Category: Ancient Web

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

Posted on by

Ancient Web: Does Anyone Remember Websites? Asked the Right Question in 2017

In 2017 somebody asked a question that sounds increasingly less sarcastic every year:

Does anyone remember websites?

Read the original TTTThis page

The argument was simple. Search engines had become incredibly good at finding specific answers, but something got lost in the process: wandering into pages built by hobbyists, academics and technically inclined weirdos because they cared about something.

The page contrasts that older web with a commercial web where personalized sites are harder to surface among optimized pages built to rank, sell or convert.

It is not a claim that everything used to be better. Plenty of old sites were broken, ugly, abandoned or useless.

The point is narrower and better: searching is not the same thing as surfing.

Finding what you did not know to ask for

When the page hit Hacker News in November 2017, it drew hundreds of comments and nearly 900 points. The discussion immediately became a catalog of people remembering the exact experience the essay described: clicking through somebody’s handwritten pages and discovering useful material that no search query would have occurred to them to type.

That distinction still matters.

Modern search is optimized around intent. You ask for a thing and receive increasingly polished attempts at that thing.

The old personal web often worked sideways. You arrived looking for batteries and ended up reading a homemade test bench. You searched for one obscure technical term and discovered somebody’s entire research notebook. Links were not merely citations. They were trails.

That is why small directories, webrings, blogrolls and manually curated lists keep reappearing. They solve a different problem from search engines.

They help you discover something before you have language for it.

CacheRat’s 1,967 Ancient Web Domains research list is basically an industrialized version of that impulse: stop asking the web only for answers and start poking through its forgotten sheds again.

The original TTTThis page is flaky now, which is almost too appropriate.

The idea survived anyway.

Try the original page

Posted on by

Ancient Web: Philip Kraus Kept a Career in Operetta Online

Personal professional websites used to function like an expandable paper résumé: one page for the biography, one for photographs, one for contact information, and as many additional pages as the career required.

Philip Kraus’s site is a surviving example.

Visit Philip Kraus’s site

Kraus’s stage-director biography describes a career spanning opera, operetta, and musical theater. In 1980 he founded Light Opera Works, a professional company devoted to operetta classics, and served as artistic director for 19 seasons, directing more than 50 productions.

The list includes revivals and Chicago premieres of works such as Orpheus in the Underworld, Naughty Marietta, The Gypsy Baron, The Grand Duke, The Chocolate Soldier, The Duchess of Chicago, and several Kurt Weill stage works.

The site also documents his long specialization in Gilbert and Sullivan. Kraus trained at Northwestern University in music and theater, performed G&S roles, and directed the repertoire repeatedly. His 1986 Elizabethan concept for The Mikado was later revived elsewhere.

He also directed serious opera, worked with Pamiro Opera, taught and directed university opera programs, and produced English performing translations of European works.

This is not a famous institutional archive.

It is one person’s website, and that is exactly why it is useful.

A career without a feed

Modern performing-arts careers are scattered across ticket sites, social networks, press releases, dead theater pages, streaming clips, and search results that gradually reorder themselves.

A static personal site does something different: it freezes the person’s own structure.

Kraus decided which productions mattered enough to list, which roles and institutions belonged together, which photographs should illustrate the work, and which translations deserved mention. The result is not neutral scholarship, but it is a primary-source record of how a working director presented decades of professional history.

That distinction matters.

Researchers often preserve the famous production but lose the connective tissue: who revived an obscure operetta, who translated it, where it was staged, and which regional companies carried that repertoire forward.

Old personal sites often contain exactly those details.

CacheRat’s 1,967 Ancient Web Domains research list exists for pages like this because cultural history does not live only in major archives. Sometimes it survives in a hand-built biography page that nobody thought to redesign.

The page is plain.

The career is not.

Browse Philip Kraus’s site

Posted on by

Ancient Web: Georg Potthast Kept DOS Machines on the Network

Getting DOS onto a modern-ish network is the kind of problem that sounds simple until you are staring at an unidentified PCI network card and an operating system that predates most of the hardware in the machine. Georg Potthast’s SIOUX pages exist for exactly that unpleasant afternoon.

Visit Georg Potthast’s SIOUX pages

The site became known in DOS and FreeDOS circles for packet drivers: small pieces of software that let DOS networking programs talk to Ethernet hardware. Potthast collected drivers for PCI network adapters and provided tools for identifying what chipset a card actually used.

One of the important utilities is NICSCAN.EXE. FreeDOS documentation specifically points users to Potthast’s site for NICSCAN and explains the practical reasoning behind it: different retail network-card models often use the same underlying chipset, so hunting for a driver by the exact name printed on the box can be a waste of time. Identify the chipset first, then find the appropriate packet driver.

That is old-computer troubleshooting in one sentence.

The FreeDOS networking guides still reference SIOUX when explaining how to bring up Ethernet. A typical configuration loads a packet driver from AUTOEXEC.BAT, often using software interrupt 0x60, and then hands networking off to software such as mTCP, WatTCP-based applications, FTP clients, DHCP tools, or browsers.

Potthast’s pages are especially interesting because they sit at the boundary between two eras. DOS networking started in a world of ISA cards, dial-up connections, Novell stacks, and carefully assigned interrupts. The SIOUX material helped extend that world into PCI Ethernet and USB hardware that arrived much later.

In other words, the site is not merely preserving obsolete drivers. It documents the work required to make an obsolete operating system cooperate with hardware it was never designed to understand.

That has practical value in 2026. Retrocomputing increasingly involves real machines rather than emulators alone. Moving files, reaching an FTP server, connecting an old PC to a local network, or simply proving that a 486 can still speak TCP/IP often depends on exactly this kind of forgotten middleware.

Modern vendor support pages are usually useless here. The card manufacturer moved on twenty years ago. The operating system vendor is gone or irrelevant. The useful answer survives because one person collected the ugly little .COM files and wrote down how to use them.

That is the old Web doing technical support after everyone with a budget stopped caring.

Explore the SIOUX DOS networking archive

Posted on by

Ancient Web: The Fred Project Built a Ventriloquist Dummy in Public

The Fred Project starts with a problem ventriloquists have had for a very long time: buy the same stock figure everybody else can buy, spend serious money on a custom maker, or build the thing yourself and accept responsibility for whatever comes out of the workshop.

Visit The Fred Project

Al Stevens chose the third option and documented the process.

The project describes building a professional ventriloquist figure using head and hand casting kits from figure maker Mike Brose and Puppets and Props. The figure was named Fred by Brose’s wife. From there the site walks through the kind of work a finished dummy normally hides: casting parts, materials, the head shell, jaw mechanics, eyes, hands, finishing, assembly, and the endless small decisions required to turn inert parts into a character that can sit on someone’s knee and appear disturbingly alive.

That level of documentation is why the page kept circulating. Puppet builders were still pointing people toward it in the mid-2000s, hobbyists rebuilding figures cited it years later, and even modern 3D-printing sellers continue recommending the Fred Project as a guide to the mechanics inside a ventriloquist figure.

The page is especially valuable because it documents a physical craft rather than merely talking about it. A lot of old Web material describes software or culture that can be copied perfectly. Fred exists at the opposite end of the spectrum. Brass rod has to fit. A jaw has to pivot. Eyes have to sit correctly. Paint, glue, filler, fabric, and hand tools eventually have to behave in the real world.

The Web was extraordinarily good at this kind of knowledge transfer. One person built something difficult, photographed the intermediate states, wrote down what worked, and left the pages online. You did not need a publisher to decide there was a large enough market for a book titled Here Is Exactly How I Built This Particular Ventriloquist Dummy.

The niche itself did the publishing.

For modern makers, the Fred Project also shows why step-by-step pages age better than short-form video. You can stop at the exact stage you need, compare the photograph with the object on your bench, reread the materials list, and continue without scrubbing through twelve minutes of introductory chatter.

There is a lot of Internet history hidden in pages like this: not famous websites, but durable instructions that quietly became part of a craft’s shared memory.

Read Al Stevens’s Fred Project

Posted on by

Ancient Web: WhipNet Archived Real Men of Genius Before Streaming Platforms

There was a time when hearing an ad you liked again required somebody on the internet to save the audio for you.

WhipNet volunteered.

Visit WhipNet

Its surviving Bud Light section is a fan-built archive of the radio campaign first called Real American Heroes and later Real Men of Genius.

The page is aggressively early-2000s: black backgrounds, giant type, tiny graphics, tables, MP3 links, lyric pages and a completely unnecessary amount of visual enthusiasm.

Good.

The archive links individual spots such as Mr. Pro Wrestling Wardrobe Designer, Mr. Jelly Donut Filler, Mr. Wrecking Ball Operator and Mr. Underwear Inspector Number Twelve. It also preserves lyrics and a copy of a November 12, 2003 St. Louis Post-Dispatch article about the campaign moving from radio into television.

That article notes that the campaign had debuted about four years earlier and that its original “Real American Heroes” name was changed after the September 11 attacks.

A fan page became the replay button

Before YouTube, Spotify and official brand archives made media instantly replayable, fan pages often did the practical preservation work.

Somebody captured the file, named it, uploaded it, wrote down the lyrics and built links around it.

WhipNet even includes the necessary disclaimer that it was not affiliated with Bud Light or Anheuser-Busch. The page says it was assembled from information gathered from other pages and asks visitors to send additional commercials the webmaster had not found.

That is classic distributed-web preservation: incomplete, unofficial, enthusiastic and surprisingly durable.

The historical value is not just the ads themselves. The page preserves how people collected media before everything lived inside a few giant services. MP3 files sat directly on web servers. Fans traded missing examples. Search engines led you to a weird table full of blue links and you hoped somebody had paid the hosting bill.

CacheRat’s 1,967 Ancient Web Domains research list exists because pages like this routinely outlive the media ecosystems they were built around.

An advertising campaign became pop culture.

A fan with web space turned it into an archive.

Dig through WhipNet

Posted on by

Ancient Web: Splitting Images Is a Static-HTML Celebrity Double Agency

Somewhere between a casting directory and a fever dream sits one of the web’s more durable business models: hire a person who looks vaguely like somebody famous.

Splitting Images built a website for exactly that.

Visit Splitting Images

The UK agency describes itself as a provider of celebrity lookalikes, famous doubles, and professional impersonators for commercials, corporate events, promotions, parties, television, and other appearances. Its homepage lists clients including McDonald’s, Virgin, Marks & Spencer, BBC, ITV, and MTV.

The site’s real archaeological value is in the individual pages.

They are wonderfully direct. A celebrity name, one or more photographs, the impersonator’s name, and the agency contact information. George Clooney. Patrick Stewart. Wayne Rooney. Pamela Anderson. Peter Andre. Eminem. The structure barely changes because it does not need to.

This is a database built out of ordinary HTML files.

Before every business became a platform

Today a talent agency might hide its roster behind JavaScript filters, embedded booking software, analytics tags, social feeds, and a customer relationship system that requires seventeen cookies before showing a face.

Splitting Images just made pages.

That simplicity has an unintended archival advantage. Search engines can still index individual performers years later. Old discussion forums from 2006 and 2007 link directly to specific lookalike pages while users argue over which doubles are convincing and which are catastrophically funny.

Those conversations are part of the artifact too.

They show the site operating as an actual piece of mid-2000s web culture, not merely a modern page designed to imitate one. A link could point to one exact celebrity double. People passed it around. The joke, the business, and the URL all traveled together.

There is also something refreshingly literal about the site’s architecture. A lookalike agency needs pictures of lookalikes. So the website is mostly pictures of lookalikes.

No thought leadership section has been detected.

CacheRat’s 1,967 Ancient Web Domains research list exists to surface pages like this because ordinary businesses are part of web history too. The internet was not built only by universities, hackers, and famous startups. It was also built by small agencies figuring out how to put their catalog online.

Splitting Images did that and, improbably, the catalog still answers the phone.

Browse Splitting Images

Posted on by

Ancient Web: TouringCarRacing.net Became a Citation Source for Racing History

Some fan sites become references because somebody keeps collecting the boring details everybody else skips. TouringCarRacing.net became one of those sites: a huge specialist database covering touring-car racing history from the 1950s into the early 1990s.

Visit TouringCarRacing.net

The site is associated with Dutch racing historian Frank de Jong, whose work is repeatedly cited by later motorsport databases, Wikipedia articles, sim-racing projects, and racing-history researchers. Sources using the site describe it as a massive database of touring-car events and results from roughly 1952 through 1993.

Its value is not one famous article. It is accumulation.

Championship tables, individual rounds, car specifications, driver combinations, lap times, photographs, and results were gathered across series that are much harder to reconstruct than Formula One. European Touring Car Championship seasons, British saloon and touring-car racing, Australian events, national championships, Group A years, and obscure one-off races all leave traces in the archive.

That depth is why the site escaped the usual fate of hobby pages. Researchers cite individual pages as sources. Sim-racing modders use its photographs and specifications to recreate historic cars and grids. Race-history pages point back to de Jong for championship tables and event results that are otherwise scattered across period magazines, programs, private collections, and memory.

The story also has a more human ending. Frank de Jong died in 2018. Later racing communities explicitly dedicated work to his memory and continued referring to TouringCarRacing.net as one of their best historical resources. Other sites now describe the database as being maintained as his legacy.

That changes what the domain represents. It is no longer simply a fan site that stayed online. It is the residue of years of one person’s specialist research, now woven into other people’s work.

This is exactly why little independent reference sites matter. A commercial publication may cover the famous championship or winning driver. A dedicated amateur historian is more likely to record who finished seventh at Zolder, what car they drove, who shared it, and whether the lap time agrees with a program from 1984.

Search engines are good at finding famous facts. They are much less useful when the question contains a year, an obscure race, a chassis, and three names nobody has searched this month.

TouringCarRacing.net exists for that question.

Explore the surviving TouringCarRacing.net database

Posted on by

Ancient Web: 2D53 Preserved North Wales Railways in Photos and Paperwork

Railway history gets dramatically more useful when somebody saves the paperwork too.

Visit 2D53

Dave Plimmer’s site documents the main-line railways of North Wales, especially the period when British Rail still looked, sounded and operated like the railway many enthusiasts actually remember.

The core photographs were mostly taken between 1975 and 1983. Around them, the site preserves timetables, timetable graphs, signal-box locations, track and signalling information, headcodes, route availability material, trip diagrams and extracts from official documents issued between 1960 and 1992.

The geographic coverage is broad: Chester, Rhyl, Colwyn Bay, Llandudno, Bangor, Holyhead, Wrexham, Shrewsbury, the Cambrian Main Line and Cambrian Coast among others.

This is not merely a gallery of locomotives.

It is operational context.

The boring documents are the treasure

A photograph tells you what was there.

A working timetable, headcode explanation or signalling diagram starts telling you how it worked.

That distinction is why sites like 2D53 age so well. Modern social media is perfectly capable of circulating a beautiful train photo. It is much worse at preserving the obscure document that explains why the train was carrying a particular code, which route it could use, or how the service fit around everything else on the line.

The site was still being updated in 2025, adding signal-box closure material and more railtour pages. That is another old-Web trait worth noticing: a website does not have to be frozen to be historical. It can simply keep accumulating carefully organized evidence for decades.

Other railway reference pages still point readers toward 2D53 for headcodes and timetable examples. That is what happens when a personal specialist site becomes infrastructure without ever needing to call itself a platform.

CacheRat’s 1,967 Ancient Web Domains research list exists for this exact species of website: enthusiast knowledge that becomes more valuable every year because the original systems, documents and firsthand memories keep getting farther away.

The web is very good at remembering famous trains.

2D53 remembers how the railway actually ran.

Explore the North Wales railway archive

Posted on by

Ancient Web: rcgldr Is What a Personal Homepage Used to Be

A personal website used to mean exactly that: a website containing the person’s stuff.

No brand architecture required.

Visit rcgldr

The homepage explains the name immediately: rcgldr is Jeff A. Reid’s unique abbreviation for “radio control glider.” That handle became his long-running online identity because, as his older companion site explains, there are plenty of Jeff Reids on the internet but only one rcgldr.

What follows is the kind of category collision that modern site builders are trained to eliminate.

Radio-control videos sit next to local car-show photos. Gymnastics videos and photos share space with racing-game clips, savefiles, mods, Hayabusa pictures, Caterham pictures, Corvette Z06 galleries, and a downloadable Need for Speed: Shift setup guide and savegame.

The older JeffReid.net page goes even farther, linking programming material on polynomial curve fitting, merge sorts, Windows multithreading, finite-field code, and other technical work.

There is no attempt to pretend these subjects belong to one market segment.

They belong to one person.

The homepage as a filesystem with opinions

That design philosophy is worth remembering.

Social networks split identity into posts. Corporate websites split information into funnels. Search engines surface isolated pages with little context. The old personal homepage often did the opposite: it let one human accumulate material over years without demanding that every addition justify the site’s “purpose.”

The result can look disorderly, but it preserves relationships that would otherwise disappear.

Reid’s radio-control identity leads into his technical interests. Racing games lead into savefiles and configuration material. Car enthusiasm leads into galleries. Old video formats remain alongside newer GitHub work. The site is not a frozen museum; it is a long-lived workbench.

That is also why tiny pages like this matter to web archaeology. Their value is not necessarily fame. Their value is continuity.

A 2015 Genesis Owners forum post, for example, links directly to an rcgldr-hosted WMV test video while discussing video playback in a Hyundai Genesis. The explanation includes encoding settings and then casually notes that the clip shows Reid’s radio-control glider flying above a canyon.

One URL connects hobby aviation, car electronics, video encoding, and forum troubleshooting.

CacheRat’s 1,967 Ancient Web Domains research list exists for precisely this sort of site: pages too individual to fit neatly into a modern content category, but too useful and specific to dismiss.

The web was better when a homepage could simply mean, “Here is my stuff.”

Browse rcgldr

Posted on by

Ancient Web: Mercenary’s Helper Thought the Other Star Fox 64 Sites Were Crap

The creator of Mercenary’s Helper looked around at the existing Star Fox 64 pages in 1997, concluded that most of them were “crap,” and decided to build a better one. That is about as pure a mission statement as the old fan Web ever produced.

Visit Mercenary’s Helper

The site says it was born August 19, 1997, only weeks after the North American release of Star Fox 64. It reads like a fan site built while the game was still new enough that strategies, secrets, high scores, voice clips, and community knowledge were being assembled in real time.

There are guides to fighting Star Wolf, level material, multiplayer information, miscellaneous help, fan fiction, polling results, and a high-score FAQ. The FAQ is not merely a list of cheats. It explains techniques for squeezing more points out of stages, exactly the sort of knowledge that once moved from player to player through FAQs, message boards, and personal sites before video walkthroughs became the default format.

The update log is part of the history. In December 1998, the webmaster finally fixed the fan-fiction button so it would stop producing a 404. On Christmas Day he uploaded 368 voice clips and improved the quality of many of them. In March 1999 he noted that a host failure had broken access to sounds and other files, so he changed the site to make them reachable again.

That kind of note is mundane until you realize what it records: the maintenance problems of a late-1990s fan site as they happened. Free hosting disappears. Links break. A teenager has a busy senior year. Somebody sends in an FAQ. Audio files get re-encoded. The site grows sideways because people contribute things.

The page still shows its period construction openly. Image-based navigation, tiny graphics, old advertising remnants, and hand-written update text remain in place. More importantly, the content still loads. A page created when the Nintendo 64 was current hardware can still tell you how to fight Star Wolf without asking you to accept cookies first.

Modern game databases are broader and more polished, but they rarely preserve the moment when a community was still discovering the game. Mercenary’s Helper does. Its tone, priorities, missing pages, corrections, and accumulated files are all part of the record.

The creator wanted to do better than the other Star Fox 64 sites. Nearly thirty years later, the fact that his version is still readable gives him a fairly strong closing argument.

Dock your Arwing at Mercenary’s Helper