Category: Ancient Web

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

Posted on by

Ancient Web: Wonderful World of Windows 3.1 Kept a Real 1990s PC Useful

Retrocomputing pages are often museums.

This one reads more like somebody refusing to stop using the machine.

Visit the Wonderful World of Windows 3.1

The site documents a real Windows 3.1 setup in detail: an AMD K6-2 system, 128 MB of memory, a Matrox G200, Sound Blaster AWE64, IBM DOS 7 underneath Windows 3.1/3.11, and a collection of period software that remained genuinely useful long after Microsoft expected anyone to care.

There are separate pages for applications, graphics, sound and video, games, notes, and word processing.

Not emulation. A living setup.

The applications page lists Netscape Navigator 3.04 Gold, Pegasus Mail, WS_FTP, NetTerm, and mIRC. The graphics section discusses Paint Shop Pro 3.11, Corel Photo-Paint 5, GIF Construction Set, and other tools. The multimedia page explains that the system can still play WAV, AU, RealAudio, MOV, AVI, MIDI, and even MP3 files.

The games page is even better: Doom, Duke Nukem 3D, SkiFree, Minesweeper, Zork, Commander Keen, tiny desktop toys, and odd Windows shareware all coexist on the same machine.

What makes the site historically valuable is the author’s attitude toward limitations.

A notes page asks whether Windows 3.1 can see everything on the modern Internet and answers plainly: no. New video formats, JavaScript, CSS, PNG support, and newer hardware drivers create problems.

But the conclusion is not that the computer is therefore useless.

That distinction matters.

Old software is often judged against everything modern software can do instead of against the jobs the old system still performs well. This site documents the latter from inside the experience rather than from a museum label.

CacheRat’s 1,967 Ancient Web Domains research list includes plenty of dead platforms. Wonderful World of Windows 3.1 is more interesting because the platform was being actively used.

It is preservation through stubbornness.

Browse the surviving Windows 3.1 setup

Posted on by

Ancient Web: Philip KaluÄ‘erÄŤić’s SDF Homepage Keeps the Shell-Account Web Alive

Not every Ancient Web page has to be abandoned.

Some people are still deliberately using the same publishing model.

Visit Philip KaluÄ‘erÄŤić’s SDF homepage

The page lives under a classic tilde-user URL on SDF, the long-running public-access Unix community. It identifies its owner, links to PGP, IRC, Signal, Codeberg, GitHub, and the fediverse, and then exposes the rest of the site almost like a home directory.

There are folders for notes, blog posts, and source code. There is a large links page. Modification dates are visible. Apache’s directory-listing aesthetic remains part of the experience.

In 2026, that feels almost confrontationally simple.

The homepage as a Unix directory

The old personal Web grew partly out of multi-user systems where publishing could be as simple as putting files in the right directory.

A URL like sdf.org/~pkal/ still carries that history in its shape.

KaluÄ‘erÄŤić’s page is modern in content but old-school in ownership. The site points outward to public code, email, IRC, and federated social accounts, yet the homepage remains the stable center. Projects can move between platforms while the personal URL continues to say: this is where my work lives.

That architecture is surprisingly resilient.

A platform profile gives you whatever fields the platform allows. A personal directory gives you files, links, essays, notes, code, odd experiments, and whatever organization makes sense to the author.

The site is also actively maintained. Directory timestamps in 2026 show ongoing additions to notes, blog material, source files, and the homepage itself.

So this is not nostalgia cosplay. It is evidence that the old model still works.

CacheRat’s 1,967 Ancient Web Domains research list is mostly archaeological, but pages like this show continuity too.

The personal homepage did not die. It became unfashionable.

There is a difference.

Browse the SDF homepage and directory

Posted on by

Ancient Web: Carl’s Friends Kept an 1847 Sourdough Starter Moving Through the Mail

This is preservation in the most literal possible sense: the thing being preserved is alive.

Visit Carl’s Friends

Carl’s Friends exists to keep Carl Griffith’s 1847 Oregon Trail sourdough starter in circulation.

The site says more than 108,000 starts have been distributed through it since 2000. The method is gloriously low-tech: volunteers maintain the culture, dry samples, receive mailed requests, and send starter back through the postal system.

The Web handles instructions, history, FAQs, photographs, recipes, and coordination. The yeast and bacteria still travel in envelopes.

Usenet became a preservation society

The story is tied directly to older Internet culture.

Members of the rec.food.sourdough Usenet group formed a preservation effort after Carl Griffith died. The site’s own history describes the group as a creation of the Internet: people who largely did not know each other in real life but shared an interest in keeping the starter alive and freely available.

The site also preserves material from rec.food.sourdough itself. After FAQ maintainer Darrell Greenwood died in 2016, Carl’s Friends copied and relinked the old sourdough FAQ so years of practical discussion would not disappear with a personal site or a dying newsgroup.

That combination makes the project unusually important as Web archaeology.

It preserves both a biological culture and an Internet culture.

The starter has to be fed. The documentation has to be maintained. Mailing instructions change. Volunteers age. Old links break. Postal rates rise. The website records those changes rather than pretending the project exists outside time.

As of fall 2026, the site was still active enough to warn that social-media attention had generated thousands of requests and temporarily overwhelmed distribution.

CacheRat’s 1,967 Ancient Web Domains research list contains plenty of digital artifacts, but few are attached to something you can put in flour and water and watch wake up.

Carl’s Friends turned a website into part archive, part instruction manual, part volunteer network, and part biological relay station.

Read the history and preservation instructions

Posted on by

Ancient Web: Prince of Persia Unofficial Website Became a Fan-Run Technical Archive

Some fan sites age into documentation projects.

Visit the Prince of Persia Unofficial Website

Jalal Noureddine’s site is one of them.

The homepage organizes material around the original Prince of Persia, Prince of Persia 2, and Prince of Persia 3D, but the real depth appears once you start clicking. There are manuals, maps, version differences, passwords, screenshots, ports, fan patches, translations, mods, magazine scans, technical fixes, and community-made material covering years of releases across multiple platforms.

The site identifies itself as the Prince of Persia Unofficial Website and dates its copyright history from 2002 through 2021.

By 2021 it was still posting new material.

Fandom turned into preservation

A casual fan page usually disappears when the author’s interest moves on.

This one accumulated enough detail that it became useful to anyone trying to understand or run old versions of the games.

The Game Gear section, for example, preserves level passwords and manual scans. Other sections document Game Boy and Sega Master System versions, complete level maps, DOS-era tricks, fan fixes, unofficial sequels, language translations, and obscure differences between ports.

That kind of work is difficult to recreate after the fact because commercial publishers rarely preserve every regional manual, patch, odd port, fan discovery, and obsolete compatibility note in one place.

Fans do.

The site’s old-school structure helps. Information is attached to specific games and versions rather than buried in a timeline. A visitor can still move deliberately from one platform to another and compare what changed.

CacheRat’s 1,967 Ancient Web Domains research list contains many examples of fan enthusiasm quietly becoming historical infrastructure.

Prince of Persia Unofficial Website is a particularly good one because the archive includes not only nostalgia, but the practical details needed to study and preserve software.

Explore the surviving Prince of Persia fan archive

Posted on by

Ancient Web: Tokyo Toilet Map Reviewed Public Restrooms in 1996

In 1996, somebody had already built the kind of hyperlocal utility that would later become an app category.

Visit the Tokyo Toilet Map

Yuuji Hayashi’s site photographed and reviewed toilets at train stations and other locations around Tokyo. The English-language page divides them into clean and dirty toilets, includes graffiti, offers a Tokyo map, and even linked to a Flash explanation of how to use a Japanese-style toilet.

The page says visitors had been counted since January 15, 1996, and shows an update date of July 14, 1997.

That makes it an unusually pure snapshot of early practical web publishing.

A location database before location apps

The premise is funny, but the utility was real.

The site tells travelers whether toilets are clean, whether paper is available, what conditions are like, and where the facilities can be found. Contemporary discussion of the page recognized exactly that value. By 2002, a mailing-list post pointed to the project as the obvious basis for a mobile location service, noting that it had already been running since 1996 and had accumulated hundreds of thousands of visits.

The site also captures the tone of the period. It is personal, blunt, sometimes crude, visually simple, and not worried about fitting into a brand system.

A modern equivalent would probably live inside a larger mapping platform with standardized ratings and user accounts. Tokyo Toilet Map instead exposes one person’s judgment directly. The categories, photos, jokes, awkward English, and navigation all belong to the creator.

That individuality is part of the historical value.

It also shows how quickly people understood what the Web could do. Hyperlocal databases did not have to wait for smartphones. Once ordinary people could publish linked pages, they began mapping the mundane parts of city life immediately.

CacheRat’s 1,967 Ancient Web Domains research list contains plenty of grand projects, but pages like this are just as revealing.

The Web was useful because nothing was too trivial to document.

Browse the 1990s Tokyo Toilet Map

Posted on by

Ancient Web: F2.org Built a Personal Web Archive Before Link Rot Won

F2.org looks like somebody tried to save every useful, funny, technical, and inexplicable thing they encountered online.

Visit F2.org

That is not far from what happened.

The site, primarily associated with Fred Curtis, accumulated software links, Windows freeware, Linux resources, coding notes, mathematics, science, books, music, origami, cooking, Australian reference pages, humor, odd documents, and personal material.

Its change log is even more revealing than the homepage.

In 2002, F2.org explicitly created an archive for defunct web pages and freeware considered worth preserving, along with newsgroup articles referenced elsewhere on the site. Years earlier, the maintainer was already documenting broken-link repairs, maintaining a “Graveyard of Defunct Links,” adding internal link checks, and reorganizing sections as the site expanded.

Link rot was already obvious

This matters because the problem was visible early.

The Web was never permanent simply because pages had URLs. Hosting vanished. Software disappeared. Companies moved documentation. Personal pages died when accounts closed. F2.org responded at the small-site level: save copies, update references, document the failures, and keep useful things reachable.

The result is wonderfully messy.

A visitor can move from Perl or Java to nonsense poetry, ancient crafts, science links, freeware utilities, Sydney information, a “Crypt of Eyesores,” and preserved documents that would otherwise be hard to find.

Modern sites are usually encouraged to narrow their topic. F2.org belongs to the older tradition where a homepage could instead become an externalized brain.

The site itself records server crashes, temporary relocations, repairs, and years of maintenance. Its homepage was still being changed in 2012 even though large parts of its structure look much older.

CacheRat’s 1,967 Ancient Web Domains research list exists for the same reason F2’s archive did: useful things disappear unless somebody bothers to notice.

F2.org noticed early.

Explore F2.org’s surviving maze

Posted on by

Ancient Web: FOVICKS Photographed the Los Angeles River as Industrial Archaeology

Some people looked at the concrete Los Angeles River and saw an ecological mistake. Greg Ercolano looked at it and saw a gigantic machine worth studying.

Visit FOVICKS

The wonderfully overbuilt acronym stands for Friends Of Vast Industrial Concrete Kafkaesque Structures.

Ercolano describes the project as a photo journal created by an amateur industrial archaeologist. He follows the Los Angeles River flood-control channel through photographs and commentary, paying attention to survey markers, bridge supports, training walls, storm drains, channel geometry, debris handling, water flow, and the sheer scale of the system.

The page is not merely a gallery of ominous concrete.

It tries to understand why the concrete is there.

Looking closely at infrastructure

Ercolano explains how urbanization changed drainage in the Los Angeles basin and why flood control became such an enormous engineering problem. He studies features that most people pass without noticing: tapered bridge supports, distance markers painted above outlets, confluences between channels, slots in the riverbed, ramps for maintenance vehicles, and the way structures attempt to manage fast-moving debris during floods.

The captions regularly wander into calculations and hypotheses. A surveyor’s number becomes an explanation of station notation. A wall becomes an excuse to think about turbulence. A bridge support becomes a lesson in what happens when a telephone pole comes downstream at speed.

That curiosity is the site’s real value.

The old Web was full of pages produced by people who decided a subject was interesting without first asking whether a market existed for it. FOVICKS is almost a perfect specimen of that instinct.

It is also a reminder that personal photography can become historical documentation. Infrastructure changes. River restoration projects advance. Neighborhoods change. What begins as one person’s strange photo journal can later preserve details nobody else thought to record.

CacheRat’s 1,967 Ancient Web Domains research list is full of pages where obsessive attention accidentally became preservation.

FOVICKS is industrial archaeology with jokes, trigonometry, and a lot of concrete.

Walk the concrete river with FOVICKS

Posted on by

Ancient Web: Virtual Saskatchewan Built an Independent Ezine About One Province

Before local culture was flattened into social feeds, a person could build an entire publication around one place and simply keep adding to it.

Visit Virtual Saskatchewan

Virtual Saskatchewan is a surviving example of that model.

The site identifies itself as an independent ezine about Saskatchewan people and places, with a copyright history running from 1997 through 2013. Writer Dave Yanko used it to publish pieces about local characters, craftspeople, natural curiosities, communities, landmarks, and the kinds of stories that rarely fit neatly into national media.

The archive includes features on dog sledding, Saskatchewan artists, crooked aspen trees, unusual rural hobbies, local businesses, famous Saskatchewanians, petroglyphs, and even a man who built model combines inside bottles.

That range is exactly why the site matters.

A province became the editorial beat

Virtual Saskatchewan feels different from a tourism bureau because the subject is not just destinations. It is texture.

One article can be about a wilderness park. Another can follow a wood-turning artist. Another can document a strange stand of twisted trees. The site treats regional identity as something assembled from hundreds of small observations rather than from slogans.

It also preserves the old independent-ezine structure: editorial writing, photographs, issue-style navigation, and direct links to individual stories. There is no infinite scroll deciding which town or person deserves attention next.

Some of the pieces have outlived their original publication context and become reference material themselves. Later sites and organizations have quoted or cited Yanko’s profiles, including material on Saskatchewan artists and cultural history.

That is the quiet strength of old regional publishing. A local web magazine can eventually become part of the historical record for the place it covered.

CacheRat’s 1,967 Ancient Web Domains research list includes sites like this because the older Web preserved geography at human scale. Towns, regions, clubs, and eccentric local landmarks got their own permanent corners.

Virtual Saskatchewan made a province feel browsable.

Explore the surviving regional ezine

Posted on by

Ancient Web: MĂ­mir Made Icelandic Grammar Navigable by Hyperlink

Some old websites feel less like websites and more like tools somebody quietly left on a workbench.

Visit MĂ­mir, the Icelandic Grammar Notebook

MĂ­mir is exactly that kind of survivor.

The site describes itself as a hypertext-based tool for quickly searching and referencing Icelandic grammar. Instead of trying to become a full online course, it does one thing well: it breaks the language into linked sections covering adjectives, adverbs, articles, conjunctions, nouns, numbers, prepositions, pronouns, and verbs.

Each section is meant to be reached directly. That sounds ordinary now, but it is worth remembering what a good fit early hypertext was for grammar. Languages are full of rules that point sideways into exceptions, related forms, examples, and definitions. A linked document naturally mirrors that structure.

Hypertext before everything became an app

MĂ­mir is useful partly because it never needed an account, dashboard, progress meter, gamification layer, or recommendation engine.

You arrive with a question. You click the relevant grammatical category. You read.

That simplicity is the historical point.

Early educational web publishing often worked like this. Someone with specialized knowledge built a reference, linked its sections together, and left the reader in control of the path through the material.

The site credits John Tebbutt and licenses its pages under Creative Commons. Its design is almost entirely structural: headings, links, examples, and navigation. There is very little between the reader and the subject.

That makes MĂ­mir a good example of a kind of educational Web that deserves preservation. It is not impressive because it is visually elaborate. It is impressive because the Web was used for exactly what hypertext was good at.

CacheRat’s 1,967 Ancient Web Domains research list contains many pages like this: small specialist references that are more useful than their appearance suggests.

MĂ­mir is a language notebook that survived because plain linked information ages surprisingly well.

Browse the surviving Icelandic grammar notebook

Posted on by

Ancient Web: Mike’s Woodwork Kept a Project Log Alive for Decades

A personal website does not have to become obsolete just because it gets old.

Sometimes it simply becomes a better record.

Visit Mike’s Woodwork

Mike’s Woodwork is a long-running project site from Villa Park, California, built around the things Mike makes in his shop and the things he learns while making them.

The homepage is not arranged like a modern portfolio. It reads more like a workshop notebook that happens to be public.

Projects pile up chronologically: boxes, tables, drawers, veneers, dovetails, reclaimed wood, tool handles, signs, repairs, CNC experiments, jigs, finishes, and tutorials created because a particular build taught him something worth explaining.

That accumulation is the site’s real strength.

The project log is a form the Web should never have abandoned

On social media, a finished piece tends to appear as a photograph and a caption.

Mike’s site preserves the mistakes, measurements, materials, design changes, finishing problems, tool choices, and decisions that happened before the final photograph.

A chest of drawers is not just shown. The page explains dimensions, secondary woods, drawer construction, dovetails, guide bars, wax, finish, and how long the build took. A CNC machine arrives and the site records the learning curve, firmware update, router choice, collets, dust boot, spoilboard, and first attempts.

That is useful information because it is specific.

The homepage even asks visitors to email and say how they found the site, which feels wonderfully foreign in the age of invisible analytics. The author is not watching a dashboard. He is genuinely curious who wandered into the shop.

And the site is still being updated.

That is what makes it such a good Ancient Web artifact. It has not been preserved by freezing it. It has been preserved by continuing to use it in roughly the same spirit: one person, one domain, years of projects, and no need to turn the work into a content brand.

CacheRat’s 1,967 Ancient Web Domains research list contains many sites that became historical because their owners stopped touching them.

Mike’s Woodwork is more interesting.

It became historical because the owner never stopped.

Browse Mike’s woodworking project archive