Category: Ancient Web

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

Posted on by

Ancient Web: Richard Eckert’s Binghamton Homepage Left a Computer Science Classroom Online

University websites used to be full of doors that opened directly into somebody’s office.

Richard R. Eckert’s Binghamton University homepage preserves that style of academic Web: a named directory on a department server, personal material, rĂ©sumĂ© information, and course resources that remained reachable without first passing through a corporate learning platform.

Eckert’s background is unusually broad. His rĂ©sumĂ© lists a Ph.D. in experimental high-energy physics, an M.S. in experimental solid-state physics, and a B.S. in physics before a long career teaching computer science. He joined SUNY Binghamton in 1983 and taught there for decades.

That history matters because the site reflects the era when computer science departments routinely published teaching material directly to the Web.

When the course URL was the course platform

Surviving Eckert course material includes introductions to Web applications, browsers, servers, HTTP, HTML, and JavaScript. One set of notes explains the browser/server relationship, walks through URL structure, describes a basic HTTP request, and points students toward simple HTML examples hosted in Eckert’s own course directories.

The examples are gloriously literal.

A page can be called mypage0.html because its job is to teach a student what the first page looks like, not win an SEO contest.

This kind of academic site was common and quietly important. Faculty pages became public textbooks, lab notebooks, assignment shelves, bibliographies, and software archives. Search engines indexed them. Other teachers linked to them. Students outside the university could stumble into a lecture they were never enrolled in.

Modern learning-management systems are better at grades, permissions, enrollment, and institutional administration. They are considerably worse at accidentally leaving useful knowledge exposed to the entire Internet for twenty years.

That is the preservation tradeoff.

CacheRat’s 1,967 Ancient Web Domains research list repeatedly runs into old university ~username directories because academia once produced an enormous public shadow library almost by accident.

Nobody called it open educational infrastructure.

Professors just put their files on the server.

Explore Richard Eckert’s surviving Binghamton pages

Posted on by

Ancient Web: AutoHistories Put Poland’s Beskid Prototype Online in 1996

The Web was barely old enough to have bad habits when Bolko Rawicz published a page about an obscure Polish prototype car.

The surviving Beskid page says plainly: “Page created on 3 Sept. 1996.”

That date matters because the page is not a recent nostalgia article retrofitted onto an old-looking site. It is a genuinely early Web document built to collect technical history that was difficult to find elsewhere.

The subject is the Beskid 106, an experimental small car developed in Poland during the 1980s by BOSMAL for FSM, the maker associated with the Polish Fiat 126p.

A prototype that never became ordinary

Development began in the early 1980s under engineer Wieslaw Wiatrak. The body, designed by Krzysztof Meissner, was carefully wind-tunnel tested and reached a drag coefficient of roughly 0.29—impressive for a tiny economy-car project of the period.

The page documents multiple prototypes, different engine sizes, independent suspension, Fiat 126p-derived components, a smooth underside, and a body designed to reduce the number of steel stampings required for production.

It also records the disappointing ending: development stopped in 1987 and the Beskid never entered mass production.

The car later became visually famous for looking uncannily like the Renault Twingo that arrived years later, earning it the nickname “Polish Twingo.” The page is careful to present that resemblance as resemblance rather than proof of some grand stolen-design conspiracy.

What makes the website especially useful is its sourcing. Rawicz credits the BOSMAL research center for photographs and information and says additional details came from 1986 issues of the Polish magazine AUTO – Technika Motoryzacyjna.

That is old-Web gold: a hobbyist page acting as a bridge between factory material, print magazines, and international readers.

CacheRat’s 1,967 Ancient Web Domains research list exists to surface pages like this—small, specific documents that became more valuable simply by refusing to disappear.

In 1996, finding information about a failed Polish prototype was not a three-second search.

Somebody had to build the page.

Read the original Beskid history page

Posted on by

Ancient Web: Bimshwel Preserved Moraff’s Dungeons of the Unforgiven in All Its Garbage-Can Glory

There are professional game retrospectives, and then there is Bimshwel’s page about Moraff’s Dungeons of the Unforgiven, which appears to have been written under the influence of prolonged exposure to rainbow garbage cans.

The page says its original version appeared around mid-2000. It documents a 1993 DOS dungeon crawler from Moraffware with screenshots, quoted marketing copy, technical observations, old correspondence, and an unreasonable amount of commentary about the game’s monsters.

The monsters deserve it.

Among them are walking garbage cans, toxic-waste containers with limbs, vaguely diseased creatures, and enemies whose animation is effectively produced by flipping graphics back and forth.

A review accidentally became preservation

The page is funny, but the historical value is underneath the jokes.

It records the game’s support for an absurd range of display modes: Hercules, CGA, EGA, VGA, Super VGA, and high-resolution modes that required choosing from old video-adapter chipsets.

It reproduces the old FILE_ID.DIZ shareware description advertising “HOTTEST SVGA EVER!!” and “25 levels of the most realistic 3D imaging ever seen in a computer game.” It also preserves the later disclaimer insisting that the game was not a Doom clone.

That is pure shareware archaeology.

The author also records contacting Moraffware and being told that Steve Moraff had moved away from the old DOS dungeon games and had no plans to update them.

What began as mockery therefore became a primary record of a software line already slipping out of circulation.

The site even notes that the page itself moved between hosts over the years. That gives it the familiar old-Web double layer: an article about abandoned software that nearly became abandoned Web content.

CacheRat’s 1,967 Ancient Web Domains research list is full of pages where obsessive humor preserved things more effectively than formal institutions did.

Nobody appointed this page the official historical archive of Dungeons of the Unforgiven.

It just kept existing longer than the thing it was making fun of.

Read the Bimshwel Moraff retrospective

Posted on by

Ancient Web: Mysterium Kept a Controversial Mithras Thesis Online

The old Web was full of pages that were not merely summaries. They were arguments.

Mysterium.com’s Mithras page preserves material associated with David Ulansey’s interpretation of the Roman mystery cult of Mithras, especially the idea that Mithraic imagery can be read through ancient astronomy and the precession of the equinoxes.

Ulansey’s work became well known through his 1989/1991 book The Origins of the Mithraic Mysteries and related writing. His astronomical interpretation has been debated rather than universally accepted, which is exactly why preserving the original argument matters.

A theory becomes much harder to evaluate when all that survives are other people describing it.

The Web as a shelf for scholarship

The Mysterium material has remained linked and cited from later pages discussing Mithraism, axial precession, ancient cosmology, and the relationship between astronomical symbolism and Roman religion.

That gives the page a different kind of value from a normal reference site.

It is not simply trying to tell readers what Mithraism “was.” It preserves one scholar’s attempt to explain a difficult body of evidence.

That distinction matters. Ancient religions are reconstructed from incomplete archaeological and textual remains. A page presenting a specific interpretation should be read as an interpretation, not as settled fact.

The older Web was often surprisingly good at making room for this kind of material. Scholars, independent researchers, and obsessive specialists could publish long-form arguments without waiting for a platform to decide whether the subject was broadly marketable.

Twenty-five years later, the page itself becomes part of the history of how the argument circulated.

CacheRat’s 1,967 Ancient Web Domains research list is not just a pile of retro design. It includes sites where ideas were published, challenged, linked, and left standing long enough to become references in their own right.

You do not have to agree with Ulansey’s theory to appreciate the preservation problem.

A disputed argument is still evidence of intellectual history.

Read the surviving Mysterium Mithras material

Posted on by

Ancient Web: The QP7 Programming Page Preserved a GeoCities Coding Course

Before coding education became an industry of subscription platforms, bootcamps, influencer funnels, and twelve-minute videos with thirty seconds of useful information, people routinely learned from somebody’s personal homepage.

The surviving QP7 Programming Page mirror is a clean example.

The site lived in GeoCities’ SiliconValley neighborhood and offered tutorials and references for languages including C++, Visual Basic, QBASIC, and Pascal.

Its C++ material is laid out like a small book. Sections cover arrays, structures, pointers, file I/O, algorithms, practical applications, classes, object-oriented programming, constructors, destructors, inheritance, and related topics.

No account required. No progress dashboard. No badge for completing lesson three.

Just pages.

A homepage could be a textbook

Old programming sites were often written by people who had recently learned the material themselves and wanted to reorganize it for the next person.

That made them uneven, but it also made them unusually direct.

The QP7 pages were linked from other programming resource lists alongside author pages, tutorials, college material, and hobbyist references. That is how discovery worked: one useful page linked to another useful page, and eventually you had twenty browser windows open and knew slightly more C++ than when you started.

The important historical detail is not that every explanation on QP7 should be treated as modern best practice. Programming languages, compilers, libraries, and conventions changed.

What the mirror preserves is how programming knowledge circulated on the Web.

A teenager or hobbyist could create a structured tutorial, place it in a GeoCities neighborhood, and become part of somebody else’s self-taught curriculum halfway around the world.

When GeoCities disappeared, pages like this became easy to dismiss as obsolete clutter. But they record the educational infrastructure of an Internet generation.

CacheRat’s 1,967 Ancient Web Domains research list is useful precisely because many important old pages were not institutions. They were one person’s contribution to a chain of links.

The old Web did not have a learning platform.

It had view source, bookmarks, and somebody in SiliconValley/Bay/5707 who decided to explain pointers.

Browse the surviving QP7 Programming Page mirror

Posted on by

Ancient Web: PieMan Built ‘Pie-berspace’ for Performance Art

Some websites preserve software manuals. Some preserve military history. PieMan.com preserves a man doing performance art with pies.

This is why the old Web remains superior as an archaeological dig.

The surviving page is dated October 31, 2000 and calls itself “Pied By An Angel.” It documents PieMan at San Francisco’s Castro Street party and then opens into a tiny archive of newspaper clippings, old photographs, pie performances, definitions, and audio/video material.

There is even a 1993 pie rap.

Of course there is.

The footer calls the site “Pie-berspace.” That phrase alone should probably be placed in a museum.

The personal website as evidence locker

The important thing is that PieMan.com did not merely advertise a performer. It accumulated evidence that the performance happened.

The page points toward early photographs, archival newspaper coverage, pieces of a 1993 Pie Show, and recordings from a period when putting video or audio on a personal website was still a deliberate technical act rather than pressing an upload button.

That makes the site a useful example of how obscure cultural material survived online.

A strange local performance could have disappeared into boxes of clippings and unlabeled tapes. Instead, somebody gave it a domain name and built a public index.

The site does not need a polished retrospective explaining why PieMan mattered. The surviving artifacts do the work. You can see the dates, the performances, the press references, and the cheerful commitment to inserting the word “pie” into nearly every available noun.

Modern platforms are very good at distributing a performance while it is happening. Personal websites were better at leaving behind a stable little room containing the leftovers.

CacheRat’s 1,967 Ancient Web Domains research list contains many pages that look ridiculous until you realize they are now the primary surviving record of something nobody else thought to archive.

PieMan appears to have understood the Web perfectly.

If history will not preserve your pie-formance, build Pie-berspace and do it yourself.

Enter PieMan’s surviving archive

Posted on by

Ancient Web: Queen Christina Survived the Death of Ancient Sites

There is a specific kind of old website that feels less like an article and more like somebody quietly building a miniature museum because they cared about one subject too much to leave it alone.

Tracy Marks’s Queen Christina of Sweden site is one of those.

Copyrighted in 1999 under Marks’s Ancient Sites identity, Torrey Philemon, the site covers Christina’s early life, abdication, later years, people around her, quotations, and the Greta Garbo film based loosely on her life.

The page also preserves its own migration story.

It states that it was originally created for The Ancient Sites Celebration of Women and that the Ancient Sites community folded on March 30, 2001. Marks then moved the surviving pages elsewhere rather than letting the work disappear with the host.

That little note is almost as interesting as the historical subject.

Early web scholarship was often personal

Christina of Sweden is a perfect old-Web subject: unconventional monarch, religious controversy, patron of the arts, abdication, Rome, and enough personality to keep a hobbyist clicking through references for months.

Marks did not need an institutional publishing platform to build a structured resource around her. A personal website was enough.

The site also reflects the Web’s old neighborhood model. Ancient Sites, like GeoCities and similar communities, grouped personal pages around themes and encouraged people to build subject-specific spaces. The result was often uneven, enthusiastic, and far more human than a standardized encyclopedia template.

When the host disappeared, the good pages became endangered immediately.

Marks migrated hers.

The page’s own footer still records that transition, including its 2001 update date and links into Marks’s wider network of Web projects. It is both a history page and evidence of the hosting culture that produced it.

CacheRat’s 1,967 Ancient Web Domains research list contains plenty of pages whose survival depended on somebody caring enough to move them manually when the landlord disappeared.

Queen Christina abdicated one throne.

Her website survived the collapse of another.

Explore the surviving Queen Christina site

Posted on by

Ancient Web: VelmaDinkley.com Escaped ACMEcity and Kept the Fan Site Alive

Most personal pages hosted on giant free-homepage services eventually died with the service. VelmaDinkley.com did something rarer: it escaped.

Visit VelmaDinkley.com

The site began in 1999 under the alias John Likeglass on ACMEcity, Warner Bros.’ short-lived attempt to build a GeoCities-like hosting network around its own entertainment properties.

ACMEcity was a strange dot-com idea even by dot-com standards. Warner Bros. and FortuneCity offered fans free web space organized into entertainment-themed neighborhoods. Instead of fighting every fan who posted copyrighted images, the company tried giving those fans a sanctioned place to build.

For a while it worked extremely well. ACMEcity reportedly grew into the hundreds of thousands of users during its first year.

Then the AOL-Time Warner merger happened.

ACMEcity became redundant beside AOL’s own hosting services, and the platform was shut down in 2001. Huge numbers of personal sites simply vanished.

Velma’s did not.

A fan page that moved its own house

Likeglass migrated the site away from ACMEcity, later used GeoCities, and eventually established the independent VelmaDinkley.com domain.

That sequence is important.

A personal website was not necessarily the same thing as the service hosting it. People who cared enough about their pages sometimes carried them from one doomed platform to another like digital furniture.

VelmaDinkley.com retained the classic fan-site mix: character material, Scooby-Doo enthusiasm, community features, and a chat area that continued functioning long after the original hosting culture disappeared. The site became interesting not just as a Scooby-Doo page but as one of the surviving artifacts of ACMEcity itself.

ACMEcity barely lasted two and a half years. The fan page outlived it by decades.

That reverses the relationship modern users are trained to expect. Today a creator usually assumes the platform is permanent and their account is disposable. The old Web repeatedly demonstrated the opposite.

CacheRat’s 1,967 Ancient Web Domains research list is full of survivors whose real historical value lies in exactly this kind of continuity.

Platforms died.

People moved their weird little websites.

Sometimes the weird little website won.

Browse the surviving Velma Dinkley fan site

Posted on by

Ancient Web: DNEP Kept Duke Nukem: Manhattan Project Mapping Alive

Some fan sites preserve screenshots. Others preserve the part that is much harder to reconstruct later: how people actually made things.

The Duke Nukem Editing Project existed to keep the editing scene around Duke Nukem: Manhattan Project usable after the original moment had passed.

That meant maps, troubleshooting, tutorials, editor quirks, and the accumulated knowledge required to wrestle with PrismEd, the game’s editing tools.

This was not a friendly modern editor with a searchable Discord history and twelve video walkthroughs. Surviving DNEP tutorials warn that learning PrismEd can be slow and frustrating, recommend reading the official material first, and document the sort of practical failures that only become obvious after somebody loses an afternoon to them.

Fan documentation became the real manual

The material associated with DNEP covers basic map construction, lighting, game tracks, powerups, monsters, forcefields, sky links, terrain tricks, triangle patches, and assorted editor problems.

That is exactly the kind of knowledge old modding communities routinely produced and then lost.

A commercial game’s official documentation might explain the supported workflow. Fan communities documented what actually happened when the editor crashed, when a feature was half-finished, when a shader behaved strangely, or when a map technically built but still refused to work.

Years later, some DNEP tutorials have been mirrored elsewhere specifically because the original forum ecosystem became fragile. That alone explains the historical value of the surviving project page.

The website is not important because Manhattan Project was the center of the gaming universe. It wasn’t. It matters because small editing communities are where a great deal of practical game-development knowledge lived.

When those communities disappeared, the knowledge often disappeared with them.

CacheRat’s 1,967 Ancient Web Domains research list contains many sites like this: places where enthusiasts quietly wrote the documentation that commercial publishers never bothered to preserve.

The glamorous part of modding history is the finished level.

The useful part is the page explaining why the damn thing would not compile.

Visit the surviving Duke Nukem Editing Project

Posted on by

Ancient Web: The GIF Pronunciation Page Has Been Arguing for ‘Jif’ Since 1998

Steve Olsen built a webpage in 1998 to settle the pronunciation of GIF. The argument is still going, which means the page has now been failing successfully for nearly three decades.

Visit The GIF Pronunciation Page

Olsen’s position is not subtle: GIF is pronounced with a soft G, like “jif.” But the page is more useful than a declaration because he kept collecting evidence.

There is the October 1987 issue of CompuServe’s Online Today, published when GIF was new, which explicitly gives the soft-G pronunciation. There is old CompuShow documentation saying the same thing. An early GIF89 image by CompuShow author Bob Berry contains embedded text telling viewers how to pronounce the format. Olsen quotes old graphics FAQs, NetBITS, people who worked around CompuServe, developers who wrote early GIF software, and later statements by GIF creator Steve Wilhite.

Then come the dissenters.

The reader-comments section is practically an archaeological layer of the argument. Some insist the G must be hard because it stands for “graphics.” Others argue that creators do not control a word forever. One correspondent gets deep into English phonology. Another proposes pronouncing JPEG incorrectly out of spite. Someone else suggests the G in GIF should simply be silent.

This is what happens when a personal webpage stays online long enough: the original argument becomes only part of the artifact.

The page accumulated the history of its own joke

Olsen says he started the site in 1998 after following the pronunciation dispute since his BBS days in the late 1980s. The page moved through hosting before landing at OlsenHome.com in 2002 and has continued receiving occasional additions.

Later culture piled onto it. The page references Strong Bad joking about nerds arguing over GIF, the Webby Lifetime Achievement Award where Steve Wilhite again used the soft pronunciation, news coverage, memes, videos, and even Jif peanut butter entering the dispute.

That gives the page a strange historical advantage over a modern explainer. It did not arrive after the controversy and summarize it. It sat inside the controversy while decades of new evidence, jokes, and arguments accumulated around it.

There is also some real file-format history buried under all the yelling: CompuServe, early image viewers, GIF89a, browser-era animated GIFs, PNG’s deliberately unambiguous “ping” pronunciation, and recollections from people who were working with the format before the Web made GIF ubiquitous.

For researchers, those firsthand recollections are the treasure. For webmasters, the lesson is simpler: a narrowly focused page can become authoritative by remaining stable, useful, and available while everybody else keeps rewriting the same subject.

CacheRat’s 1,967 Ancient Web Domains research list exists to find pages like this—documents that survived long enough to become part of the history they originally described.

As for how GIF is pronounced, CacheRat is staying out of it.

We have enough enemies.

Read the evidence and choose your side