collides — here it's run in the opposite direction, to show how few random-looking bits it actually takes to make one browser stand out from everyone else's.

== The techniques that do the most work ==
* '''Canvas fingerprinting.''' A hidden instruction tells your browser to silently render text or shapes to an invisible <code><canvas></code> element, then reads back the resulting pixel data. Tiny differences in GPU, graphics driver, and font-rendering engine mean two different machines drawing the exact same instructions produce measurably different output — a side effect nobody designed for tracking, repurposed for it anyway.<ref>[https://en.wikipedia.org/wiki/Canvas_fingerprinting Canvas fingerprinting — Wikipedia]</ref>
* '''WebGL / audio fingerprinting.''' Same idea, different subsystem — how your GPU renders a 3D scene, or how your audio stack processes a synthesized waveform, both vary just enough between hardware/driver/OS combinations to be a usable signal.
* '''Font and plugin enumeration.''' Which fonts are actually installed (checked indirectly, by measuring how text renders) says something real about your OS, region, and installed software.
* '''Everything else, stacked.''' Timezone, installed language packs, screen resolution and color depth, hardware concurrency (CPU core count), battery API quirks, and dozens of smaller signals each contribute a little.

== What actually helps, and what doesn't ==
Incognito/private browsing mode blocks cookies, not fingerprinting — your canvas and WebGL output look identical in and out of a private window. A VPN changes your IP address and rough location, which is a real and separate privacy improvement, but does nothing to the dozens of other signals a fingerprinting script reads directly from your browser. The approach that actually works is uniformity: the Tor Browser is deliberately engineered so that as many installations as possible report identical values for every fingerprintable signal, on the theory that a fingerprint that's shared by millions of other Tor users isn't a useful fingerprint at all.<ref>[https://support.torproject.org/glossary/browser-fingerprinting/ Browser Fingerprinting — Tor Project Support]</ref> Browser extensions that randomize individual signals (a slightly different canvas hash every visit) can sometimes make things worse rather than better, since "this fingerprint changes in a suspicious, non-human way every reload" is itself a distinguishing signal.

== See also ==
* [[OSINT (Open Source Intelligence)]]
* [[The Odds a UUID Ever Repeats]]
* [[Google Dorking]] — a completely different way for something you didn't mean to expose to end up findable anyway

[[Category:Psychology, Society, & Cognitive Biases]]

Why There's No Paywall-Bypass Tool Here

2026-08-07T06:52:03Z

GAMA: Converted from site HTML fragment to wikitext: citations, categories, cross-links

A specific category of tool exists purely to get around a publisher's paywall. They work. They also directly contradict the terms the publisher set for their own content, which is a different thing from reading an old cached copy of a page that's since changed. This is a deliberate line, not an oversight — if a category of link consistently shows up in every "secret websites" list and consistently doesn't show up here, this is why.

== Not all paywalls are the same wall ==
A '''hard paywall''' blocks the article entirely without a subscription (most of the Wall Street Journal). A '''metered/soft paywall''' gives a set number of free articles before cutting you off, usually tracked by a cookie or a JavaScript counter (most general-interest newspapers). A '''freemium''' model gives away most content and paywalls only specific premium pieces. The metered kind is the one bypass tools target most, precisely because the mechanism enforcing it often lives entirely in the browser — a cookie, a bit of client-side script — rather than on the server.

== This isn't a hypothetical legal question ==
In August 2024, GitHub took down a popular paywall-bypassing browser extension and its thousands of forks after a DMCA complaint from the News Media Alliance, a trade group representing over 2,200 news and media publishers.<ref>[https://www.newsmediaalliance.org/ News Media Alliance]</ref> The complaint's legal basis wasn't ordinary copyright infringement — it specifically invoked Section 1201 of the Digital Millennium Copyright Act, the anti-circumvention provision that makes it illegal to bypass a technical measure protecting copyrighted content.<ref>[https://www.copyright.gov/title17/92chap12.html 17 U.S.C. § 1201 — US Copyright Office]</ref> US courts have separately upheld those anti-circumvention provisions as constitutional. That's the live legal environment a "just add a paywall bypasser" feature would actually be shipping into, not an abstract risk.

== The honest context ==
Industry research consistently finds that only a small minority of readers globally — commonly cited around 15-20% — actually pay for online news, per the Reuters Institute's annual Digital News Report, and a large majority of readers report simply avoiding paywalled sites rather than paying or bypassing them.<ref>[https://reutersinstitute.politics.ox.ac.uk/digital-news-report/2024 Digital News Report 2024 — Reuters Institute for the Study of Journalism]</ref> Paywalls are, for most publishers, one of the few subscription-revenue mechanisms left standing as ad revenue keeps shrinking industry-wide — which doesn't make every paywall well-designed or every publisher's terms fair, but it does mean "just make the content free again" isn't a neutral position, it's taking a side in a funding fight that's already squeezing the newsroom producing the reporting in the first place. Reading an already-archived, already-public snapshot of a page (see [[Digital Preservation & Link Rot]]) is a different act entirely from a live tool built to defeat a paywall the moment it's put up — the former deals with what's already public, the latter exists to make something non-public anyway. The same underlying question — whose rules govern content you didn't write yourself — is what makes [[Terms of Service (and Why Nobody Reads Them)|Terms of Service]] worth actually reading in the first place.

== See also ==
* [[Digital Preservation & Link Rot]]
* [[Terms of Service (and Why Nobody Reads Them)]]

[[Category:Psychology, Society, & Cognitive Biases]]

The Mpemba Effect: When Hot Water Freezes First

2026-08-07T06:52:03Z

GAMA: Converted from site HTML fragment to wikitext: citations, categories, cross-links

Sometimes, under the right conditions, hot water beats cold water into ice. Nobody fully agrees on why. That's still genuinely one of the more entertaining open questions in physics.

== The kid who wouldn't let it go ==
In 1963, thirteen-year-old Erasto Mpemba was making ice cream at Magamba Secondary School in Tanzania. Short on freezer space, he shoved his mix in while it was still hot, instead of waiting for it to cool first like his classmates did. His froze first. When he asked a teacher why, he reportedly got told he was simply wrong about what he'd just watched happen. That was a bad answer, because he kept testing it and kept getting the same result.<ref>[https://en.wikipedia.org/wiki/Mpemba_effect Mpemba effect — Wikipedia]</ref>

== He wasn't the first to notice ==
He was, however, the reason anyone followed up. The observation itself is centuries old. Aristotle wrote in his ''Meteorology'', in the 4th century BCE, that water which has already been heated freezes faster. Francis Bacon noted in the ''Novum Organum'' that "water a little warmed is more easily frozen than that which is quite cold." Descartes wrote much the same thing. All three observations sat around unexplained and largely ignored for hundreds of years, because nobody built a real investigation around them.

== Making it official ==
Mpemba eventually got taken seriously by physicist Denis Osborne, who worked with him to actually test the effect properly. They co-published a 1969 paper in the journal ''Physics Education'', titled, with admirable directness, "Cool?"<ref>[https://iopscience.iop.org/article/10.1088/0031-9120/4/3/312 Cool? — Physics Education, 1969]</ref> The phenomenon has carried Mpemba's name ever since.

== Where it stands now ==
Still genuinely contested, which is the fun part. Proposed explanations range from faster evaporation in the hot sample (less water left to freeze), to differences in convection currents, to dissolved gas content, to supercooling behavior. A 2016 paper in ''Scientific Reports'' directly challenged whether the effect even reproduces reliably under tightly controlled lab conditions at all.<ref>[https://www.nature.com/articles/srep37665 On the Mpemba effect and its statistical significance — Scientific Reports, 2016]</ref> So the honest current answer isn't "here's why it happens," it's "it's been observed for over two thousand years, one teenager forced science to actually look at it, and the mechanism is still being argued about." That's a good ratio of "old" to "unresolved" — the same combination that keeps [[The Voynich Manuscript Nobody Can Read|the Voynich Manuscript]] and [[Kryptos: The Puzzle the CIA Still Hasn't Confirmed|Kryptos]] interesting for completely different fields of study.

== What happened to the kid ==
Mpemba didn't stay in a lab. He went on to become Principal Game Officer for Tanzania's Ministry of Natural Resources and Tourism and vice-chairman of the African Forestry and Wildlife Commission, spending most of his career in wildlife conservation rather than physics. He died on May 14, 2023, at age 73, with a genuine, still-unresolved physics phenomenon carrying his name because he refused to accept "you're wrong" as an answer to something he'd watched happen with his own eyes.

== See also ==
* [[The Myth of Flowing Glass]]
* [[The Voynich Manuscript Nobody Can Read]]

[[Category:Physics & Perception]]
[[Category:The Unknown & Cosmic Questions]]

Steganography

2026-08-07T06:52:02Z

GAMA: Converted from site HTML fragment to wikitext: citations, categories, cross-links

Cryptography hides what a message says. Steganography hides that a message exists at all — an encrypted file announces "there's a secret here," a steganographic one doesn't. The word is Greek for "covered writing," and the technique predates computers by millennia: Herodotus records a Greek noble tattooing a message onto a slave's shaved scalp, then waiting for the hair to grow back over it before sending him through enemy territory.<ref>[https://en.wikipedia.org/wiki/Steganography Steganography — Wikipedia]</ref> The digital version most modern tools use — LSB (least-significant-bit) encoding — swaps the lowest, least visually significant bit of each color channel in an image for a bit of the hidden message, nudging every affected pixel's color by an amount too small to see. It's also fragile on purpose: anything that recompresses the image (a JPEG re-save, an upload to a site that reprocesses photos) scrambles those low bits and destroys the message, which is why LSB tools generally only ever output PNG.

== From pigeons to pixels ==
The technique kept reinventing itself with whatever medium was available. During the siege of Paris in the Franco-Prussian War (1870–71), photographer René Dagron used microfilm to let a single carrier pigeon smuggle huge volumes of correspondence past the blockade — not steganography in the strict "hidden" sense (everyone knew pigeons carried messages), but the direct ancestor of the microdot: a photograph reduced until it's the size of a printed period. Emanuel Goldberg formalized the two-stage reduction process for true microdots in 1925, and by World War II, German intelligence was mailing them concealed on ordinary-looking letters, embedded in the depth of a "." on a typed page, or hidden on couriers' clothing and even inside dolls carried as gifts — the FBI's wartime Special Intelligence Service spent years unraveling espionage rings across Latin America built partly around exactly this delivery method.<ref>[https://www.fbi.gov/history/famous-cases/nazi-saboteurs FBI History: Famous Cases — Nazi Espionage]</ref>

== A case the medium itself gave away ==
Digital image steganography isn't just a classroom exercise — it's shown up in real intelligence work. In 2010, the FBI's decade-long Operation Ghost Stories ended with the arrest of ten Russian intelligence officers living under deep-cover American identities. Part of how they communicated with Moscow was embedding encrypted messages into ordinary-looking images and posting them to public websites, where handlers would download the same images and extract the hidden payload with matching software.<ref>[https://www.justice.gov/archive/opa/pr/2010/June/10-nsd-708.html Ten Alleged Secret Agents Arrested in the United States — US Department of Justice, 2010]</ref> The steganography itself reportedly held up — what didn't hold up was everything around it (tradecraft mistakes, physical surveillance, a cooperating source), which is the pattern in essentially every real steganography case that's ever become public: the hidden-message layer is rarely the point of failure.

== How steganalysis actually catches it ==
Detecting LSB steganography specifically doesn't require breaking any code — it requires noticing that "real" pixel data (from a camera sensor, or ordinary JPEG compression) has statistical noise patterns that are subtly different from noise generated by flipping bits to encode a message. Techniques like chi-square attacks and RS (regular/singular) analysis look for exactly that mismatch: an image that's had data embedded in its low bits tends to have measurably different statistical structure in those bits than one that hasn't, even though nothing is visible to the eye.

== Beyond images ==
LSB-in-images is the easiest version to build and to explain, but it's one technique among many. Audio steganography hides data in similarly inaudible bit-level or frequency-domain changes; video steganography gets an entire extra dimension (time) to spread a payload across; and text steganography ranges from crude (invisible Unicode characters, trailing whitespace patterns) to genuinely clever (choosing between synonymous words or sentence structures according to a hidden bit pattern). Statistical structure that looks meaningless but resists every attempt to explain it away is also the entire premise of the [[The Voynich Manuscript Nobody Can Read|Voynich Manuscript]] — the open question there is whether anyone will ever find the equivalent of a working steganalysis technique for a 600-year-old handwritten codex.

== See also ==
* [[The Voynich Manuscript Nobody Can Read]]
* [[Kryptos: The Puzzle the CIA Still Hasn't Confirmed]]

[[Category:History]]

OK: The Word Everyone Uses and Nobody Can Agree On

2026-08-07T06:52:01Z

GAMA: Converted from site HTML fragment to wikitext: citations, categories, cross-links

Today I learned that the single most-used word in the English language, more universal than "hello," more universal than any brand name, might just be a joke one newspaper editor made up in 1839 and never walked back. Isn't that fun?

== The paper trail ==
On March 23, 1839, the ''Boston Morning Post'' ran a satirical item poking fun at a group called the Anti-Bell-Ringing Society, written (almost certainly) by editor Charles Gordon Greene. Buried in it: "o.k.," short for "all correct," deliberately misspelled as a joke.<ref>[https://en.wikipedia.org/wiki/OK OK — Wikipedia]</ref> In the late 1830s it was genuinely fashionable among younger, educated writers to intentionally misspell a phrase, then abbreviate the misspelling, and use that as slang: the 19th-century equivalent of typing "totes" instead of spelling out "totally." "OK" started life as one entry in that fad, "oll korrect," shortened to two letters.

== How a joke became permanent ==
Most of that era's deliberately-misspelled abbreviations died with the fad. "OK" didn't, because it got picked up somewhere with actual reach: the 1840 U.S. presidential campaign. Supporters of incumbent Martin Van Buren, nicknamed "Old Kinderhook" after his New York hometown, adopted "O.K." as a rallying cry, forming "O.K. Clubs" that doubled the joke as a reference to Van Buren's own initials.<ref>[https://en.wikipedia.org/wiki/Martin_Van_Buren#1840_presidential_campaign Martin Van Buren: 1840 presidential campaign — Wikipedia]</ref> Van Buren lost the election. The word didn't need him to win, though. It had a second, unrelated adoption boost waiting in telegraph operators, who needed a fast way to confirm a message had gone through cleanly, and two letters beat spelling out "all correct" every time.

== The theories that didn't hold up ==
Because "OK" is so short and so old, it's attracted a pile of competing origin stories over the decades: that it comes from Andrew Jackson misspelling "all correct" on official documents, from the Choctaw word ''okeh'' ("it is so"), from the Wolof phrase ''waw kay'', or from "Aux Cayes," a Haitian port whose rum was supposedly good enough to be shorthand for quality. None of these have the thing the Boston Morning Post story has: an actual dated, printed, still-readable source. That's the work of Allen Walker Read, a Columbia University English professor who spent years in the 1960s tracing "OK" through 19th-century newspaper archives and settled the debate about as definitively as etymology ever gets settled. The joke-abbreviation slang fad, not any of the folk stories, is the version with the paper trail. Allan Metcalf, a linguist at MacMurray College, later turned Read's research into a full book, ''OK: The Improbable Story of America's Greatest Word'' (2010).<ref>[https://global.oup.com/academic/product/ok-9780195377934 OK: The Improbable Story of America's Greatest Word — Oxford University Press]</ref> His framing is basically: this should have died as a bad pun within a year, and instead it became the most recognized word on the planet.

The same "the boring, documented answer beats the fun folk story" shape shows up in [[Dord: The Word the Dictionary Invented by Accident|Dord's]] story too — a word can end up real for reasons that have nothing to do with anyone intending it.

== See also ==
* [[Dord: The Word the Dictionary Invented by Accident]]
* [[Where "Foo" and "Bar" Came From]]

[[Category:Language, Communication, & Internet Culture]]
[[Category:History]]

Digital Preservation & Link Rot

2026-08-07T06:52:00Z

GAMA: Converted from site HTML fragment to wikitext: citations, categories, cross-links

The average webpage doesn't survive very long — a meaningful fraction of links go dead within a few years. The Internet Archive's Wayback Machine crawls and stores snapshots of pages over time specifically to fight this, and anyone can manually save a page's current state to it. If a page matters enough to cite or rely on, it's worth archiving before it isn't there anymore.

== How bad, specifically ==
A 2024 Pew Research Center study tracking a broad sample of webpages found that 38% of pages that existed in 2013 were no longer accessible a decade later, and roughly a quarter of all pages that existed at any point between 2013 and 2023 had gone dead by the end of that window.<ref>[https://www.pewresearch.org/internet/2024/05/17/when-online-content-disappears/ When Online Content Disappears — Pew Research Center, 2024]</ref> The same study found 54% of Wikipedia pages contain at least one dead reference link, 23% of news pages do, and roughly half of the URLs cited in US Supreme Court opinions no longer resolve to their original content. A separate, independently run study of nine years of published links found more than 66% dead by the end of that period.<ref>[https://www.hbs.edu/faculty/Pages/item.aspx?num=53419 The Half-Life of Links — Harvard Business School Working Paper]</ref> None of these numbers agree exactly, but they all land in the same range: link rot isn't a rare edge case, it's closer to the default outcome for a link nobody actively maintains.

== Two different failure modes ==
"Link rot" usually gets used for two genuinely different problems. '''Reference rot''' is the simple case — the URL returns a 404, the domain expired, the site shut down entirely. '''Content drift''' is quieter and arguably worse: the URL still works, but the page behind it has been edited, replaced, or repurposed since it was cited, so a link that looks fine is silently pointing at something other than what the citing author actually referenced. A snapshot service defends against both; a live URL defends against neither.

== The tools, beyond the Wayback Machine ==
The Internet Archive's Wayback Machine is the largest and longest-running general-purpose web archive, but it's not the only option.<ref>[https://web.archive.org/ Wayback Machine — Internet Archive]</ref> '''archive.today''' (also seen as archive.ph/archive.is under various mirror domains) takes a single, permanent, on-demand snapshot the moment you submit a URL — useful specifically because it can't be affected by a site's <code>robots.txt</code> asking crawlers to stay away, which the Wayback Machine has historically respected. '''Perma.cc''', built and run by a consortium of law libraries, exists specifically to solve the Supreme Court citation problem above: it gives academic and legal citations a permanent link that won't rot.<ref>[https://perma.cc/ Perma.cc]</ref>

== The practical habit ==
Archiving a page takes seconds and costs nothing; recovering a page that's genuinely gone (not indexed anywhere, never archived) is often simply impossible. The asymmetry is the whole argument for doing it before you need it, not after. The same asymmetry, on an institutional scale, is the entire story behind every burned library and every archive lost to a single fire or flood: a record that only ever existed in one place, gone the moment that one place is.

== See also ==
* [[Why There's No Paywall-Bypass Tool Here]]
* [[OSINT (Open Source Intelligence)]]
* [[Reverse Image Search & Metadata]]

[[Category:History]]

The Voynich Manuscript Nobody Can Read

2026-08-07T06:52:00Z

GAMA: Converted from site HTML fragment to wikitext: citations, categories, cross-links

Somewhere in Yale's rare book library sits about 240 pages of handwriting nobody on Earth can read, illustrated with plants that don't appear to match any known real species.

== What it physically is ==
The manuscript lives at Yale's Beinecke Rare Book & Manuscript Library, catalogued as MS 408. It's named after Wilfrid Voynich, the rare-book dealer who acquired it in 1912, not an author, since nobody knows who wrote it.<ref>[https://en.wikipedia.org/wiki/Voynich_manuscript Voynich manuscript — Wikipedia]</ref> It's a real, physical, handwritten codex on vellum, filled with an unknown script in a fluent, confident hand (whoever wrote it wasn't struggling or improvising letterforms) alongside illustrations sorted loosely into herbal, astronomical, cosmological, pharmaceutical, and recipe-like sections.

== When it's actually from ==
Radiocarbon dating performed by the University of Arizona in 2009 narrowed the vellum's age to 1404–1438, placing it firmly in the early Renaissance and ruling out older guesses that had floated around for decades.<ref>[https://library.medieval.org/voynich-radiocarbon University of Arizona radiocarbon dating of the Voynich manuscript, 2009]</ref> One honest caveat worth keeping in mind: that dates the parchment, not necessarily the ink on it. It was historically possible, if unusual, to write on older stock. Nobody's found real evidence that happened here, but it's the kind of gap a careful source doesn't gloss over.

== The decoding graveyard ==
Since Voynich's acquisition in 1912, cryptographers, linguists, historians, and enthusiastic amateurs have thrown real effort at this, including codebreakers with WWII- and Cold War-era training, the sort of people who'd cracked actual military ciphers for a living. None of it has produced a decoding that held up under scrutiny. The Beinecke reportedly still gets a steady trickle of emails from people convinced they've finally solved it. So far, every single one has fallen apart on closer inspection.

== The hoax theory, and its own problem ==
In 2003, Keele University computer scientist Gordon Rugg proposed that it might not be a code at all. Using a Renaissance-era encryption tool called a Cardan grille combined with tables of meaningless syllables, he showed a single person could hand-generate a document with Voynichese's basic statistical texture in a matter of months, no hidden message required.<ref>[https://www.cs.keele.ac.uk/staff/g_rugg/voynich/ Gordon Rugg — Voynich manuscript research, Keele University]</ref> He pinned the manuscript on Edward Kelley, a 16th-century Englishman with a documented history of running cons, allegedly to fleece Holy Roman Emperor Rudolf II. It's a genuinely clever demonstration with a genuinely awkward hole in it: Cardano's grille wasn't invented until roughly a century after the manuscript's vellum was carbon-dated to, and Rugg's simplest version of the method still can't reproduce the text's more complex statistical patterns. Neither problem has been fully resolved by supporters of the theory.

== Why that's the actually interesting part ==
It would almost be less strange if the hoax theory just won outright. Plenty of "codes" throughout history have turned out to be exactly that. What keeps the Voynich manuscript sitting differently is that whichever way you lean, the writing has real statistical structure: consistent word-length patterns, letter-frequency behavior that resembles genuine language rather than random scribble, patterns Rugg's own method struggles to fully fake. Whatever it is, it wasn't produced carelessly. It just still hasn't given anything up — the same stalemate [[Kryptos: The Puzzle the CIA Still Hasn't Confirmed|Kryptos]]'s K4 panel has been sitting in for over three decades now, just with a different medium and a much smaller archive to draw suspects from.

== See also ==
* [[Kryptos: The Puzzle the CIA Still Hasn't Confirmed]]
* [[Steganography]]
* [[The Mpemba Effect: When Hot Water Freezes First]]

[[Category:The Unknown & Cosmic Questions]]
[[Category:History]]

The Odds a UUID Ever Repeats

2026-08-07T06:51:59Z

GAMA: Converted from site HTML fragment to wikitext: citations, categories, cross-links

A version-4 UUID has 128 bits total, but not all of them are random: 6 are fixed by the format itself (4 bits mark it as version 4, 2 more mark the variant), leaving 122 genuinely random bits.<ref>[https://en.wikipedia.org/wiki/Universally_unique_identifier Universally unique identifier — Wikipedia]</ref> Birthday-paradox math — the same math behind "how many people need to be in a room before two share a birthday" — says you'd need to generate roughly 2.71 quintillion (2.71 × 10<sup>18</sup>) version-4 UUIDs before the odds of any two colliding crossed 50%. Every UUID this site has ever generated, and every one it ever will, is functionally guaranteed unique — not because the math forbids a collision, but because the number of UUIDs that would need to exist for one is bigger than most estimates of the total data ever stored by humanity.

== Not all UUIDs are random ==
"UUID" names a format (128 bits, that familiar <code>8-4-4-4-12</code> hyphenated hex layout), not a single method for filling it. The version number — a single hex digit baked into the string itself — says how it was actually generated, and the versions in real use trade off very differently:

* '''v1 (timestamp + MAC address).''' Encodes the generating machine's network MAC address and a high-precision timestamp directly into the UUID. Guaranteed unique per machine as long as its clock doesn't move backward, but it leaks two things a random UUID doesn't: roughly when the record was created, and a hardware identifier for the machine that created it.
* '''v4 (random).''' What most tools generate by default today. No embedded information beyond "this was made to be a UUID."
* '''v5 (name-based, SHA-1).''' Deterministic: the same input name plus the same namespace always produces the same UUID. v3 is the same idea with the older, weaker MD5 instead of SHA-1, kept mainly for backward compatibility.
* '''v7 (Unix-timestamp-prefixed, random tail).''' A newer addition to the standard, formalized in RFC 9562 (2024), built specifically to fix a real operational problem with v4: database indexes built on a v4 primary key fragment badly over time because insert order has nothing to do with sort order, whereas a v7 UUID sorts roughly chronologically by creation time the way an auto-incrementing integer does.<ref>[https://datatracker.ietf.org/doc/html/rfc9562 RFC 9562: Universally Unique IDentifiers (UUIDs) — IETF, 2024]</ref>

== The alternative that isn't a UUID at all ==
Before v7 existed, teams that wanted both "sortable like a timestamp" and "collision-resistant like a UUID" often reached for a non-standard format instead — ULID (Universally Unique Lexicographically sortable ID) and Twitter's Snowflake ID are the two most common. Both encode a timestamp plus randomness or a machine/sequence counter into a fixed-width identifier, sort naturally, and predate v7's standardization.

== What actually goes wrong in practice ==
Real UUID collisions in production systems are, unsurprisingly, essentially never caused by the random-number math failing — they're caused by something upstream of the math: a bad or predictable random number generator, a v1 UUID generated on a virtual machine that cloned its MAC address, or — far more common than either — a caching or retry bug that generates the same UUID twice by accident rather than the generator itself repeating. If two UUIDs in a real system ever collide, the math above is not where to start debugging.

The underlying idea — that a few dozen bits of real entropy is enough to make a collision practically impossible — is the same math [[Browser Fingerprinting]] runs in reverse: there, a much smaller entropy budget (tens of bits, not 122) is enough to make a browser practically identifiable instead.

== See also ==
* [[Browser Fingerprinting]]

[[Category:Mathematics, Probability, & Game Theory]]

On Time, and Why It's Hard

2026-08-07T06:51:58Z

GAMA: Converted from site HTML fragment to wikitext: citations, categories, cross-links

'''The epoch.''' Machines count seconds from January 1, 1970, 00:00:00 UTC, and everything since is an incrementing integer — Unix time.<ref>[https://en.wikipedia.org/wiki/Unix_time Unix time — Wikipedia]</ref> ISO 8601 is the closest thing to a universal human-readable format for the same information — see GAMAYUN's own Date Stamp Recon tool for the actual converter.

'''The Year 2038 Problem.''' A lot of systems store that Unix timestamp as a signed 32-bit integer. That format runs out of road on January 19, 2038, at 03:14:07 UTC — one second later, the count wraps back around to a large negative number, which most software reads as December 1901.<ref>[https://en.wikipedia.org/wiki/Year_2038_problem Year 2038 problem — Wikipedia]</ref> Same failure mode as the Y2K panic, different root cause: Y2K was a storage-shortcut problem (two-digit years), this is an arithmetic-overflow problem (a fixed-width integer running out of bits). Most modern 64-bit systems already sidestepped it. Plenty of embedded and legacy 32-bit systems have not, and won't notice until they do.

'''Timezones ruin everything.''' They aren't clean hour-wide slices — India runs UTC+5:30, Nepal runs UTC+5:45, and some Pacific islands hop entire calendar days depending which side of the International Date Line their government prefers this decade.<ref>[https://en.wikipedia.org/wiki/UTC%2B05:45 UTC+05:45 (Nepal Time) — Wikipedia]</ref> Add daylight saving time (which not every country observes, and which some countries have started and then abandoned more than once) and "what time is it in the other office" stops being arithmetic and starts being a lookup against a constantly-revised political database.

'''Leap seconds.''' Earth's rotation is slowing down, almost imperceptibly, so every so often an official leap second gets inserted to keep clock time lined up with the actual planet.<ref>[https://en.wikipedia.org/wiki/Leap_second Leap second — Wikipedia]</ref> In 2012, a leap second exposed a bug in the Linux kernel's futex implementation that pushed CPU usage on affected servers to 100% — Reddit, Mozilla, and Yelp all had a bad night over one extra second.<ref>[https://www.wired.com/2012/07/leap-second-bug-wreaks-havoc-with-java-linux/ How a Leap Second Wreaked Havoc — Wired, 2012]</ref> Several major tech companies now "smear" the leap second across an entire day of very slightly slower ticks instead of inserting it all at once, specifically so this doesn't happen again. The bugs got annoying enough, often enough, that on November 18, 2022, the world's governments voted at the General Conference on Weights and Measures to retire the leap second entirely by 2035 — after which UTC will simply be allowed to drift out of sync with the Earth's actual rotation rather than being corrected.<ref>[https://www.nist.gov/news-events/news/2022/11/no-more-leap-seconds-until-least-2035 No More Leap Seconds Until at Least 2035 — NIST, 2022]</ref> Russia was the one notable vote against it, reportedly over concerns about its own satellite navigation systems' dependence on the current scheme.

'''How your computer's clock actually stays right.''' Left alone, a computer's internal clock drifts — cheap quartz oscillators gain or lose real seconds per day. NTP (Network Time Protocol), designed by David L. Mills starting in the 1980s, is the quiet background process that fixes this: your machine periodically asks a server it trusts what time it actually is and nudges its own clock toward that answer, smoothly rather than jumping.<ref>[https://en.wikipedia.org/wiki/Network_Time_Protocol Network Time Protocol — Wikipedia]</ref> Those servers form a hierarchy of "strata" — stratum 0 is an actual physical reference clock (atomic or GPS-synchronized), stratum 1 servers are directly connected to one of those, and everything else is a few hops further down the chain.

'''Why cron starts the week on Sunday.''' Cron's day-of-week field runs 0–6 with Sunday as 0, not Monday. It's inherited from POSIX C's <code>struct tm</code>, where <code>tm_wday</code> has always used the same 0-for-Sunday convention, which itself traces back to older US calendar conventions treating Sunday as the start of the week. (ISO 8601 goes the other way entirely for its own week-numbering system — Monday is day 1, and "week 1" of a year is defined as whichever week contains that year's first Thursday.)<ref>[https://en.wikipedia.org/wiki/ISO_8601#Week_dates ISO week date — Wikipedia]</ref>

== See also ==
* [[Calendars Are Political]]
* [[The Odds a UUID Ever Repeats]] — a different kind of "the numbers say this basically never happens"
* [[Browser Fingerprinting]] — another article built on bits of entropy, just applied to identity instead of timestamps

[[Category:Physics & Perception]]

Kryptos: The Puzzle the CIA Still Hasn't Confirmed

2026-08-07T06:51:58Z

GAMA: Converted from site HTML fragment to wikitext: citations, categories, cross-links

There's a sculpture standing on actual CIA grounds that stumped the agency's own cryptanalysts for years. Then, in 2025, its answer reportedly turned up in a completely different archive, and the man who built it still won't just tell everyone.

== The sculpture ==
Kryptos is a copper-and-granite installation by artist Jim Sanborn, dedicated on November 3, 1990, at the George Bush Center for Intelligence, CIA headquarters, Langley, Virginia. Sanborn collaborated with Ed Scheidt, a retiring CIA cryptographic officer, to design the actual encryption.<ref>[https://en.wikipedia.org/wiki/Kryptos Kryptos — Wikipedia]</ref> Carved into the sculpture's curved copper screen are four separate encoded passages, roughly 869 characters combined, labeled K1 through K4.

== Three down ==
K1 and K2 use a Vigenère cipher run through a keyed, non-standard alphabet ("KRYPTOSABCDEFGHIJLMNQUVWXZ"), with the keywords PALIMPSEST and ABSCISSA respectively. K3 is a transposition cipher. All three were actually solved quietly, more than once, before anyone outside a small circle knew it. An NSA team led by Ken Miller cracked them internally in late 1992, and CIA analyst David Stein solved them by hand in 1998, but neither disclosure went public. Credit usually goes to computer scientist Jim Gillogly, who solved and published all three independently in 1999, making him the first to actually tell anyone. K2's decoded text is its own small reward. It includes a set of geographic coordinates pointing to a spot roughly 174 feet from the sculpture itself.

== The one that wouldn't crack ==
K4 is short, about 97 characters, and that's part of what makes it brutal. Its cipher method has never been publicly confirmed, and it resisted every serious public attempt for more than three decades, making it one of the most famous unsolved ciphertexts in the world, sitting in plain view of an intelligence agency that specializes in exactly this.

== The 2025 wrinkle ==
In September 2025, journalists Jarett Kobek and Richard Byrne reported finding documents in the Smithsonian's own archives that appeared to contain K4's plaintext.<ref>[https://www.wired.com/story/kryptos-cia-sculpture-k4-code-smithsonian/ Has the CIA's Kryptos Sculpture Finally Been Cracked? — Wired, 2025]</ref> Sanborn didn't confirm it publicly. Instead, that November, he auctioned off his personal archive, his own working notes on the piece, for just over $962,000, well above its estimated value, rather than simply releasing the answer.<ref>[https://www.rrauction.com/ RR Auction — Jim Sanborn Kryptos archive]</ref> He's also stated that a fifth message, K5, exists and won't surface until K4 is publicly acknowledged as solved. As of this writing, nothing has been officially confirmed by Sanborn or the CIA, so "unsolved" is still technically the accurate word for K4. It's just not a clean one anymore.

== See also ==
* [[The Voynich Manuscript Nobody Can Read]]
* [[Steganography]]

[[Category:The Unknown & Cosmic Questions]]
[[Category:History]]

OSINT (Open Source Intelligence)

2026-08-07T06:51:57Z

GAMA: Converted from site HTML fragment to wikitext: citations, categories, cross-links

Gathering information from publicly available sources — search engines, public records, social media, breach-notification services — rather than hacking or stealing anything. A breach checker like Have I Been Pwned is OSINT. A site that sells you the actual leaked passwords is not OSINT, it's trafficking in stolen data. A huge amount of "hacker tool" content online blurs "checks if you were exposed" together with "lets you access what was stolen." They're not the same thing legally or ethically, and knowing the difference is the whole skill.

== Older than the term ==
"OSINT" as a phrase is generally traced to the US military and intelligence community, gaining common use through the late 1980s and early 1990s.<ref>[https://en.wikipedia.org/wiki/Open-source_intelligence Open-source intelligence — Wikipedia]</ref> The practice it names is much older than the name itself — military intelligence services have drawn on newspapers, journals, press clippings, and public radio broadcasts for generations, and both sides of the American Civil War openly read each other's newspapers for troop movements and supply information. What changed with the internet wasn't the concept, it was the volume and speed: a technique that used to mean waiting for tomorrow's paper now means a public social media post can be geolocated and cross-referenced within minutes of going live.

== A real investigation, not a hypothetical ==
In 2014, the volunteer-run investigative outlet Bellingcat took on the downing of Malaysia Airlines Flight 17 over eastern Ukraine using nothing but open sources: photos and videos posted publicly online, satellite imagery, and mapping tools like Google Earth. By cross-referencing the visible background of a video showing a Buk missile launcher on a flatbed truck — specific buildings, road markings, terrain — against satellite imagery, the team traced its route back to a Russian military base near Kursk.<ref>[https://www.bellingcat.com/news/uk-and-europe/2014/11/08/origin-of-the-separatists-buk-a-bellingcat-investigation/ Origin of the Separatists' Buk: A Bellingcat Investigation]</ref> Four years later, in 2018, the Dutch-led international Joint Investigation Team's own official findings confirmed the same conclusion Bellingcat had reached with public information alone, before any government investigation had concluded.<ref>[https://www.jit-mh17.com/ Joint Investigation Team MH17]</ref> It's one of the clearest demonstrations that OSINT isn't a lesser, amateur version of "real" intelligence work — done rigorously, it can out-pace agencies with classified resources, because the constraint was never access, it was whether anyone put in the work to actually cross-reference what was already public.

== A rough version of the actual method ==
Professional OSINT work generally follows something like the intelligence cycle: define what you're actually trying to find out, collect from multiple independent public sources, cross-reference and corroborate rather than trusting any single source, and only then draw a conclusion — explicitly flagging what's confirmed versus what's merely plausible. The corroboration step is the one amateur "internet detective" efforts skip most often, and it's the one that turns a plausible-looking coincidence into either a real finding or a debunked one — the same corroboration gap that undoes most amateur attempts at [[Reverse Image Search & Metadata|reverse-image verification]] too.

== Where the legal and ethical line actually sits ==
Looking up what's already public is generally legal — it's the digital equivalent of reading a published newspaper. The line moves once a method stops being "look at what's already visible" and starts being "get access to something that wasn't meant to be visible": logging into an account that isn't yours, using a leaked-credential database to test other accounts, or treating [[Google Dorking]]'s results (technically public, functionally exposed by accident) as an invitation rather than a finding to report responsibly. Most OSINT tools are ethically neutral — a reverse image search doesn't care whether you're debunking a hoax or stalking someone — which is exactly why the intent and the target matter more than the tool.

== See also ==
* [[Google Dorking]]
* [[Reverse Image Search & Metadata]]
* [[Browser Fingerprinting]]
* [[Digital Preservation & Link Rot]] — the archived-page side of the same verification habit

[[Category:Psychology, Society, & Cognitive Biases]]

Where "Foo" and "Bar" Came From

2026-08-07T06:51:56Z

GAMA: Converted from site HTML fragment to wikitext: citations, categories, cross-links

The two most-repeated origin stories — a WWII acronym, and a 1930s comic strip — both check out, and the more surprising finding (documented in an actual IETF standards document written specifically to settle this, RFC 3092) is that the comic strip almost certainly came first.<ref>[https://datatracker.ietf.org/doc/html/rfc3092 RFC 3092: Etymology of "Foo" — IETF]</ref>

== The comic strip nobody expects ==
"Foo" shows up in print well before any wartime acronym: in ''Smokey Stover'', a wildly popular American newspaper comic strip that ran from roughly 1930 to 1952, where the title character (a firefighter) used "foo" as a nonsense catchphrase constantly.<ref>[https://en.wikipedia.org/wiki/Smokey_Stover Smokey Stover — Wikipedia]</ref> The strip was popular enough to spin off over 500 unofficial "Foo Clubs" and work its way into contemporary songs — it was, briefly, just a piece of mainstream American pop culture, the way a meme phrase spreads today.

== Where FUBAR actually fits ==
FUBAR ("[messed] up beyond all recognition/repair") is real WWII-era US Army slang, and "foo fighter" was genuinely used by WWII radar operators for unexplained blips — the same term that later got attached to UFO sightings, and, much later, to the rock band.<ref>[https://en.wikipedia.org/wiki/Foo_fighter Foo fighter — Wikipedia]</ref> For years the assumed story was that "foo" in computing came from an informal shortening of FUBAR. RFC 3092's actual research points the other way: contemporary accounts tie WWII "foo fighter" directly back to the already-famous comic strip, not the other way around, and there's a real possibility FUBAR itself picked up its spelling under the influence of "foo" (with a nod to the similar-sounding German word ''furchtbar'', "terrible") rather than "foobar" being a sanitized version of FUBAR. In short: the acronym almost certainly didn't invent the word, it more likely borrowed it.

== How it got into code ==
The jump from slang to programming is credited to MIT's Tech Model Railroad Club, around 1960 — the same group whose model-railroad-control jargon fed heavily into what became early hacker culture and, eventually, the ''Jargon File'' itself.<ref>[http://catb.org/jargon/html/F/foo.html "foo" — The Jargon File]</ref> "Foo" needed a placeholder pair once code examples needed a second arbitrary thing, and "bar" — already sitting right next to it via FUBAR — was the obvious choice. From there the family just grew: <code>baz</code>, <code>qux</code>, <code>quux</code>, and others exist purely because someone needed a third, fourth, or fifth interchangeable nothing-word and reached for whatever sounded like it belonged in the same set. None of them mean anything. That's the entire point of a metasyntactic variable — a name that signals "this could be replaced with literally anything" precisely by not meaning anything itself.

Foo, bar, and dord all end up in the same place, incidentally: a word that only exists because of how it was used, not because anyone set out to define it. See [[Dord: The Word the Dictionary Invented by Accident]] for the version where that happened entirely by accident, inside a real published dictionary.

== See also ==
* [[OK: The Word Everyone Uses and Nobody Can Agree On]]
* [[Dord: The Word the Dictionary Invented by Accident]]

[[Category:Language, Communication, & Internet Culture]]