Category: Ancient Web

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

Posted on by

Ancient Web: Friends of the Creeks Kept Walnut Creek’s Waterways in Public Memory

Most local waterways become invisible once a city builds roads, parking lots, flood-control channels, and neighborhoods around them. Friends of the Creeks has spent decades making sure Walnut Creek, California, remembers that the creeks are still there.

Read the Friends of the Creeks history material

The organization grew out of a city planning effort that began in 1991, when Walnut Creek appointed a citizen task force to develop a plan for improving downtown creeks. The resulting Creeks Restoration and Trails Master Plan was completed in 1993, and members of the task force went on to form Friends of the Creeks to help put those recommendations into practice.

That history matters because the group’s website does more than announce cleanup days.

It preserves the story of how the city’s waterways were changed, constrained, bypassed, restored, and gradually reconsidered as pieces of both infrastructure and habitat.

A local website can preserve the map underneath the city

Walnut Creek’s watershed has been shaped by development and flood-control projects for generations. Railroad alignments, concrete channels, road crossings, diversions, and later restoration work altered how residents encountered the water.

The old Friends of the Creeks history page became useful enough that later local-history projects still point readers back to it when explaining how the modern city grew around the creek system.

The organization itself remained practical. Its work has included annual downtown creek cleanups, invasive-plant removal, habitat restoration, public education, creek-history talks, advocacy during development planning, and efforts to expand restoration beyond the downtown area.

A 2014 watershed report recorded at least 190 volunteers participating in the group’s 25th annual downtown cleanup, covering roughly 2.5 miles of creek. By then the work had shifted partly from simply removing trash toward improving habitat because the amount of garbage had declined.

That is a satisfying kind of progress to see preserved in old reports and Web pages.

There is also a broader Internet-history lesson here. Small civic organizations often hold extremely specific local knowledge that does not fit neatly into a city government’s official history. Their websites preserve maps, talks, volunteer reports, explanations of landscape changes, and institutional memory that can disappear when the people involved retire or move away.

CacheRat’s 1,967 Ancient Web Domains research list exists partly to find this quieter kind of Web history. Not everything valuable online is a famous homepage or lost piece of software. Sometimes it is the page explaining why the creek behind a shopping center bends where it does.

Cities pave over memory quickly.

Volunteer webmasters sometimes put it back.

Explore Friends of the Creeks

Posted on by

Ancient Web: Pekka Väänänen Rendered Quake on an Oscilloscope Through an Audio Jack

There are normal ways to play Quake. Pekka Väänänen chose to make a Hitachi V-422 oscilloscope display it through an audio connection.

Read “Quake on an oscilloscope: A technical report”

The 2014 write-up is exactly what a good personal technical page should be: not merely a video of a weird result, but a record of how the weird result worked and what went wrong along the way.

An oscilloscope in XY mode does not receive a conventional video frame. Two voltages control the horizontal and vertical position of the beam. Väänänen’s solution was therefore to reduce Quake‘s rendered scene to line geometry, convert the points into two audio channels, and let the left and right voltages steer the beam.

That sounds simple right up until the details arrive.

Lines need to be drawn at reasonably constant speed or their brightness changes. Geometry has to be culled aggressively because an oscilloscope cannot sensibly redraw every polygon in a Quake level. Traversal order matters because jumping from one unrelated line to another produces unwanted connecting streaks. The signal needs enough sample rate to draw a useful number of points before persistence fades.

Väänänen started with an XY oscilloscope simulator in Processing, then modified the DarkPlaces Quake source. The final system used audio output at high sample rates, line interpolation, view-frustum and back-face culling, and other compromises to turn a textured 3D game into a glowing wireframe display.

The page includes recordings and implementation details instead of hiding the trick behind a finished demo.

The explanation is better than the stunt

The video is the part people remember. The article is the part worth preserving.

A short viral clip could show Quake running on an oscilloscope and leave every interesting question unanswered. Väänänen documented the signal model, the renderer changes, practical limitations, test tools, and later updates. A technically curious reader can follow the path from “would this work?” to an actual implementation.

This is a later Web artifact than many Ancient Web entries, but it belongs to the same independent publishing tradition: one person builds something unreasonable, writes down the engineering, hosts the files, and leaves the page available long after social posts about the project would have vanished down a feed.

For programmers, the page is also a nice example of constraints producing clarity. Strip away textures, framebuffers and modern displays, and the problem becomes brutally physical: where should the electron beam be next?

CacheRat’s 1,967 Ancient Web Domains research list includes technical pages like this because obscure experiments often contain better engineering explanations than the polished summaries that later reference them.

Somebody made Quake output voltages until an oscilloscope became a monitor.

The least we can do is keep the notes.

Explore Väänänen’s oscilloscope Quake report

Posted on by

Ancient Web: Paul Burgess Kept the Personal Homepage as a Place for Ideas

A personal homepage used to be allowed to contain whatever happened to be inside one person’s head. Paul Burgess’s site still follows that rule: writings, fiction, interests, philosophical material, and what the homepage itself calls “pseudo-philosophical ramblings.”

Visit Paul Burgess’s personal website

The domain dates to 2001, but the material reaches across a broader period of personal Web culture in which the homepage was not a résumé and did not need a content strategy.

Burgess wrote fiction, essays, definitions, commentary, and material touching on subjects such as semiotics and philosophy. One older essay, Why Triadic?, concerns the structure of Charles Sanders Peirce’s semiotic theory and has been cited in later academic work on Peirce and artificial intelligence.

Elsewhere the site is simply personal.

That mixture is the point.

Before every interest needed its own account

Burgess also maintained a blog called Let the Finder Beware. In his final post in November 2007, he wrote that he had been blogging for about three years, had produced 951 posts, and finally felt that he was running out of fresh things to say.

That ending is almost quaint now.

The modern Web is built around systems that would strongly prefer nobody ever conclude they have posted enough. Personal sites worked differently. You could write hundreds of entries, stop, leave the archive standing, and move on without an algorithm sending concerned notifications about your engagement metrics.

The relationship between Burgess’s homepage and blog is also typical of the transitional Web of the early 2000s. The personal domain remained the stable home for longer-form material and identity, while Blogger handled the chronological stream. The platforms were tools attached to a personal site rather than replacements for one.

What survives is therefore more than a collection of essays. It is a model of publishing in which a person’s interests did not need to be commercially coherent.

A philosophy page could sit next to fiction. A blog could link back to an old essay. A definition written years earlier could later appear in somebody else’s bibliography. The website accumulated a shape because the author did, not because a template designer had decided where the “content pillars” belonged.

CacheRat’s 1,967 Ancient Web Domains research list exists for exactly these personal sites. They are difficult to summarize because summarization was never the point.

The homepage was the category.

Browse Paul Burgess’s surviving personal site

Posted on by

Ancient Web: The Web’s First Japanese Pizza Page Has Been Judging Mayonnaise Since 1995

On September 5, 1995, somebody in Tokyo decided the Web needed a page explaining Japanese pizza to the rest of the world. Thirty-one years later, the page is still here, and the mayonnaise has not become less alarming.

Visit the Web’s First Japanese Pizza Page

The author lived near Pizza Studio in Setagaya, Tokyo, and built the page around the restaurant’s menu. Instead of offering a broad history of pizza in Japan, the site does something much more valuable: it records what an ordinary neighborhood pizza place was actually selling in the mid-1990s.

The menu includes familiar ingredients such as pepperoni, mushrooms, bacon, onions, and garlic.

Then things begin moving sideways.

A Seafood Special combines shrimp, squid, octopus, clams, scallops, and crab legs. Another seafood pizza pairs shrimp, squid, tuna, and mayonnaise. The Japanese Style pizza includes pork, shimeji mushrooms, bamboo shoots, nori, and shiso. The author also explains that sweet corn, boiled egg, potato-and-mayonnaise combinations, and squid-ink sauce were common enough to deserve mention.

A restaurant menu accidentally became food history

The page is funny because the author clearly has opinions. It opens by suggesting that some forms of Japanese pizza may qualify as an offense against mankind, then promises to avoid further value judgments before immediately struggling with that promise for the rest of the page.

But underneath the jokes is a useful cultural snapshot.

Restaurant menus change constantly. Small restaurants close. Printed flyers disappear. Delivery chains standardize tastes. A homemade Web page from 1995 can preserve combinations and marketing language that would otherwise survive only in somebody’s memory.

The page also became a reference in its own right. Later food directories, Japan-resource pages, discussion forums, and even a published pizza guide pointed readers back to it. A 2003 Japan forum discussion about unusual toppings linked to the page as the obvious existing reference.

This is a perfect example of early-Web authority emerging from specificity rather than credentials. The author was not trying to create the definitive scholarly history of Japanese pizza. He had a neighborhood pizza shop, a Web page, and concerns about squid ink.

That was enough.

CacheRat’s 1,967 Ancient Web Domains research list exists for pages like this because tiny contemporary observations often become better historical evidence than later nostalgia.

The page was created in 1995.

The squid-and-mayonnaise combination remains under investigation.

Browse the original Japanese Pizza Page

Posted on by

Ancient Web: ROT8000 Built ROT13 for the Unicode Generation

ROT13 has one serious design flaw: it assumes the Internet speaks English. ROT8000 looked at the other tens of thousands of Unicode characters and decided this injustice could no longer stand.

Visit ROT8000

The project describes itself as “rot13 for the Unicode generation.” Traditional ROT13 shifts each Latin letter thirteen places through the alphabet, making the same operation both encode and decode the text. It was famously useful on Usenet for hiding spoilers, punchlines, offensive material, and anything else readers should have to deliberately unscramble.

ROT8000 applies the same general gag to the Unicode Basic Multilingual Plane.

Instead of rotating twenty-six letters, it shifts eligible code points roughly halfway across that 16-bit space while leaving whitespace and surrogate ranges alone. Apply the transformation again and the original text returns.

The result looks spectacularly broken while remaining intentionally reversible.

An art project escaped into programmer culture

ROT8000 was created for the pl41nt3xt pavilion of text-based works in The Wrong, an online digital-art biennale curated in this case by A Bill Miller.

The implementation existed first in C# and later moved to JavaScript, allowing the transformation to happen directly in the browser. The project’s source repository also includes tests checking reversibility across the supported character space.

Then people outside the art project noticed it.

ROT8000 circulated through Hacker News, Boing Boing, and commentary from Cory Doctorow. It eventually appeared as an operation in tools inspired by CyberChef, turning a deliberately excessive spoiler cipher into a recognizable bit of hacker utility.

The project is funny because it solves a problem nobody urgently had with far more Unicode than anyone needed.

It is also a neat example of Web-native software culture. A tiny interactive page can be artwork, programming joke, educational demonstration, and reusable library at the same time. There is no product funnel. Nobody needs an account. You paste text into a box, the browser turns it into apparent typographic damage, and the concept explains itself.

That simplicity is worth preserving even though ROT8000 is younger than many Ancient Web subjects. It belongs to the same independent-web lineage: make a strange thing, publish the source, explain the idea, and leave a stable page behind for whoever wanders in next.

CacheRat’s 1,967 Ancient Web Domains research list includes plenty of sites whose value comes from exactly that kind of unnecessary ingenuity.

ROT13 hid spoilers.

ROT8000 makes them look like your character encoding suffered a workplace accident.

Try ROT8000 and read how it works

Posted on by

Ancient Web: Daring Armstrong Got an Entire Fake Hero Homepage on Something Awful

In 2003, Something Awful did not merely write jokes about a fictional idiot named Daring Armstrong. It gave him an entire personal homepage so he could lie about his adventures properly.

Visit Daring Armstrong’s Awesome Page of Heroism

The page is written entirely in character as Armstrong, a catastrophically self-impressed American time adventurer. He explains that the site celebrates his journeys “through time, through love” and whatever other exciting nouns he cannot currently remember. His fictional intern, Laura Abromowitz, is supposedly responsible for turning his journals and scrapbooks into a website where grateful citizens can thank him for repeatedly saving the world.

Navigation is part of the joke.

There are Comics of Adventure, a Photo Album of Adventure, a Journal of Adventure, the Timelab Crew of Mild Excitement, trophies, and a section cataloging women he claims to have seduced. Random “Dareotoids” deliver useful hero knowledge. The trophy cabinet includes awards for everything from skin moisture and bocce ball to time travel, steak consumption, film editing and being extremely attractive.

The joke works because it is not presented as a modern article explaining a character.

It is presented as his website.

The fake homepage is the artifact

Daring Armstrong appeared in Something Awful stories about aliens, time travel, political stupidity and other genre nonsense. The hosted site takes the parody one step further by giving that fictional universe its own in-world Web presence.

That format belongs to a particular period of Internet comedy. A character did not need a social account with scheduled posts. You could build the character a terrible homepage, stuff it with fake photographs and self-authored biography, and let readers explore the joke sideways.

The surviving page also preserves the architecture of early Web humor: image-heavy navigation, separate HTML sections, little bits of lore scattered across subpages, and enough commitment to the premise that the reader has to poke around to find everything.

There is a useful preservation lesson hiding under the stupidity. Comedy websites are easy to dismiss as disposable because they were built around current jokes. But years later they document formats, voices, design conventions, and communities that serious institutional archives rarely bothered to collect.

CacheRat’s 1,967 Ancient Web Domains research list is not limited to historically respectable material. The old Internet also contained a great deal of meticulously constructed garbage, and some of that garbage was excellent.

Daring Armstrong would presumably claim he invented the Web himself during a classified adventure.

Fortunately, his homepage is more credible than he is.

Explore Daring Armstrong’s Awesome Page of Heroism

Posted on by

Ancient Web: Don Worth Put Beneath Apple Manor Back Online Himself

One of the earliest procedurally generated computer role-playing games is still connected to the personal homepage of the guy who wrote it. That is almost offensively convenient for Internet archaeology.

Visit Don Worth’s UCLA homepage

Don Worth was a systems programmer at UCLA when he created Beneath Apple Manor for the Apple II. The game was first published in 1978, two years before Rogue, and generated new dungeon layouts rather than making players traverse one fixed map.

That design choice eventually made Beneath Apple Manor important to discussions of roguelike history and procedural generation.

Worth’s own page matters because it keeps the creator attached to the artifact.

Rather than relying entirely on later game databases, researchers can follow the trail back to his personal university-hosted site, where a PC version of the game was made available and where Worth described his computing background and the game’s origins.

Before “roguelike” had anything to be like

Worth had previously written games on UCLA’s IBM 360 mainframe and was a Dungeons & Dragons player. After buying an Apple II, he began translating tabletop ideas into software: random dungeon rooms, hidden areas, monsters, treasure, character statistics, combat, and the uncertainty of exploring spaces the player could not yet see.

The working title was reportedly Underground Adventure.

The released game sends a lone adventurer into the cellars beneath a ruined manor in search of a Golden Apple. Each run can produce a different dungeon. Monsters move outside the player’s visible area, and death has consequences that later generations would recognize immediately from roguelike design.

None of this required a giant studio.

It required a programmer, an Apple II, Integer BASIC, some 6502 machine language for speed-sensitive portions, and the desire to make a computer behave more like a tabletop dungeon master.

The personal homepage also demonstrates something important about software preservation. At one point the original publisher and commercial context stopped mattering. The creator himself could simply put a version of his old program online and let people keep studying and playing it.

That direct creator-to-Web path has preserved countless pieces of computing history that would otherwise depend on abandoned corporate archives.

CacheRat’s 1,967 Ancient Web Domains research list exists for pages like Worth’s because sometimes an unassuming university homepage contains a primary source for an entire genre.

The dungeon keeps regenerating.

Fortunately, the homepage did not.

Explore Don Worth’s page and Beneath Apple Manor material

Posted on by

Ancient Web: Stefanie Wiele’s Homepage Preserved a World Championship in Shorthand

There are personal websites about bands, video games, cars, and pets. Stefanie Wiele built one around competitive stenography, which is a useful reminder that the old Web gave every serious hobby exactly as much room as its owner was willing to type.

Visit Stefanie Wiele’s stenography site

The German-language site collects material about Stenografie—shorthand—along with text processing, typing competitions, training material, and Wiele’s experiences in organized international competition.

Its most historically interesting section concerns the 43rd Intersteno Congress, held in Hanover from July 28 through August 3, 2001.

Contemporary material from Wiele’s stenography association records that she and Jochen Schönfeld produced what the organization described as the greatest competition success in its history, bringing world-championship honors to Oldenburg. The achievement was significant enough that they were received at city hall by the mayor.

The association specifically directed readers to Wiele’s website for her firsthand account, photographs, a radio interview, and the mayor’s speech.

That is exactly the sort of material personal websites were good at preserving.

A sport most people forgot was a sport

Competitive shorthand and text entry have a long history built around speed and accuracy. Intersteno traces international competitions back well before personal computers, through eras of handwritten shorthand, typewriters, electronic word processing, and modern keyboarding.

Wiele’s site sits in the middle of that transition.

The pages are useful not merely because they explain shorthand systems but because they document the community around them: clubs, congresses, practice, records, competition results, and the pride surrounding a skill that became increasingly invisible as office technology changed.

The site also has a wonderfully old-Web quality in 2026. Search discussions still point beginners toward it when somebody finds a mysterious page of German shorthand and asks what the symbols mean. A personal reference page created by a world-class practitioner continues answering questions years after most people stopped encountering shorthand in ordinary life.

That is preservation through usefulness.

CacheRat’s 1,967 Ancient Web Domains research list exists for sites like this because disappearing skills often survive first in the personal pages of the people who cared enough to teach them.

The office world replaced shorthand pads with laptops.

Wiele kept the strokes online.

Explore Stefanie Wiele’s shorthand and typing pages

Posted on by

Ancient Web: Galactic Server Is a Surviving Channeling-Era Cosmology Hub

Galactic Server looks exactly like a website called Galactic Server ought to look: a dense constellation of links to UFO contacts, crop circles, channeling, extraterrestrial claims, spiritual cosmology, Internet radio, memorial pages, and subjects that require several additional dimensions before breakfast.

Visit Galactic Server

The site describes its purpose as distributing “spacepeople related material” based mainly on channeling and personal experience, with a focus on galactic affairs and spiritual or interdimensional topics.

Those are the site’s claims and beliefs, not established facts.

That distinction matters, but it does not make the website historically uninteresting. Quite the opposite.

Alternative-belief communities were among the earliest groups to discover how useful the Web could be for connecting material that traditional publishing rarely touched. A personal server could hold transcripts, contact stories, scanned photographs, newsletters, radio links, biographies, theories, and references to other researchers without needing approval from a publisher or broadcaster.

Galactic Server is a particularly dense survivor of that style.

A homepage as a cosmology map

The main page links outward to sections with names such as UFO Contacts, Central Cropcircle Base, How to Contact Extraterrestrials, channeling collections, and material associated with figures in UFO and New Age circles. It also contains apparently unrelated personal interests, including a Kate Bush resource page.

That collision is pure personal Web.

A modern content system would separate these into categories, brands, channels, or accounts. Galactic Server instead behaves like one person’s mental map rendered directly into hyperlinks. The visitor is expected to follow whichever branch looks interesting and eventually emerge somewhere else entirely.

Older references identify the site’s operator with Tore Alfstad and list Galactic Server under several domains over the years, suggesting the material migrated as hosting arrangements changed. That kind of domain-hopping was common for personal archives that outlived their original providers.

For researchers, the value is not deciding whether the site’s extraordinary claims are true. It is preserving how a particular Internet subculture documented itself: the terminology, networks of influence, people being cited, subjects being connected, and presentation style used before social platforms flattened these communities into feeds and groups.

CacheRat’s 1,967 Ancient Web Domains research list exists for precisely this kind of artifact. The stranger the worldview, the more useful it is to preserve the original source rather than rely on somebody else’s summary of it.

You do not have to believe in interdimensional spacepeople.

You only have to believe somebody spent a spectacular amount of time hyperlinking them.

Explore Galactic Server

Posted on by

Ancient Web: Doctor Nostromo Has Been Running the Same Hobby Site Since 1995

Doctor Nostromo’s homepage says it plainly: established in 1995 and still updated daily. That alone puts the site in a fairly ridiculous class of Internet survivor.

Visit Doctor Nostromo’s Home Page

The author also explains why the site still looks the way it does. This is a hobby. It earns no income. There are no ads, no registration, no tracking, no visitor-data collection, and no cookies being shoved into the browser for marketing purposes. He likes simple pages and has no interest in replacing them with the sort of modern interface that requires a map to navigate.

Thirty-one years later, that decision has become part of the site’s identity.

The content is sprawling but personal rather than algorithmic. There are more than 1,900 short movie reviews, large curated music playlists, an image gallery, Mystery Science Theater 3000 material, an Avengers timeline, game projects, and sections that have accumulated because the person maintaining the site happened to care about them.

The gaming material is especially useful historically. DrNostromo hosts the official Hacx site and material related to older Doom projects including STRAIN and the Nostromo’s Run Project. The Hacx section preserves files, history, and notes from a modding community that is itself now approaching three decades old.

There is even a notice asking fans to stop requesting unreleased Hacx files because everything the author still has has already been made available.

That is an extremely old-Web problem to still be having.

The homepage never became a content strategy

One of the stranger things about a site this old is that its original publishing model still makes sense.

There is no need to ask what the “brand” is. The brand is one guy’s interests. Movies sit next to music. Doom sits next to pictures. A game called To Be King, which the author says he worked on for more than thirty years, now has its own section and public release.

The homepage was updated in July 2026, while individual sections continue receiving their own newer dates. This is not a museum copy of a 1995 page. It is a living personal website that simply never accepted the premise that personal websites needed to evolve into social-media accounts.

For a webmaster, that is worth noticing. The technical simplicity is not preventing the site from publishing. It may be part of why the site remains maintainable after three decades.

CacheRat’s 1,967 Ancient Web Domains research list contains abandoned sites, archived sites, and sites that survived by barely changing. DrNostromo is the more interesting fourth category: an old site that is still alive because somebody still enjoys running it.

No pivot required.

Browse DrNostromo.com