Category: Ancient Web

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

Posted on by

Ancient Web: The X-Files Episode Guide Still Looks Like a Fan Built It Because One Did

The X-Files generated exactly the kind of fandom that the old Web was built to accommodate: episode guides, quotes, conspiracy theories, character notes and people willing to maintain all of it long after sane bedtime hours.

Visit the X-Files Episode Guide

This surviving fan guide covers every episode of the series and was updated beyond the original nine-season run to include the 2016 and 2018 revival seasons. The author also selected a favorite or notable quote for episodes and openly invites arguments over those choices.

That last part is important. This is not a sterile database. It is somebody’s guide.

Before Streaming Apps Did the Indexing For You

Modern viewers can usually get an episode title, season number and synopsis from a streaming interface. Older fan sites went further because the fans had to build the information layer themselves.

The guide includes season-by-season pages, episode summaries, related links and printable pages. The fourth-season archive, for example, runs from “Herrenvolk” onward with plot summaries written for readers who were navigating the show as a body of episodes rather than clicking a thumbnail grid.

The site also preserves the aesthetic vocabulary of television fan pages: show logos, promotional imagery, simple tables and links to official and unofficial resources living side by side.

Fandom as Indexing Labor

Before giant entertainment databases absorbed most search traffic, fans did enormous amounts of unpaid cataloging. They transcribed credits, identified production codes, organized continuity and argued over quotations because nobody else was going to do it for them.

That work made the Web useful.

The X-Files is an especially fitting subject because the series itself rewarded obsessive cross-referencing. Monster-of-the-week episodes sat beside a sprawling mythology, and knowing where something happened often mattered.

This guide remains a compact record of that fan-maintained information culture. It is neither the largest nor the slickest X-Files resource ever made. It does not need to be.

It survived, it still works, and it still reads like somebody cared enough to keep track.

Browse the X-Files Episode Guide

Posted on by

Ancient Web: Brave New World as a Chapter-Indexed HTML Book

Before online reading meant an account, an app, a cookie wall and a toolbar asking whether you would like the toolbar to stop existing, a book on the Web could be eighteen numbered links and some HTML. Hedweb’s Brave New World page is a surviving example of exactly that approach.

Visit the Hedweb Brave New World page

The page presents Aldous Huxley’s 1932 novel as a sequence of 18 linked chapters. There is no simulated paper, page-turn animation, reading-progress database or recommendation carousel. Chapter numbers sit at the top and bottom. Click one. Read. Click the next one.

That simplicity makes the page interesting even apart from the text it contains.

The Book Is the Interface

The page is mostly typography and hyperlinks. It opens with the title, Huxley’s name and dates, then immediately exposes the chapter navigation. At the bottom, the reader can jump into a wider collection of Huxley-related material, including a character guide, links concerning Island, material about Brave New World Revisited, a Huxley photogallery and other essays and projects.

In other words, the page behaves less like an ebook product and more like a small hand-built research neighborhood.

That distinction matters when looking at older Web publishing. HTML already had nearly everything needed for a basic digital book: headings, paragraphs, anchors and separate documents. A long text did not require a dedicated reader. The browser was the reader.

Durable Because There Is Almost Nothing to Break

The page also demonstrates one of the recurring lessons in the Ancient Web collection: boring technology ages extremely well.

A numbered link to another HTML file does not care which JavaScript framework is fashionable. It does not require an API endpoint to remain funded. There is no client-side application that has to compile against a dependency tree from fifteen years ago. The structure is visible to browsers, search engines, accessibility tools and human beings without much negotiation.

The legal availability of books can vary by country and copyright term, so CacheRat is not treating this as a recommendation about where a particular reader should obtain the novel. What is worth documenting here is the web form that survived: a complete long-form reading interface reduced to ordinary documents and links.

Modern publishers spend a lot of effort recreating properties the Web already had. Hedweb’s page is a useful reminder that sometimes the correct amount of interface is barely any interface at all.

Return to the chapter-indexed Brave New World page

Eighteen chapters. Eighteen links. Somehow the browser survives the workload.

Posted on by

Ancient Web: Mike Willegal Rebuilt the Apple I and Revision-0 Apple II From the Board Up

There is collecting old computers, and then there is rebuilding the earliest Apple machines closely enough that the circuit board itself becomes a historical document. Mike Willegal’s site lives firmly in the second category.

Visit Mike Willegal’s Apple II pages

Willegal became known in the vintage-computing community for projects recreating the Apple I and the earliest revision of the Apple II motherboard. His pages preserve far more than finished glamour shots: board layouts, assembly information, troubleshooting notes, scans, downloadable documentation and years of project history.

One of the best surviving pieces is an Apple II assembly and bring-up guide. It reads like real bench documentation because it is. The guide walks through tools, donor components, ROMs, memory, monitors and the practical reality of powering up newly assembled vintage-compatible hardware.

The Mimeo 1 and Revision-0 Apple II

Willegal’s Apple I project became known as the Mimeo 1. Contemporary posts from the late 2000s show him working through component sourcing and PCB reproduction while trying to keep the design faithful to the original machine.

His Revision-0 Apple II work similarly documented the earliest board configuration rather than merely producing a generic Apple II clone. Other preservation projects later built on that work, which is a strong sign that the documentation mattered beyond the original batch of boards.

The site also preserves correspondence with Steve Wozniak about early Apple design decisions. One exchange discusses the Apple I video circuitry and Wozniak’s recollection that it grew out of his earlier terminal, Pong and Breakout-related design experience rather than Don Lancaster’s TV Typewriter.

That kind of firsthand technical context is hard to replace once scattered forum posts and personal pages disappear.

A Workshop Notebook That Accidentally Became History

The best thing about Willegal’s site is that it was not built as a museum exhibit after the fact. Much of it was written while the work was happening.

You can see problems being found, boards revised, documentation improved and obscure early-Apple details being argued over in public. That makes the site useful to restorers, collectors and historians in a way a cleaned-up retrospective often is not.

Explore the Apple I Mimeo project

Posted on by

Ancient Web: TheOldNet Put the Archived Web Back on Vintage Computers

TheOldNet attacked a problem that digital preservation usually ignores: an archived 1990s website is not much fun if the archive itself requires a modern computer to view it. The project built a layer between archived pages and old hardware so vintage browsers could actually get back onto the old Web.

Visit TheOldNet

The project appeared in 2019. Its creator described two annoyances with browsing old pages through the Internet Archive. First, archive pages can include modern scripts and interface code that old browsers cannot handle. Second, an archive is easy to search when you already know what you want, but less useful for aimless exploration.

TheOldNet’s answer was a proxy and page parser sitting in front of archived material. It stripped away modern additions and presented the old pages in a form that vintage systems could deal with more easily. The project also built discovery features so users could wander instead of supplying a perfect historical URL.

Old Computers, New Plumbing

The best part is that the project did not stop at a themed website for modern browsers.

TheOldNet developed services around actual retro hardware. A serial Wi-Fi modem emulator can connect an old computer to modern Wi-Fi while pretending to be a Hayes-compatible dial-up modem. To the vintage machine, there is no magical modern network card. A terminal program still sends familiar AT commands and “dials” a network address.

There is also a Commodore 64 BBS project. Its text browser was designed to reach TheOldNet from PETSCII-era hardware, with additional BBS directory and chat features. The project’s documentation lists baud rates from 300 through 115200, which is a delightful reminder that “web performance” means something different when the computer on your desk considers 2400 baud a respectable afternoon.

Preservation You Can Actually Use

This approach is valuable because software preservation is not only about storing files. Context matters. A Netscape-era page viewed in a modern Chromium window is preserved information, but it is not the same experience as loading it through hardware and software that existed near the time the page was made.

TheOldNet provides some of that missing bridge. It lets old computers participate rather than sit powered off beside a modern machine showing screenshots of them.

The main site has had periods of instability and its current front page can be extremely minimal, but the surrounding project remains documented through its code, proxy work, BBS and modem tools. That makes it worth recording as more than a nostalgia skin.

Explore TheOldNet and its vintage-computing services

The Web was designed to work across different machines. Making it work again on machines thirty years old is a pretty good demonstration of the idea.

Posted on by

Ancient Web: Abandoned But Not Forgotten Cataloged Empty America State by State

Before urban-exploration photography became an endless social feed, Abandoned But Not Forgotten was already collecting empty malls, hospitals, factories, schools and strange roadside ruins into state-by-state HTML pages.

Visit Abandoned But Not Forgotten

The site’s structure is gloriously direct. Pick a state. Pick a place. Look at the photographs. The archive spans everything from Pennsylvania institutions and New Jersey ruins to Missouri’s abandoned Bannister Mall and forgotten industrial buildings.

The project also has a very clear rule: do not ask for locations. The site explains that many photographs were submitted by other people and often arrived without exact location information anyway. That practical limitation became part of the site’s identity.

Built From Submissions, Not a Content Calendar

The site’s index thanks Rob Pless for starting the project and says it was later taken over so the material could remain preserved. It openly asks for new submissions and credits contributors on many individual galleries.

That creates a different kind of archive from a modern commercial photography site. The pages vary wildly. Some contain a line or two of context. Some are almost entirely photographs. Some document places that have since been demolished or redeveloped.

That inconsistency is part of the value.

One Missouri page lists abandoned malls, buildings, barns, silos, a shoe factory, a water-slide park and Wheatley Sanitarium. The old Bannister Mall page records the shopping center after abandonment and notes how surrounding businesses had also closed. Elsewhere, the site preserves large image sets from places such as Pennhurst State School and Rosewood State Hospital.

A Web Archive of Physical Disappearance

Most preservation projects worry about websites disappearing. Abandoned But Not Forgotten has the opposite problem too: many of its subjects disappeared in the physical world after the photographs were posted.

That makes the site useful twice over. It is an artifact of an older style of community Web publishing, and it preserves visual evidence of buildings that may no longer exist.

The design is old, repetitive and image-heavy. Good. A slick redesign would not make the photographs more valuable.

Browse the Abandoned But Not Forgotten archive

Posted on by

Ancient Web: Historical Boys’ Clothing Became a 30,000-Page Social History Archive

Historical Boys’ Clothing is one of those websites where the name radically undersells the thing. What began as a project about children’s clothing has grown into a sprawling independent archive covering fashion, childhood, countries, schools, youth organizations, photography, economics and social history.

Visit Historical Boys’ Clothing

The site’s own history says the project began in 1996, initially on free hosting. Christopher Wagner was interested in children’s fashion and needed help with HTML; the collaborator who took on the web work had previously taught history and social sciences. The project gradually expanded from clothing into a broader attempt to document how children actually lived through different historical periods.

That expansion got out of hand in the most productive possible way.

Tens of Thousands of Pages

The site currently estimates roughly 30,000 pages and more than 40,000 images across HBC and associated sections. Earlier versions of the homepage already described tens of thousands of pages. Most of the material is arranged through old-fashioned hierarchical navigation rather than a modern infinite feed.

Visitors can work by garment, chronology, country, activity, photography, biography, school, uniform and other categories. The archive reaches across more than 100 countries and includes material on both boys and girls when broader context is useful.

A large part of the collection is photographic. Period portraits and snapshots are used not simply as decoration but as evidence: what children wore, how formal clothing changed, when certain garments disappeared, and how ideas about childhood differed between places and decades.

The underlying argument is sensible. Clothing is visible social data. A photograph of a child in a sailor suit, school uniform, work clothes or Sunday best can tell you things that a conventional political history usually does not bother recording.

A Website Built by Accumulation

The site is also a remarkable example of how enormous the independent Web could become without turning into a conventional media company. HBC moved away from free hosts because of banners, pop-ups and other clutter, then spent decades adding pages under its own control.

Its pages are not polished into a single modern design system. There are spelling mistakes, deeply nested navigation trails and sections accumulated at different periods. But the information architecture is understandable, the URLs are linkable, and a huge amount of niche material remains reachable through ordinary HTML.

The site was still showing updates in 2026. That means a project founded before Google has survived through portals, blogs, social networks, mobile apps and multiple rounds of people announcing that websites were dead.

Researchers interested in clothing or childhood history can use it directly. Webmasters can study something else: what happens when a website is allowed to become a life’s work instead of being redesigned into oblivion every three years.

Explore the Historical Boys’ Clothing archive

Thirty thousand pages later, apparently the answer is: you end up with an encyclopedia.

Posted on by

Ancient Web: A Family Trip Through Ontario’s Bruce Peninsula, Preserved as a Plain HTML Travel Page

Some old travel pages are more useful than polished destination guides because they preserve exactly what somebody noticed while actually being there. This University of Waterloo-hosted page about Ontario’s Bruce Peninsula is one of those.

Visit the Bruce Peninsula travel page

The page is a small collection of family travel notes from places around the peninsula, including Stokes Bay, Tobermory, Dorcas Bay and Singing Sands Beach. The writing is compact and practical: whether the water was warm, whether the beach was sandy or rocky, whether children could comfortably wade, and whether a place was difficult to find.

There is no attempt to turn the trip into lifestyle content. Nobody optimized a headline around “ten hidden gems.” The page simply records what happened.

The Kind of Detail Search Results Usually Lose

One note about Stokes Bay says the first beach the family found had no signs or toilets and that they were unsure whether it was public or private. Another says a local person had to point them toward Black Creek Provincial Park Beach. At Singing Sands, the writer remembers water so shallow that young children had trouble finding a spot deeper than their waists.

Those are small observations, but they are exactly the sort of details that vanish when travel information becomes standardized.

The page also preserves spelling mistakes, uneven phrasing and the author’s own priorities. That matters. Personal Web pages were often documents first and products second—or not products at all.

A Tiny Piece of Geographic Memory

The page survives under the University of Waterloo’s public web space, a reminder of how universities once hosted large amounts of personal material alongside research and institutional information.

For researchers interested in the old Web, pages like this show how ordinary people used server space: family trips, photographs, practical notes and local knowledge arranged in simple HTML.

For travelers, it remains a snapshot of how these places were experienced before review platforms became the default way of describing them.

Explore the Bruce Peninsula page

Posted on by

Ancient Web: Andrew Clem Built a Baseball Stadium Atlas by Hand

Andrew Clem’s website is the sort of personal domain that makes modern profile pages look malnourished. It contains years of writing, birding records, travel material and politics, but the spectacular part is a hand-built atlas of baseball stadiums going back more than a century.

Visit Andrew Clem’s website

The baseball section describes itself, reasonably enough, as “baseball stadium heaven.” Its core is a large set of stadium pages covering current and former Major League Baseball parks back to 1909, plus miscellaneous and non-MLB parks Clem has visited or researched.

Stadium Diagrams as a Long-Term Obsession

These are not copied seating charts. Clem draws the parks from a consistent top-down perspective and keeps refining them as better information becomes available.

The diagrams document field shape, grandstand geometry, decks, dimensions and historical changes. On some pages a mouse rollover can switch between different configurations of the same stadium. His Mile High Stadium page, for example, can show football and baseball layouts from different eras. Older stadium pages include chronology notes showing when their diagrams were revised.

That revision history is part of the charm. This is not a database dumped once and forgotten. It is somebody returning to the same research project for years, correcting the shape of an upper deck because apparently peace was never an option.

Clem also standardized the home-plate position in newer diagrams so parks can be compared side by side or overlaid. That is the kind of small design decision that turns a hobby site into a genuinely useful research tool.

Then There Are the Birds

Baseball is only one branch. The site also contains a long-running wild-bird diary with photographs, life lists, annual arrival dates and species indexes. Clem’s bird-photo pages let visitors browse photographs by species, sex, age and other variations.

There are also travel photographs, personal archives and years of blog material. The site’s navigation still reflects the older idea that one domain could simply be a person’s entire information universe rather than one narrowly branded project.

The candidate page CacheRat found sits inside the site’s old archive structure, but following it outward reveals the larger point: the archive never became a dead end. The domain is still being maintained, and some baseball pages show updates and visits from 2026.

For webmasters, there is a lesson here. A specialized reference does not need a huge backend or a corporate editorial staff to become valuable. It needs consistent structure, durable URLs, accumulated expertise and an owner stubborn enough to keep fixing things.

Explore Clem’s baseball stadium diagrams

The result is exactly the kind of thing search engines are supposed to help us find: one person’s extremely specific knowledge, organized well enough that strangers can use it.

Posted on by

Ancient Web: Kozmic Dreams, Art Cars and a Handmade Personal Web

Kozmic Dreams is what happens when somebody treats a personal website like a house with too many rooms and keeps adding rooms anyway. The site mixes artwork, photography, pets, collections, a virtual museum and a pair of heavily documented art cars into one long-running independent web project.

Visit Kozmic Dreams

The site says its material dates back to the 2000s, and its current navigation still feels like a personal homepage rather than a portfolio assembled by a template service. You can enter through ordinary links or take a “mystery tour” by clicking objects in the artwork on the front page. That is an extremely old-Web solution to navigation: make the interface part of the fun and trust the visitor to poke around.

Two Cars Became Giant Canvases

The best material is probably the art-car section. The creator describes transforming a 1988 Pontiac 6000 STE into the Kozmic Dream Machine in 1998, spending roughly 150 hours airbrushing it with a cosmic design. The car later appeared in five Woodward Dream Cruises before the engine finally died.

The successor was stranger and more ambitious. A 1989 Ford Thunderbird inherited from the creator’s grandfather became the Music Machine, with about 200 hours spent on the taping and painting alone during the summer of 2003.

The hood carries a large Frank Zappa portrait. Zappa’s hair turns into a musical staff that runs along the sides of the car, with part of the guitar solo from “Rat Tomago” worked into the design. The car later won its class at the 2005 Detroit Autorama and won Best in Class at the 2008 Autorama Extreme.

The site does not merely show the finished objects. It preserves the process: body work, painting, detail photographs, old scanned images and event pages. That makes it more useful than an Instagram gallery would have been. The sequence survives.

A Personal Site That Never Flattened Itself

Elsewhere are watercolors, ceramics, glasswork, pyrography, macro photography of insects and spiders, Detroit art, family artwork and years of photographs. Some pages still carry the small-image, link-heavy structure of the personal Web even though the site has continued to receive updates.

That is the real value here. Kozmic Dreams did not become a generic artist landing page with three giant photographs and a contact form. It remained a collection of interests.

For anyone studying personal publishing, that matters. The old Web was full of sites where the organizing principle was simply this is the stuff I care about. Search engines could index every odd little page, and a visitor could wander from airbrushed Thunderbirds to spiders without being redirected into a feed.

Explore the artwork, photography and art-car archive at Kozmic Dreams

Somebody kept the doors open. That is enough reason to go wandering.

Posted on by

Ancient Web: Kirby’s Dream Site Rebuilds the Fansite Internet on Neocities

Not every Ancient Web find has to be a page that physically survived from 1997. Kirby’s Dream Site is a modern Neocities fansite that deliberately brings back the older Web’s habits: handmade navigation, site buttons, fan shrines and links to other people’s pages instead of an algorithmic feed.

Visit Kirby’s Dream Site

The site’s “Friends of Kirby” section is especially revealing. It points visitors outward to other Kirby resources, including Kirby’s Rainbow Resort, which it describes as probably the longest-running Kirby fansite, along with a surviving official Kirby Super Star website and smaller fan projects.

There is even an old-fashioned site button made by another webmaster and a page preview presented through an iframe. None of this is accidental. It uses the visual and social grammar of the personal Web.

The Link Is Part of the Content

Modern social platforms encourage creators to keep visitors inside the platform. Old fansites routinely did the opposite. A page about Kirby naturally linked to another page about Kirby because discovering somebody else’s shrine was part of the fun.

Kirby’s Dream Site recreates that structure. The value is not simply the information on one page; it is the small network of independent sites around it.

That makes Neocities projects like this interesting to an Internet-history project even when they are newer than the material they imitate. They show that the old model of web publishing is not technically obsolete. People can still make a weird little corner of the Internet, decorate it however they want and send visitors somewhere else without asking permission from a platform.

The result feels less like browsing a content product and more like finding someone’s room with the door open.

For Kirby fans, it is another fan-built resource. For webmasters, it is evidence that the fansite format survived because the format still does something centralized feeds do badly: it gives an interest a place.

Explore Kirby’s Dream Site and its outgoing fan links