Category: Ancient Web

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

Posted on by

Ancient Web: The Totally Awesome Homepage Was Bad Web Design on Purpose

J. Geoff Malta’s “My ToTaLlY AwEsOmE Homepage” is not an example of somebody accidentally making a terrible 1997 website. It is somebody deliberately making a terrible 1997 website and then discovering that a surprising number of visitors could not tell the difference.

Visit My ToTaLlY AwEsOmE Homepage

The page, copyrighted 1997–2004, is a concentrated parody of personal-homepage habits from the period. There is mixed-case spelling, an “under construction” attitude, material about cars and cats and friends, a fake Web-design business, encouragement to make the page your browser’s home page, visitor-counting culture, and the general visual confidence of somebody who has discovered HTML and intends to use all of it immediately.

The best part is that the site explains the joke—and still had to keep explaining it.

Malta preserved messages from visitors going back to 1997 who criticized the site’s design, offered unsolicited professional advice, or otherwise treated the page as a sincere failure. Those emails turn the parody into something more useful than a gag. They show that the conventions being mocked were common enough that intentional bad design could pass as normal personal publishing.

That is hard to recreate after the fact. A modern designer can imitate tiled backgrounds, counters, gratuitous GIFs, and clumsy navigation, but parody written while those things were current captures a different kind of evidence. It records what Web users already considered annoying, embarrassing, overused, or ridiculous while the habits were still alive.

The page also preserves something the modern Web has mostly standardized away: individual sites were allowed to be weird in incompatible ways. A bad homepage could be bad because one person made twenty highly specific choices. Today’s bad pages often feel bad because they all installed the same five scripts.

For anyone studying early Web aesthetics, this is worth keeping beside the sincere examples. Archaeology needs jokes too. Satire tells you what a culture recognized about itself.

Experience the intentionally awful homepage

Posted on by

Ancient Web: The GIA Preserved SquareSoft News From the PlayStation Boom

The Gaming Intelligence Agency covered Japanese games during the years when SquareSoft seemed capable of turning every RPG announcement into an event. Its old SquareSoft subdomain survives as a piece of that reporting environment rather than a cleaned-up retrospective written years later.

Visit The GIA’s SquareSoft archive

The GIA operated around the turn of the millennium and became known for dense coverage of role-playing games, imports, previews, rumors, and Japanese industry news. The SquareSoft section is especially useful because it catches the company during the PlayStation boom: Final Fantasy, experimental side projects, new labels, staff announcements, and the steady churn of information surrounding games before release.

One surviving text item dates back to 1996 and reports Square launching the AQUES brand name. Other GIA material from the late 1990s and early 2000s circulated apparent design information and rumors surrounding games such as Final Fantasy IX. Contemporary RPG sites also cited GIA reporting on staff and production details for Final Fantasy X.

This material is valuable precisely because it is not hindsight. Modern game histories know which rumors became true, which games shipped, and which projects mattered. Contemporary pages did not. They preserve uncertainty: leaks that might be real, announcements with incomplete details, tentative release information, and the way fans understood a company before the outcome was obvious.

That makes old game-news sites unusually useful research sources. They capture not only what happened but what people knew at a particular moment. A polished database can tell you that a game was announced. A surviving 1999 news item can tell you what the announcement actually looked like before fifteen years of wiki cleanup removed the confusion.

The GIA itself closed in the early 2000s, but its pages continued to be cited and discussed by later writers and former readers. Pieces of its reporting survive under old hostnames like this one, which is exactly why ancient URLs are sometimes worth more than a modern summary.

CacheRat’s 1,967 Ancient Web Domains research list includes many such stranded fragments of old media. The GIA’s SquareSoft pages are a reminder that yesterday’s disposable news can become tomorrow’s primary source.

Explore The GIA’s old SquareSoft pages

Posted on by

Ancient Web: Robotech Research Built an Unofficial Mecha Bible

Robotech Research is a good example of what happens when a fan site stops being a shrine and starts behaving like an unofficial technical manual.

Visit Robotech Research

Associated with Kenneth Olson, the site accumulated material on Robotech mecha, the RDF and REF, role-playing scenarios, equipment, aircraft, and specifications pulled from a mixture of published sources, interpretation, and fan-created material. That last distinction matters. Robotech fandom has always had enough overlapping television edits, RPG books, Japanese source material, and unofficial extrapolation to make “canon” an occupational hazard.

Robotech Research lived directly inside that mess.

Its pages were useful because they tried to turn scattered material into structured reference entries. Machines received designations, performance numbers, armament lists, operational roles, and explanatory notes. For somebody running a Robotech RPG campaign or trying to understand how one vehicle related to another, this was much more useful than a collection of character pictures.

It was also influential enough that later fan discussions still have to untangle its legacy. One example is the VF-4 Lightning. Later Robotech and Macross discussions point back to Olson’s material when explaining how unofficial or speculative equipment became treated by some fans as if it belonged naturally in the Robotech Expeditionary Force inventory. Other technical pages across the fan Web cited Robotech Research for aircraft and equipment details as well.

That does not make the site an official source. It makes it historically interesting for the opposite reason. It shows how decentralized fandom actually manufactured reference culture. A well-organized personal site could become authoritative in practice long before anyone stopped to ask whether every line came from an official book.

The original domain is less reliable today than it once was, but its fingerprints remain in forum discussions, copied specifications, RPG resources, and citations scattered across other old fan pages. That is a familiar form of Web survival: the source weakens while its information keeps circulating.

For researchers, Robotech Research is therefore useful in two ways. It documents Robotech fandom, and it documents how online communities built their own technical continuity around fictional machines.

Explore the original Robotech Research domain

Posted on by

Ancient Web: Pieter Built a Real Anonymous FTP Server in 2026

Most retro-computing projects emulate old software.

Pieter decided to bring back the protocol.

Visit Pieter’s FTP server

The page is wonderfully explicit: this is a real FTP server with anonymous login, part of a non-commercial preservation project. The server lives at ftp://ftp.pieter.com and exposes directories for applications, games, MIDI files, MP3s and other material.

That matters because modern browsers quietly removed FTP support years ago. Chrome, Firefox and other mainstream browsers stopped behaving like general-purpose clients for ftp:// links, turning a once-routine part of using the internet into something younger users may never have seen.

Pieter’s page therefore does two jobs at once.

It hosts files, and it demonstrates the older network model directly.

Before every file needed a web app

Anonymous FTP was brutally simple.

You connected to a server, logged in as anonymous, moved through directories and downloaded files. No account dashboard. No JavaScript bundle. No infinite-scroll card grid. No recommendation engine deciding which ZIP file had better engagement.

Just directories and files.

Pieter’s server deliberately preserves that experience. The web-facing index explains that browsers no longer support FTP and points visitors toward software that still does. It also makes clear that the server exists for preservation and community use rather than as a commercial download service.

The surrounding Pieter.com project goes even further into living retro-computing: old operating systems, dial-up experiments, IRC and other technologies are presented as things you can still interact with rather than merely read about.

That distinction is useful.

A screenshot can show what an FTP client looked like. A working server teaches you what connecting to one actually feels like.

The server index was still being updated in September 2026, so this is not a corpse somebody forgot to unplug. It is a modern reconstruction built with enough fidelity to remain useful as a teaching tool and historical toy.

CacheRat’s 1,967 Ancient Web Domains research list is mostly about surviving old pages. Pieter’s FTP server represents the other side of preservation: sometimes the best way to remember an old system is to make the damn thing run again.

The ancient web was not only HTML.

Some of it was a blinking cursor asking for a username.

Connect through Pieter’s FTP project

Posted on by

Ancient Web: Crashland Preserved Naughty Dog’s Old Crash Bandicoot Material

Corporate game websites are very good at promoting the next release and surprisingly bad at preserving the last one.

Fans often end up doing the archaeology.

Visit Crashland’s preserved official-site material

Crashland hosts a surviving collection of pages and assets associated with the old official Crash Bandicoot web material from the Naughty Dog era. The archive includes production material, character pages, images, and development-era artwork connected to the original PlayStation games.

Its value becomes obvious when you trace individual images.

Animator and character designer Charles Zembillas, who worked on Crash Bandicoot, published some of his old development drawings on his own blog years later. In the comments to a 2013 post showing Baby T concept art from Crash Bandicoot 3, a reader linked directly to Crashland’s preserved copy of the older official material. Another commenter described the original site as a personal “Crash Bible” from the period when it still lived under Naughty Dog’s control.

Other writing about the technical development of Crash Bandicoot has also used Crashland’s preserved production pages as image sources.

That is exactly how a fan mirror quietly becomes historical infrastructure.

The promotional site became a primary source

When these pages were current, they were marketing and fan-service material.

Years later, their function changed.

Production sketches, character-development art, launch graphics, screenshots, and old explanatory pages can help reconstruct how a game was presented while it was still new. They preserve not just the finished product but traces of the development process and the studio’s own public storytelling around it.

The company no longer needs those files to sell a 1990s PlayStation game.

Researchers and fans still need them to understand one.

That mismatch is why so much corporate web history survives only because an individual copied it before the original disappeared.

There is an important distinction here: Crashland is not the original Naughty Dog server. It is a preservation copy maintained elsewhere. That should be stated plainly rather than blurred away. The historical value comes precisely from the fact that material removed from its original context was saved by somebody who thought it mattered.

CacheRat’s 1,967 Ancient Web Domains research list exists for this kind of trail: old corporate pages, fan mirrors, personal citations, and surviving fragments that together reconstruct what the live web once contained.

The official site moved on.

The fan kept the evidence.

Browse Crashland’s preserved Crash Bandicoot material

Posted on by

Ancient Web: Zzickle Transcribed Television One Episode at a Time

Some old fan sites collected pictures. Zzickle did something much more labor-intensive: it sat down and transcribed television episodes line by line.

Visit Zzickle

The site, styled as “The Grand Duchess of Mai’jik,” contains episode transcripts for series including Queen of Swords and Monarch of the Glen. These are not automatically scraped subtitle dumps. The pages explain the practical reason for doing the work: the creator’s father had trouble following Scottish accents and colloquialisms in Monarch of the Glen, while the available DVD closed captions were awkward to use and difficult to read in the setup they had.

So the solution was extremely old-Web: make the missing resource yourself.

Closed captions helped with the transcription, but the result became something different from the television display. Dialogue could be read, searched, checked, and revisited without scrubbing through a disc. Character names and uncertain phrases could be corrected over time. A family accessibility problem turned into a public reference other viewers could use.

What makes Zzickle especially interesting is that it did not simply freeze after the shows ended. Transcript pages have continued receiving revisions years later, including edits in 2026. That is an absurdly long maintenance tail for material that began around turn-of-the-century television fandom, and it demonstrates the quiet advantage of publishing on your own site: nobody needs to approve another correction twenty years later.

There is also a preservation lesson here. Streaming services can make a show feel permanently available right up until licensing changes. Fan-made transcripts are much smaller, simpler artifacts. They do not replace the program, but they preserve names, dialogue, scene order, linguistic details, and searchable textual evidence that video platforms are surprisingly bad at exposing.

The Web once contained thousands of projects like this: somebody had one narrow problem, solved it carefully, and then left the solution where strangers could find it. No growth strategy required.

The larger 1,967 Ancient Web Domains research list contains many such personal reference projects. Zzickle is a particularly good example because the work is still being tended.

Explore Zzickle’s television transcripts

Posted on by

Ancient Web: Early Web Links Is Building a Card Catalog for the Personal Web

The old web did not only lose websites.

It lost ways of finding websites.

Visit Early Web Links

Early Web Links is a modern attempt to rebuild that missing layer. Its premise is blunt: before the internet went corporate, people made websites because they had something to share, and many of those sites are now difficult to discover through ordinary search.

So the site catalogs them by hand.

At the time of this article, the directory reports more than 12,000 sites and sorts them into the sort of categories a human actually browses: literature, programming, retro computing, zines, pets, collecting, astronomy, amateur radio, paranormal material, local communities, recipes and dozens more.

It also has a Random Site button, guestbook, site-of-the-week feature and a stream of new additions.

In other words, it behaves like a directory instead of apologizing for being one.

Discovery without pretending to know what you want

Search engines are excellent when the user has a query.

Directories solve a different problem.

They let somebody browse a neighborhood of ideas without already knowing the destination. That matters for personal websites because their value is often impossible to summarize into a high-volume keyword.

A Viennese zoologist’s giant shrew site, somebody’s tilde page, a homebrew electronics archive or a regional history page may be exactly what you wanted to find ten seconds after you discover that it exists.

Before that moment, there was no query.

Early Web Links leans into period aesthetics, but the useful part is underneath the Netscape buttons and under-construction graphics: human classification. Somebody looked at the site, decided what it was about and placed it near other things that might send a curious person sideways.

That is closer to a card catalog than an algorithmic feed.

CacheRat’s 1,967 Ancient Web Domains research list is obviously operating in the same archaeological territory. The difference is useful: CacheRat is a salvage list; Early Web Links is growing into a browsable directory.

Both exist because the modern web has an absurd amount of information and still manages to hide the interesting stuff.

The solution may not be smarter ranking.

Sometimes you just need shelves again.

Browse Early Web Links

Posted on by

Ancient Web: Opsopaus Built a Hypertext Pythagorean Theology

The old web was extremely good at one specific kind of behavior: somebody becoming obsessed with a subject and publishing the entire rabbit hole as linked HTML.

John Opsopaus did that with Pythagorean theology.

Visit the Pythagorean Theology project

Opsopaus is a writer and longtime practitioner of modern Pagan and esoteric traditions who has also worked in computer science. His site combines those interests in a sprawling study of Pythagorean and later Platonic religious philosophy, symbolism, ritual, theology, and historical sources.

The project should not be mistaken for a neutral academic encyclopedia. It is explicitly interpretive and reconstructive, and some sections move from ancient textual evidence into modern spiritual practice and esoteric theory.

That is part of what makes the site historically interesting.

It shows the web being used not merely to summarize scholarship but to build a personal intellectual system in public.

Hypertext was made for this kind of thing

Dense religious and philosophical traditions are full of names, symbols, cross-references, quotations, competing interpretations, and concepts that make little sense in isolation.

A linked website handles that naturally.

Opsopaus’s project breaks the material into sections and lets readers move sideways through related ideas rather than forcing everything into a single linear essay. The result feels closer to a private research notebook made public than to a modern SEO article.

The site has also had a long afterlife through citation and discussion. Forum posts and later blogs still link directly into individual chapters. A 2024 podcast appearance by Opsopaus points listeners back to the same site and specifically to the first chapter of the theology project.

That is more than simple survival.

It means the old URL is still functioning as the author’s reference edition decades after the web around it changed.

There is another useful lesson here: preservation does not require agreement.

A researcher can treat the site’s historical claims critically while still valuing the artifact. The important thing is to preserve what the author actually argued, how the argument was organized, and which sources and interpretations were connected together.

CacheRat’s 1,967 Ancient Web Domains research list exists for pages like this because unusual intellectual projects are especially vulnerable to disappearing once the person maintaining them stops paying for hosting.

The modern web likes short answers.

Opsopaus built a labyrinth instead.

Explore the Pythagorean Theology project

Posted on by

Ancient Web: EyeEye Robots Documented Walking Machines Before YouTube

Before every garage engineering project had a build video, EyeEye Robots documented the slow, expensive, occasionally ridiculous business of making machines walk into televised combat arenas.

Visit EyeEye Robots

The site belonged to Team EyeEye, a British robot-building group associated with machines such as Jim Struts, Miss Struts, Arnold, and several Techno Games entries. Their specialty was particularly awkward: walking robots. Wheels are easy. Legs introduce balance, timing, linkage geometry, actuator problems, and many more opportunities for something expensive to stop moving five minutes before filming.

That difficulty is why the surviving documentation matters. EyeEye’s old pages and construction diaries record the machines as engineering projects rather than television characters. Miss Struts, for example, was built for Robot Wars and later won a walker battle. Contemporary material associated with the team describes hundreds of hours of work before auditions and then still more changes afterward. Arnold and later variants continued the same basic obsession with making a heavy competition robot move on legs reliably enough to survive an arena.

The team also entered Techno Games, where walking, climbing, sprinting, and other specialized machines competed in events that rewarded engineering rather than destruction. Names including Tina Treks, Rita Rocks, A.A.T., and Jimmy Struts survive in records connected with the group.

The useful thing about an old site like this is that it preserves the process. Television tends to compress fabrication into a few shots of sparks, a voice-over, and then the machine appears in the arena. Personal Web pages kept the ugly middle: mechanisms, revisions, photographs, build notes, failures, and the sort of detail another hobbyist could actually learn from.

A modern video channel could document all of this more dramatically, but it would also be harder to scan twenty years later. A plain page of labeled photographs and technical notes is surprisingly durable research material.

The domain is not always cooperative today, and some of the site’s value now survives through references, mirrors, and later Robot Wars documentation. That makes the original even more worth recording while its trail is still reconstructable.

Explore the surviving EyeEye Robots site

Posted on by

Ancient Web: History of Cuba Turned the Virginius Affair Into a One-Man Reference Project

HistoryofCuba.com is the kind of site that makes the phrase “personal website” sound misleadingly small. Jerry A. Sierra built a sprawling independent reference on Cuban history, and one of its surviving pages covers the 1873 Virginius incident in the middle of the Ten Years’ War.

Read the Virginius page at History of Cuba

The Virginius was a former Confederate blockade runner later used to carry men and supplies to Cuban insurgents fighting Spanish rule. Spanish forces captured the ship in 1873, and executions of crew members and passengers created an international crisis involving Spain, the United States, and Britain. It is exactly the sort of episode that gets compressed into a paragraph in broad histories but rewards a dedicated page of its own.

Sierra’s larger project is the more interesting Web artifact. He has described HistoryofCuba.com as a one-person operation maintained after hours and on weekends. A count he performed on the site found 1,021 files spread across folders and subfolders, including 450 text files and 571 graphics. The project contained about 207,000 words by that count, arranged as timelines, articles, image galleries, book excerpts, and historical timetables.

That scale explains why a page about one nineteenth-century maritime incident can exist without needing to justify itself to an editor. The old Web was unusually good at this. If somebody thought the Virginius deserved its own page, they could simply make one and connect it to a larger body of work.

The project also shows the maintenance burden hidden inside long-lived personal sites. Sierra later wrote publicly about refactoring old HTML, multiple generations of stylesheets, XHTML and HTML 4.01 pages, and JavaScript galleries that had accumulated over years. Some obsolete pieces were removed; others were deliberately kept because nostalgia was itself part of the archive.

That is useful context for anyone preserving an old site today. Longevity is not the same thing as freezing every byte forever. Sometimes survival means repairing the structure while keeping the historical material intact.

Return to the Virginius incident at History of Cuba