Category: Ancient Web

The most ancient times of the internet: 90s, early 2000s.

Posted on by

Ancient Web: Zompist Tried to Prove English Spelling Has Rules

English spelling has spent centuries earning its reputation as a practical joke, and Mark Rosenfelder’s old Zompist page responds with the deeply Internet-era reaction: fine, let’s model the damn thing.

Visit Zompist’s English spelling page

The page, dated 2000, is titled “Hou tu pranownse Inglish.” Its argument is not that English spelling is elegant. It is that the system contains far more regularity than people usually give it credit for.

Rosenfelder lays out sound values, dialect assumptions, letter combinations, vowel rules, consonant rules, exceptions, historical leftovers, and the famous troublemakers such as gh and ough.

He also takes a swing at George Bernard Shaw’s famous “ghoti” joke, arguing that the proposed pronunciation only works by ignoring where English spelling rules actually permit those letter values.

The page then does something that makes it especially interesting as an old Web artifact: it turns the argument into a computational test.

Make the computer pronounce it

Rosenfelder assembled a sample lexicon of more than 5,000 English words and a set of rules for his Sound Change Applier. According to the page, the rules generated 59 percent of pronunciations perfectly and 85 percent either perfectly or with only relatively minor errors.

That does not make English spelling sane.

It does demonstrate that “English spelling is completely random” is not a particularly useful explanation.

The page is also full of old-browser archaeology. It discusses Unicode support as something the reader’s browser may or may not handle correctly and explains notation choices partly in terms of what HTML could reliably display.

That mixture is classic early specialist Web publishing: a serious subject, a personal voice, hand-built technical notation, downloadable data, and the assumption that an interested reader will happily scroll for a very long time.

Modern search results tend to chop questions like this into isolated answers: why is ough weird, why is knight spelled that way, why does c have two sounds?

Rosenfelder tried to put the machinery in one place.

CacheRat’s 1,967 Ancient Web Domains research list includes pages like this because old personal sites often preserve the full argument instead of optimizing each paragraph into a separate answer box.

English spelling is still a mess.

At least this mess comes with documentation.

Read the full spelling system

Posted on by

Ancient Web: Arkm’s World Turns Navigation Into Storytelling

Arkm’s World does not greet you with an About page, a menu, or a paragraph explaining the project. It drops you into a place.

Enter Arkm’s World

The homepage reads like a fragment of fantasy travel writing: the narrator is sailing through a place called Shadow, notices that the sky no longer has pterodactyls, and prepares to make one more “shift.” A moon-on-water image sits underneath. The next action is not presented as navigation so much as continuation.

That is the whole trick.

The website treats hyperlinks as movement through a fictional environment rather than as ordinary site controls. Instead of “Home / About / Gallery,” the visitor follows the story because clicking is how the story advances.

Arkm’s World is hosted on Neocities, the service that deliberately revived the idea of personal homepages after the collapse of GeoCities. It also has a small place in that revival’s own history: a NeoCities neighborhood chronology records Arkm’s World as the first site to request listing when the recreated Area51 neighborhood opened in January 2021.

Hypertext can still be a medium

The Web spent years standardizing navigation for good reasons. Consistency helps people use sites. But something was lost when every site began behaving like the same application.

Early Web experiments often treated the hyperlink itself as an artistic device. A link could be a door, a decision, a joke, a trap, a hidden room, or an unexplained word that only made sense after you clicked it.

Arkm’s World belongs to that tradition.

It is not technically complicated. That is part of the point. The atmosphere comes from writing, imagery, pacing, and the decision not to explain everything immediately.

For modern webmasters, this is a reminder that not every page needs to optimize for instant comprehension. If the site is art, fiction, or exploration, a little disorientation can be intentional.

Neocities has become one of the places where this older understanding of the Web survives: personal sites as worlds instead of profiles.

CacheRat’s 1,967 Ancient Web Domains research list includes both genuinely old sites and newer projects that deliberately inherit old-Web ideas. Arkm’s World is the latter.

It is small, strange, and uninterested in behaving like an app.

That is enough reason to keep it.

Enter Arkm’s World again

Posted on by

Ancient Web: Betty’s Graphics Keeps 16,713 Pieces of Web Decoration Alive

The old Web had an endless appetite for tiny pictures. Spinning mailboxes, flaming dividers, sparkly stars, tiled backgrounds, cartoon food, zodiac icons, animated buttons—if a page had three empty pixels, somebody eventually found a GIF to put there. Betty’s Graphics is keeping that ecosystem alive by the thousand.

Visit Betty’s Graphics

The Neocities site describes itself as the largest hand-curated graphics archive on the platform and currently lists 16,713 images. Its categories include animals, backgrounds, characters and people, food, holidays, objects, plants, gadgets, Web graphics, and zodiac signs.

The creator is also admirably clear about authorship: these graphics were collected and shared, not presented as original artwork. The entire archive is offered as a downloadable collection in addition to being browsable online.

This is a modern preservation project, not a claim that Betty’s Graphics itself came out of 1996. What it preserves is the visual language of the personal Web—especially the millions of tiny assets that were copied from page to page, free-graphics site to free-graphics site, often until nobody had the slightest idea where they started.

The Web used to have a junk drawer

Modern web design has become much more consistent. That is good if you need to buy a refrigerator or reset a password. It is less exciting if you miss the era when visiting someone’s homepage could result in immediate exposure to a dancing skull, six tiled roses, a glittering cursor, and a button announcing that the page was best viewed at a resolution your monitor has not used since the Clinton administration.

Graphics archives made that possible.

People did not need Photoshop skills to give a homepage an identity. They found a background tile, a divider, a few icons, maybe an “under construction” image, copied them into a directory, and built something unmistakably theirs. The result was frequently ugly. It was also frequently memorable.

Betty’s Graphics matters because these assets are individually trivial and collectively historical. A single 2-kilobyte star GIF tells you almost nothing. Sixteen thousand pieces begin to show the shared visual vocabulary people used to decorate independent websites.

The archive is not sitting unused, either. Other Neocities pages continue to credit or recommend Betty’s Graphics as a source for their own designs. In other words, this is not merely a museum. The old graphics are being put back into circulation.

That is exactly how living preservation works. Save the files, document what they are, make them easy to browse, and let somebody build something weird with them again.

CacheRat’s 1,967 Ancient Web Domains research list tracks the old sites. Betty’s Graphics preserves some of the tiny pieces those sites were made from.

Browse all 16,713 pieces at Betty’s Graphics

Posted on by

Ancient Web: AssociativeMusic.com Tried to Turn Found Noise Into a Musical Philosophy

A rat’s heartbeat, television static, wind, empty buildings, and noises too quiet for ordinary ears are not the normal ingredients of a pop-music manifesto. That was more or less the point of AssociativeMusic.com.

Visit AssociativeMusic.com

The site was associated with New York producer Nick Koenig, better known as Hot Sugar, and with a process he called Associative Music. The idea was not simply to sample unusual sounds. It was to use sounds that already carried some emotional, physical, or cultural association and turn them into musical material.

Contemporary descriptions of Koenig’s work talk about heartbeats, television static, environmental noise, resonant spaces, birds, wind, storms, and objects recorded outside conventional studios. Instead of treating those noises as background contamination, he treated them as instruments waiting to be extracted.

The old site itself is difficult to inspect reliably now, but fragments of its writing survive through links and quotations elsewhere on the Web. One passage argued that recordings should not be limited to the human ear’s normal range. Sounds beyond ordinary hearing, it suggested, could be amplified, slowed, or shifted into a range where people could experience them.

That is a wonderfully Internet-era proposition: if nature refuses to make a convenient WAV file, fix nature.

A website was part of the idea

AssociativeMusic.com mattered because the site gave the concept somewhere to exist outside individual releases.

Press coverage from the early 2010s repeatedly described Koenig as the founder of Associative Music and pointed readers back to the domain. The site acted as a manifesto, definition, and reference point for a process that was otherwise difficult to describe in one sentence.

That was a common role for independent artist websites. A record label page could sell the release. A social profile could announce it. A personal domain could explain the weird theory behind it.

The distinction is important because platforms are poor homes for ideas that need context. A timeline rewards the new post. A website can keep the explanation sitting in the same place for years.

Even after the original page becomes unreliable, its language can continue circulating through old Tumblr posts, artist bios, press releases, event listings, and music archives. That secondary trail is now part of the preservation record.

For web archaeologists, this is another example of a site whose cultural footprint is larger than the amount of HTML still easily reachable at the original address.

The larger 1,967 Ancient Web Domains research list contains many domains like this: small enough to disappear, distinctive enough that other people kept quoting them.

AssociativeMusic.com survives because the idea escaped the server.

Try AssociativeMusic.com

Posted on by

Ancient Web: The Poem Tree Preserved Metered Poetry With AOL-Era Charm

The Poem Tree is the kind of literary website that could only have grown from somebody caring more about the subject than about redesign cycles.

Visit The Poem Tree

Edited by Caleb Murdock with Rhina P. Espaillat listed as editorial advisor, the site describes itself as an online anthology with an emphasis on metered poetry—traditional and modern—with some free verse included as well. Its stated purpose is to present formalist poets to the public and help popularize metrical verse after a century dominated by free verse.

The collection ranges widely. Traditional names include John Donne, George Herbert, Keats, Shelley, Shakespeare, Tennyson, Whitman, and Wordsworth. Early-modern and twentieth-century sections include Robert Frost, Thomas Hardy, Wilfred Owen, Ezra Pound, Sara Teasdale, and Yeats. Contemporary formalists include Dana Gioia, A.E. Stallings, X.J. Kennedy, Rhina Espaillat, and others.

The site is also more than a pile of poems. It includes essays, a preface, notes for poets, recommended reading, an editor’s section, and material about writing craft. That gives it the feel of a small literary magazine fused with a reference anthology.

Then there is the technical archaeology.

The viewing instructions still advise AOL users to turn off graphics compression if they want the artwork to look right. That single sentence dates the page more effectively than any footer could. It comes from an era when a webmaster had to worry not just about browsers, but about online services silently mangling images to make dial-up pages load faster.

The site says it is no longer being expanded and shows a last update in February 2020, but that does not make it obsolete. Poetry does not expire because the CMS market moved on. The page remains useful precisely because the poems, editorial choices, and internal organization are stable.

For modern site owners, The Poem Tree is a reminder that a durable niche archive can outlive many technically superior publishing systems. Its value comes from selection and structure, not from software novelty.

CacheRat’s Ancient Web research list keeps turning up these small, stubborn archives: one person or a few people deciding a subject deserves a permanent place on the Web, then actually building one.

The Poem Tree feels less like a platform and more like somebody’s carefully maintained bookshelf. That is exactly the point.

Browse The Poem Tree

Posted on by

Ancient Web: KeyHut Still Turns Old PCs Into Cash Registers

KeyHut’s cash-register software is a reminder that an old computer does not become useless just because Silicon Valley has stopped admiring it.

Visit the KeyHut POS page

The program, maintained by Dale Harris, turns a DOS or Windows PC into a point-of-sale terminal. It handles the things an actual small shop needs: receipts, inventory, taxes, employee access, scanners, cash drawers, customer information, sales history, ticket sales, scales, networking, and assorted retail hardware.

What it deliberately does not require is almost more interesting.

The documentation says the program does not need a mouse. It does not process credit cards itself. It does not need a modern graphical interface. Keyboard navigation is treated as an advantage because cash-register work consists largely of repeating known actions quickly.

The site is equally direct about price: the software is free.

Harris’s donation pitch asks satisfied users for five dollars, donated equipment, a link from their Web site, or helper software. It reads less like a software company funnel and more like somebody keeping a useful machine alive because people keep using it.

Software that refuses to become obsolete

KeyHut is valuable as old-Web archaeology because the page and the program reinforce each other.

The site explains parallel-port printers, serial devices, pole displays, receipt printers, scanners, DOSBox, older Windows versions, networking, and file-based register configuration. It assumes the reader might still have hardware that most modern software has forgotten exists.

That is not purely nostalgic. Old retail equipment can be durable, cheap, and perfectly adequate for a small operation. A computer that would be miserable at running a modern browser can be extremely competent at adding prices, printing receipts, and opening a cash drawer.

The documentation even explains that the .POS file effectively is the register configuration, a wonderfully concrete contrast with modern cloud software where the important state may live somewhere the user cannot see.

The result is an odd kind of longevity. The interface looks ancient because the job did not require reinvention.

CacheRat’s 1,967 Ancient Web Domains research list includes sites like KeyHut because old software sometimes survives for the best possible reason: it still solves the problem.

A touchscreen, subscription, analytics dashboard, and animated onboarding sequence would certainly make this more modern.

They would not necessarily make the cash drawer open any better.

See the KeyHut cash-register system

Posted on by

Ancient Web: Math Lair Found Real Math Inside Tic-Tac-Toe

Tic-tac-toe is usually treated as the game you teach a child five minutes before the child discovers that competent play makes it boring. Math Lair looked at that and found an entire mathematics lesson hiding underneath.

Read Math Lair’s tic-tac-toe page

The page explains standard play, then quickly moves into variants: magic-square tic-tac-toe, avoidance or misère play, movable markers, multiplayer versions, 3-D tic-tac-toe, Ultimate Tic-Tac-Toe, and other rule changes that turn the familiar grid into something less trivial.

From there it gets more interesting.

Math Lair uses the game as an introduction to combinatorial game theory and generalized tic-tac-toe. One section discusses Frank Harary’s idea of choosing polyomino shapes as the winning objective and asks which shapes can actually be forced by the first player.

That leads to “winning” and “losing” animals, smallest board sizes, move counts, and the observation that once a target contains a smaller impossible shape, the larger target is impossible too.

The page even gives a clean strategy argument for why the second player cannot have a guaranteed winning strategy in generalized tic-tac-toe: if such a strategy existed, the first player could make an irrelevant opening move and then copy it.

This is what educational hobby pages do well

The page was last updated in March 2020, but its design and structure feel much older in the best possible way. It is one long, patient explanation with links, diagrams, questions, references, and almost no interface between the reader and the idea.

That matters because mathematics pages age unusually well when they are built around reasoning instead of presentation tricks.

A flashy interactive demo may break. A clearly written argument about strategy usually does not.

Math Lair also shows how an old-fashioned personal educational site can outperform a search result. Searching “tic-tac-toe” gives you millions of playable boards. This page asks why the game behaves the way it does, what happens when the rules change, and how one simple game connects to larger mathematical ideas.

CacheRat’s 1,967 Ancient Web Domains research list has many pages like this: narrow subjects explored far past the point a commercial site would bother.

The Web is good at breadth. Pages like Math Lair remind us that individuals were often better at depth.

Visit the original Math Lair page

Posted on by

Ancient Web: The Lurker’s Guide to Babylon 5 Made Fandom Searchable

Before fandom had giant wikis, searchable social platforms, recap podcasts, and databases maintained by companies, Babylon 5 fans built something remarkably close by hand.

Visit The Lurker’s Guide to Babylon 5

The Lurker’s Guide was maintained by Steven Grimm and grew alongside the television series during the 1990s. It collected episode information, analysis, schedules, interviews, background material, online discussion, and links into one organized reference system.

Contemporary Babylon 5 FAQs pointed readers to the Guide as the place to go for detailed information. That alone says quite a bit about what the site represented at the time. Usenet could contain enormous amounts of useful discussion, but finding the right message again was another matter. The Guide acted as memory.

It also existed unusually close to the production itself.

Babylon 5 creator J. Michael Straczynski participated directly in online fan discussion, which made the show’s Internet culture unusually rich but also created practical problems. Fans were repeatedly warned not to post story ideas where Straczynski might encounter them, because unsolicited ideas could create legal complications.

The Web page became a useful boundary layer between conversation and reference.

Fandom before the wiki template

The Guide is interesting now because its structure feels familiar even though the machinery underneath it is not.

Episode pages were assembled from static files. Mirrors existed in multiple countries. FTP material sat beside Web content. Fans consulted Usenet FAQs that pointed toward the Web site, while the Web site itself linked back into the larger online ecosystem.

That was normal Internet architecture in the mid-1990s: several protocols and communities loosely stitched together by people rather than one platform owning the entire experience.

The Guide eventually became significant enough that it was cited in books and magazines as an unusually strong television fan resource. Its real achievement, though, is simpler. It made a sprawling conversation navigable.

Modern fandom sites often begin with a software platform and then accumulate information.

The Lurker’s Guide began with information that was already scattered everywhere and built the platform around the need to preserve it.

CacheRat’s 1,967 Ancient Web Domains research list includes pages like this because fan sites can become primary records of how a work was experienced while it was still unfolding.

You can read about Babylon 5 anywhere now.

This site shows what it looked like when the audience was still waiting for next week’s episode.

Explore The Lurker’s Guide

Posted on by

Ancient Web: BanDHMO Proves True Facts Can Still Be Nonsense

Dihydrogen monoxide sounds exactly like the kind of chemical somebody on Facebook would demand the government remove from school lunches immediately. BanDHMO.org built an entire straight-faced campaign around that reaction.

Visit BanDHMO.org

The trick, of course, is that dihydrogen monoxide is water: H2O described with a deliberately alarming chemical name.

The site presents water through the language of hazard warnings. It can contribute to erosion. Inhaling enough of it can kill you. It appears in industrial processes. It has been found in tumors. Every one of those statements can be framed so that it sounds sinister while withholding the single piece of context that changes everything: we are talking about water.

That is why the joke has survived far longer than most novelty websites deserve.

A 2025 paper on chemophobia specifically points to bandhmo.org as an example used to encourage scientific literacy, common sense, and skepticism. The authors describe the DHMO exercise as a useful way to show how information can be technically factual while still being arranged to mislead.

The joke is really about framing

BanDHMO works because readers tend to evaluate a claim from the vocabulary surrounding it.

Call something a chemical contaminant, list scary consequences, mention industry, environmental exposure, and death, and the brain starts assembling the expected story before it has checked the chemistry. The individual facts do not need to be false. The framing does the heavy lifting.

That makes the site more relevant now than when it was just an Internet prank.

The modern Web is extremely good at stripping context from accurate information. Charts get cropped. Relative risk appears without absolute risk. A study result becomes a headline that says considerably more than the paper did. A scary chemical name gets treated as evidence that the substance itself must be unnatural or dangerous.

BanDHMO reduces that entire problem to one stupidly effective example.

It is also a reminder that “do your own research” is meaningless if the research consists only of collecting statements that support the conclusion you already wanted. You can build a terrifying dossier on water if you choose the right facts and leave out the right sentence.

The site’s supposed campaign against DHMO has been cited in classrooms, articles, forum discussions, and demonstrations of cognitive bias for years. The gag became useful because people keep falling for the underlying technique even after they know the punchline.

CacheRat’s 1,967 Ancient Web Domains research list contains plenty of strange old pages. BanDHMO is one of the rarer specimens whose joke became more valuable as the information environment got worse.

Read it once as satire. Then read it again as a manual for spotting bullshit.

Visit the original BanDHMO site

Posted on by

Ancient Web: Akrep Nalan’s 2002 Website Became a Digital Memorial

There is something different about visiting the old website of a performer after the performer is gone. What once functioned as promotion quietly becomes an archive.

Visit AkrepNalan.com

Akrep Nalan was the stage name of Turkish singer and actress Füsun Nalan Açın. She released her first album, Dağ Çiçeği, in 1991 and followed it with Zil Zurna Sevdalar in 1994. Her website’s domain dates to May 2002, when artist homepages were often still direct extensions of the person rather than standardized social profiles.

The surviving site’s own metadata is almost charmingly literal: songs, photographs, biography, home, and ways to reach her. No engagement funnel required.

The site itself has been revised over time, but it still carries the shape of an older artist website. Search-engine records from the early 2020s describe a very small homepage by modern standards, and archived technical scans show the domain serving its own images, scripts, and familiar decorative web assets rather than being absorbed into a generic music-industry platform.

Akrep Nalan died in March 2022.

That changes what the website is.

An old artist site outlived its original job

When a musician’s current website is built entirely around tour dates, streaming links, and marketing campaigns, it can become useless the moment the career stops being active. Older personal sites sometimes survive differently because they were already collections.

Biography pages, photographs, press material, discography, and old design choices remain interesting even after there is nothing new to announce.

AkrepNalan.com also preserves a period when regional celebrity websites did not necessarily resemble the English-language commercial Web. The design, language, navigation, and presentation were made for a Turkish audience around one specific performer. That local character is historically useful. The Web was never one culture, even when modern templates make it look that way.

There is also a practical preservation point here. Independent domains can outlast the services surrounding them. Social platforms rise, change ownership, lock old material behind logins, or disappear. A domain that somebody keeps renewing can remain a stable pointer for decades.

Akrep Nalan’s music itself survives through modern services, of course. But those services tell you what can be streamed. Her old website tells you how somebody chose to present her on the Web.

Those are different records.

The larger 1,967 Ancient Web Domains research list includes many personal and celebrity sites that have become accidental archives simply by remaining online longer than expected.

AkrepNalan.com is one of them. It started as a homepage. It now functions partly as a memorial built out of HTML, photographs, and a domain somebody refused to let vanish.

Visit AkrepNalan.com