Category: Ancient Web

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

Posted on by

Ancient Web: Michael Scharrer Kept the POV-Ray Source Code With the Art

A rendered image is nice. A rendered image plus the scene file that generated it is much more useful. Michael Scharrer’s POV-Ray collection keeps both, turning a gallery of strange computer graphics into a small archive of executable ideas.

Visit Michael Scharrer’s POV-Ray Scenes

The collection includes titles like Strange Crystal, Debris, Energy Pulse, Nuclear Containment, Tree of Squares, Liquid Fire, Woven Sines, Infinite Spring, Quarks, Fractal Supernova, Sphere Flake, and Caustics. They sound like filenames found on a graphics workstation somebody forgot to shut down in 1998, which is part of the charm.

The useful part is the documentation behind them. In Strange Crystal, for example, Scharrer explains that the central object is built from the union of 40 randomly rotated cubes. A chunk is removed and placed in front using constructive solid geometry. Multiple colored lights at slightly different locations fake wavelength-dependent caustics, while repeated refraction creates the crystalline glimmer.

Then the page links directly to the source code.

That is an old-school computing habit worth preserving. The image is not treated as the final object. The method is part of the work. Someone interested in POV-Ray, constructive solid geometry, procedural art, random seeds, light behavior, or simply how another person thought through the scene can inspect the machinery rather than reverse-engineering a screenshot.

Scharrer also makes the material available under a Creative Commons Attribution license. That reinforces the same idea: these are not only pictures to admire but building blocks somebody else can learn from and reuse.

POV-Ray itself comes from a style of graphics work that feels increasingly alien next to modern 3D suites. You describe a world in text—objects, materials, lights, transforms, mathematical relationships—and let the renderer turn the description into an image. There is a particular satisfaction in changing a few parameters and getting an entirely different universe back.

The website matches that philosophy. It is direct, searchable, and built around the artifacts. No infinite scroll. No account. No portfolio animation that makes you chase the actual image around the screen. Pick a scene, read how it works, take the source code if you want it.

For programmers interested in visual work, this is the kind of page that can still teach more than a gallery with ten thousand polished renders and no explanation whatsoever.

Browse the complete POV-Ray scene collection

Posted on by

Ancient Web: Tony Jiang Wrote a Complete Super Smash Bros. N64 Guide

Before every game mechanic had a wiki page, somebody had to sit down and write the damn guide.

Tony Jiang did.

Visit Tony’s Super Smash Bros. N64 guide

The site covers the original 1999 Super Smash Bros. for Nintendo 64: basic mechanics, single-player play, multiplayer strategy, items, bonuses, every playable character and every stage.

Jiang explains that he bought the game in 2003, after the N64 was already outdated, because he had enjoyed playing it with friends. Then he built a guide from his own experience.

There is an especially good period detail in his introduction. He says the guide is mostly text because he did not have Nintendo’s permission to use game images, making screenshots would take work, and images would consume space on his web server.

That sentence is practically a fossil layer.

Today bandwidth and storage are cheap enough that a strategy site can casually autoplay video. Here, screenshots were an editorial budget decision.

Tested on an emulator, verified on the real machine

Jiang says he used an emulator and ROM image to test moves because it was more convenient, then verified the results on his actual Nintendo 64.

That is a surprisingly solid little methodology for a fan guide.

The site is structured rather than dumped into one giant walkthrough. Each character gets move lists and strategy pages. Stages get their own guides. There are separate sections for single-player scoring, items, multiplayer and emulation.

It is also personal in the way old game sites usually were. The author is not pretending to be an editorial institution. He tells you what he played, why he liked it, what tools he used and what he thinks is worth reading.

That matters in 2026 because game documentation increasingly gets detached from the people who actually discovered and tested things. A surviving fan guide preserves not just information about a game, but the process by which players learned it.

CacheRat’s 1,967 Ancient Web Domains research list is full of pages like this: narrow, stubborn little knowledge bases that were useful long before somebody figured out how to monetize “content.”

Tony’s guide does not need a redesign.

It needs to stay online.

Read Tony’s N64 Smash guide

Posted on by

Ancient Web: Zat Solo’s Deep Space Outpost Kept Frontier Alive

Before game wikis flattened everything into standardized tables, a useful game site might be one person’s orbital outpost.

That is basically what Zat Solo built.

Visit Deep Space Outpost

The surviving address belongs to the University of Bologna domain and was known in old Elite and Frontier circles as Deep Space Outpost. Period Usenet posts from 2000 and 2001 repeatedly point players there for Frontier: Elite II and Frontier: First Encounters material.

One alt.fan.elite reply simply tells a player that there are “loads of links” on Zat’s Frontier page. Another message carries Zat Solo’s signature line: “Deep Space Outpost – The final Frontier.” Contemporary gaming discussion also linked the site for information on the Elite series when people were trying to compare versions, patches, utilities, and the quirks of running old DOS-era space sims on newer systems.

Even Amiga Active pointed readers toward the same Frontier resource in 1999.

That matters because the site was not operating in a world where every old game already had a polished wiki, Discord, subreddit, YouTube archive, and Steam guide.

It was part of the infrastructure.

A fan page could be the documentation layer

Games like Frontier generated unusually technical communities. Players discussed flight models, mission systems, display behavior, patches, ports, emulation, and fan-modified executables. If official support was thin or aging, the useful knowledge often migrated to personal pages and newsgroups.

Deep Space Outpost sat inside that ecosystem.

The surviving references are almost as important as the page itself because they show how information moved. A player asked a question on Usenet. Another player answered with a direct URL. Someone else linked a utility. A magazine printed the same address. Search engines were not the center of the system; people were.

That model looks primitive only if you ignore how resilient some of it became.

More than twenty-five years later, the old URL still identifies a real piece of Elite fandom and can still be traced through discussions written by the people who used it at the time.

CacheRat’s 1,967 Ancient Web Domains research list is built for exactly this kind of archaeology: not merely finding an old page, but recovering the network of people and references that made the page useful.

The modern web calls that community knowledge.

The old web just called it Zat’s site.

Visit Deep Space Outpost

Posted on by

Ancient Web: DAN-BALL Kept Browser Games Alive for Twenty Years

DAN-BALL started on April 16, 2006. Twenty years later, it is still updating games. That alone puts it in a rare category of browser-game sites that did not simply survive the death of Flash, social platforms, app stores, and several generations of Web fashion. It kept shipping.

Visit DAN-BALL

The site is managed by three people identified on its profile page: ha55ii, a game programmer; tylor, a web programmer; and otascube, an app game programmer. Their catalog is wonderfully difficult to summarize because the games often sound like experiments somebody accidentally made playable.

There is Powder Game, a falling-particle simulation where players can combine materials and build contraptions. Liquid Webtoy turns fluid simulation into a browser toy. Planet simulation plays with gravity. Compasses lets hundreds of little magnets interact. Earth Editor builds worlds from dots. 100bit is a shared drawing surface that can replay the history of how an image was made. There are also action RPGs, puzzle games, hamsters, rockets, pandas, mines, ray tracing, and enough stick figures to staff a very cheap war.

The update history is almost as valuable as the games themselves. It stretches back to the site’s launch date and records a constant stream of small releases: Rockets 1.0 on April 16, 2006; Compasses 1.0 that August; SAND MOJI 1.0 in September; Powder Game 1.0 in April 2007. The same log continues into 2026 with current versions of Stick Ranger 2 and mobile-app maintenance.

That continuity matters. Browser games are usually discussed as a dead era bounded by plugin technology. DAN-BALL shows another path: keep rebuilding, porting, maintaining, and letting the site remain the center of gravity instead of abandoning it for whatever distribution platform is fashionable this year.

The comments are still active too. People show up in 2026 talking about games they played as children, sharing creations, requesting ports, and celebrating the site’s twentieth anniversary. There is an actual community thread running from mid-2000s Web games into current browsers and phones.

The design has changed enough to keep working but not enough to erase its ancestry. There are still giant access counters, plain navigation, update logs, individual game pages, and a sense that the homepage is a directory rather than a marketing funnel.

For a webmaster, that may be the most interesting part. DAN-BALL did not need to become a giant platform. It just needed to remain a place people knew how to return to.

Explore twenty years of games at DAN-BALL

Posted on by

Ancient Web: BamSoftware Turned 10 MB Into 281 TB Without Recursion

Some old technical pages are useful because they document forgotten hardware.

Others calmly explain how to turn 10 megabytes into 281 terabytes.

Read A better zip bomb

David Fifield published “A better zip bomb” in July 2019 and kept updating the page afterward. The important trick is that this is not the familiar recursive ZIP bomb built from archives nested inside archives.

Fifield’s construction overlaps files inside the ZIP container so many output files can refer back to the same compressed data. The result expands in a single decompression pass and can reach a compression ratio above 28 million to one.

At the limits of the ordinary ZIP format, the example grows from roughly 10 MB to 281 TB. With 64-bit ZIP extensions, the numbers get even more absurd.

The page includes the explanation, source code, sample archives, diagrams, technical notes and later corrections. It became a proper security paper at the 2019 USENIX Workshop on Offensive Technologies, where it won Best Paper.

The web page is the lab notebook

What makes the BamSoftware version worth keeping is not merely the headline number.

The HTML page is a research artifact.

It shows the reasoning, links the implementation, records updates by date and leaves enough detail for another technical reader to understand what is actually happening inside ZIP and DEFLATE rather than repeating “compression bomb bad” from a security glossary.

That kind of page used to be common on the technical web. Somebody found something strange, wrote the whole thing down, published the code and kept the original URL alive.

There is also an important distinction here: the classic 42.zip depends on recursive unpacking for its ridiculous expansion. Fifield’s design specifically tries to remove that requirement. That is why the work was interesting enough to become a conference paper instead of another malware anecdote.

CacheRat’s 1,967 Ancient Web Domains research list exists for exactly this kind of technical rabbit hole: a compact page that contains far more useful information than its visual footprint suggests.

The modern web has no shortage of security explainers.

Very few of them hand you the bomb diagram, the source code and the math.

Dig into the ZIP bomb research

Posted on by

Ancient Web: NES City Preserved the 2003 NES Fan Scene

Before retro gaming became a polished YouTube category, people built entire little towns around one old console.

NES City was one of them.

Visit NES City

The surviving site says NESCity.com dates to late December 2003, when two older fan sites run by NES Monster and Weird-Edsel were merged into one project. The archive is still there, and it is less interesting as a conventional review site than as a record of what a game community actually looked like in 2003 and 2004.

Updates mention AIM icons, desktop backgrounds, Winamp skins, IRC chat rooms, a forum, an RSS feed, Flash banners, NES MP3s, sprite work, game shrines, reader submissions, merchandise, and random site-to-site advertising.

One update announces a detailed King’s Quest V shrine. Another discusses a Ground Kontrol interview. Others document a site Easter egg hunt, browser problems, forum migrations, a Jabber contact option, and the revelation that an AIM icon for DuckTales had grown beyond the terrifying limit of 7 KB.

This is what web history looks like when nobody knew they were making web history.

The useful part is the mess

Modern retro-gaming databases are much better at answering clean questions. NES City is better at answering messy ones.

What did fans make for each other? Which sites did they link to? What counted as a major update? Which technologies did hobby webmasters experiment with? What disappeared when another fan site died?

The site’s own archive casually records those answers.

Its “Fake FAQ” even explains that unfinished sections are crossed out and jokes about creating FAQ entries before anyone had asked the questions. That tone matters. Early fan sites were not publications trying to sound institutional. They were places run by recognizable people who were visibly improvising.

NES City also survives inside the broader NESHQ ecosystem, itself one of the long-running NES reference communities. That continuity gives this little 2003 project something many contemporary fan pages never got: a functioning address decades later.

CacheRat’s 1,967 Ancient Web Domains research list exists partly for this kind of page. The value is not that every old fan opinion was profound. The value is that the social texture of the web is otherwise extremely easy to erase.

A database can tell you when Super Mario Bros. 3 shipped.

NES City can remind you that somebody once had to convert its AIM icon to black and white because the file was too damn big.

Browse NES City

Posted on by

Ancient Web: The Simple Sailor Kept Ocean Voyaging Small

Roger Taylor’s website begins with a position that sounds almost hostile to the modern marine catalog: ocean sailing has become needlessly complicated, bigger boats are not automatically better, and simplicity is a serious design choice rather than a lack of equipment.

Visit The Simple Sailor

Taylor is the skipper associated with Mingming, a small junk-rigged Corribee, and later Mingming II. The site has been online in this form since at least 2007 and grew into a substantial collection of voyage reports, articles, photographs, video clips, heavy-weather material, whale encounters, and books.

The best part is that the site does not merely tell you Taylor sailed small boats into serious water. It preserves the voyage records.

His 2018 account, for example, follows Mingming II north from Whitehills on Scotland’s Moray Firth toward Svalbard and Franz Josef Land. Eighteen days and roughly 1,200 miles out, he closed the coast of Prins Karl’s Forland. He later passed north of Ross Island at roughly 80 degrees 50 minutes north and pushed farther east before turning back near 81 degrees 15.3 minutes north, where current and pack-ice risk made continuing a bad bargain.

This is not marina lifestyle writing. The site includes fog, unsounded water, Force 9 weather, bare-poles running, ice glare, navigation decisions, and the practical reasoning behind using a small, stiff, watertight boat.

The older pages also reach much farther back than the website itself. Taylor includes photographs of building his first ocean cruiser, the 19-foot Roc, in the early 1970s; of Roc after storm damage during the 1974 Singlehanded Trans-Tasman Race; and of his time aboard the ill-fated three-masted barque Endeavour II in 1971.

That gives the site an unusual quality. It is simultaneously a personal homepage, a sailing philosophy, a primary-source voyage archive, and a bridge between pre-Web experience and Web-era self-publishing.

For a modern webmaster, there is another lesson hiding in the rigging. The site does not need elaborate software to remain valuable. Its useful unit is still the same one the early Web was excellent at: a person with unusual experience writing down what happened and linking the photographs beside it.

Taylor says the website will never be finished. After nearly two decades, that may be the most old-Web sentence on it.

Read the voyages and sailing notes at The Simple Sailor

Posted on by

Ancient Web: Mister Ralls Left a GeoCities Self-Portrait Behind

Mike Ralls did not build a personal page by writing a résumé. He opened with Cuba, Thailand, Mount Fuji, cold-water endurance, free will, an African spider monkey named Gizmo, and a séance that supposedly summoned Dick Tracy. Then he admitted that this probably was not the clearest introduction.

Visit the preserved Mister Ralls page

That is immediately more useful than a modern profile card.

The page survives today through a GeoCities mirror, and some of its original images and counters are visibly wounded. References still point toward old geocities.com and Yahoo-era resources. Character encoding has taken a few bites out of the text. The layout is table-heavy. None of this makes the page less interesting. It makes the page more legible as an artifact of its time.

Ralls describes returning to Oregon after living in Japan, where he had worked as a teacher and what his business cards called a cultural ambassador for the Japanese government. He writes about the way living abroad forces the mind to accept that things do not have to work one familiar way in order to work at all.

At the time represented by the page, he was in graduate school pursuing an interdisciplinary master’s degree with history as the eventual goal. He had also started rock climbing, partly because ordinary gym exercise bored him. This is exactly the kind of context the old personal Web preserved casually and at scale: somebody’s interests, plans, jokes, intellectual obsessions, and current phase of life all occupying the same page.

The stated purpose of the site was mostly to post Ralls’s alternate history writing and whatever else he felt like sharing. The table of contents also linked to photographs and his Usenet presence. That last link is especially period-correct. A personal website and Usenet identity were not separate brands; they were different places where the same person existed online.

Modern social profiles encourage people to compress themselves into templates. GeoCities did almost the opposite. It gave you some server space and enough HTML rope to build whatever representation of yourself seemed reasonable that week.

This page survives because somebody mirrored it. That means we do not merely have a description of what a personal homepage looked like. We can still see the broken image references, the odd spacing, the anti-spam email trick, the grandiose heading “Mike Ralls – The Man and The Legend,” and the actual prose.

For Internet-history research, that messiness is the evidence.

Explore Mister Ralls’s surviving GeoCities page

Posted on by

Ancient Web: MadSci Has Been Answering Weird Science Questions Since 1995

Before search engines got good, one perfectly reasonable strategy for learning science was to ask an actual scientist and hope somebody answered.

MadSci built an entire website around that idea.

Visit the MadSci Network

MadSci says it debuted in September 1995 as part of Washington University’s Young Scientist Program, a student-run effort aimed at improving science literacy among K-12 students.

It evolved into a volunteer Ask-a-Scientist network covering subjects ranging from astronomy and biology to chemistry, computer science, engineering, earth science, and physics.

The site’s own history says that in 1999 more than 900 scientists volunteered and answered more than 25,000 questions. The current homepage advertises an archive of more than 36,000 answered questions and still offers the old core functions: browse the archive, search it, read FAQs, use the Random Knowledge Generator, or submit a question.

The examples are what make the project memorable.

Why do grapes spark in a microwave? Why does root beer foam with ice cream? Why does a black eye swell? Why might a frog urinate when somebody picks it up?

This is exactly what science outreach looks like when ordinary curiosity is allowed through the front door.

A human search engine for questions Google had not indexed yet

MadSci belongs to a specific era of the web when expert communities were experimenting with direct public access.

Instead of publishing only polished articles, the site preserved the questions themselves. That meant the archive gradually accumulated all the awkward wording, odd premises, school-project emergencies, and genuine curiosity that people actually bring to science.

That archive became more valuable with age.

A single answer might help one student. Tens of thousands of indexed answers become a searchable historical record of what people wanted to know, how experts explained it, and how online science communication worked before modern Q&A platforms and AI assistants.

The site also developed FAQs, web-resource libraries, experiments, research datasets, and educational tools around the question-answer core.

Most importantly, it kept the old structure visible.

The homepage still looks like a laboratory that escaped from 1997 and refused the redesign meeting.

CacheRat’s 1,967 Ancient Web Domains research list exists for pages like this: sites whose accumulated answers matter more than whatever visual fashion the current year says a homepage should follow.

MadSci’s copyright line now spans 1995–2026.

Thirty-one years later, the laboratory that never sleeps is apparently still combustible.

Browse the MadSci Network

Posted on by

Ancient Web: Land of Kain Built a Shadowrun Rulebook on the Web

Some old fan sites reviewed games. Others quietly became replacement manuals. Land of Kain’s Shadowrun section belongs to the second category: somebody kept rewriting, simplifying, comparing, and organizing the rules until the website itself became a working tabletop system.

Visit Land of Kain’s Shadowrun section

The deepest part of the site is a document called Shadowrun X, a heavily condensed personal ruleset built from ideas spanning several editions of Shadowrun. It is not a casual list of favorite house rules. The page runs through attributes, skills, qualities, cyberware, adept powers, spells, totems, drones, gear, armor, weapons, contacts, conversation, observation, infiltration, chases, combat, character creation, and design notes.

That last part matters. The author frequently explains why a rule was changed. Body and Strength are merged because they felt too interlinked. Perception is pulled out of the normal skill structure. Electronics and mechanics are combined for a more wireless world. Rules inherited from Shadowrun 1 through 5 are compared rather than merely copied.

The result is partly a game document and partly an archaeological record of somebody thinking through a complicated rules system over time.

The presentation is pure independent Web: long text, internal anchors, dense tables and lists, almost no attempt to turn the material into a glossy product. If you want to know how a particular subsystem works, you scroll or jump directly to it. There is no account wall, no video introduction, and no popup asking whether you would like to receive weekly cyberpunk insights in your inbox.

For tabletop researchers, that is useful. Commercial rulebooks show what publishers released. A site like this shows what players actually did with those rules afterward: what annoyed them, what they merged, what they threw away, and what they considered worth preserving.

It is also a good example of why personal reference sites age well. The design does not need to be fashionable because the value is in the accumulated thinking. A 1,600-line rules page still works perfectly well as a 1,600-line rules page.

CacheRat’s larger 1,967 Ancient Web Domains research list contains plenty of sites like this: not famous enough to dominate search results, but far too substantial to dismiss as random old pages.

If you have ever modified a tabletop system until the modifications became their own game, Land of Kain will look strangely familiar.

Explore the Shadowrun material at Land of Kain