Category: Ancient Web

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

Posted on by

Ancient Web: Ceolas Has Been Indexing Celtic Music Since 1994

Ceolas was indexing Celtic music before most musicians had websites to be indexed.

Visit Ceolas

The archive dates itself to 1994 and describes its purpose simply: collect and organize information about Celtic music on the Internet.

That grew into artist profiles, discographies, live-music guides, festivals, books, magazines, radio shows, instruments, reviews, sound samples, traditional tunes, dance material, and hundreds of links.

Before music discovery became a recommendation feed

Ceolas came out of an Internet where discovery moved through mailing lists, Usenet groups, FTP archives, hand-built web indexes, and people forwarding useful information to each other.

Its own history reflects that structure.

The project contributed to the IRTRAD-L mailing list and rec.music.celtic, maintained regional event calendars, operated material through celtic.stanford.edu, and published guides that were reused by Irish-American and Scottish-American newspapers.

A surviving September 1994 Internet-resources posting points readers toward the Ceolas archive for the current hypertext version of its Celtic-music resource list.

The tune section is particularly interesting because early music-sharing systems had to solve notation as well as discovery. Ceolas became an early home for ABC music notation, a compact text format for writing melodies that could travel through email and text-based systems long before browser audio was routine.

The site says that by November 1997 it was serving information to more than 40,000 users per month. Whether or not that number sounds large now, the important part is what those users were finding: specialist material that had previously been scattered across local scenes, printed listings, mailing lists, and personal knowledge.

CacheRat’s 1,967 Ancient Web Domains research list contains many pages from this indexing era.

Ceolas shows why they mattered.

Before search engines got good, somebody had to know where the music was.

Return to Ceolas

Posted on by

Ancient Web: AwayMessages.com Preserves the Miniature Literature of AIM

Before read receipts, typing indicators, and permanently connected phones, people sometimes had to announce that they had gone to take a shower.

Visit AwayMessages.com

AwayMessages.com is a surviving catalog of the little status notes people posted on AIM and other instant-messaging clients when they stepped away from the keyboard.

The homepage still sorts messages into categories such as bathroom, bored, food, friendship, funny, insults, partying, romance, school, shower, sleep, and inspirational.

That taxonomy is nearly a social history by itself.

A status message used to be an event

An AIM away message was part utility and part performance.

It told friends you were unavailable, but it could also be a joke, lyric, insult, inside reference, flirtation, or carefully selected fragment of teenage philosophy.

AwayMessages.com turned that behavior into a searchable library.

Its setup guide explains how to enable away messages in AIM, Yahoo Messenger, ICQ, and Google Talk. It also documents an AIM trick in which %n inside a message would be replaced with the screen name of the person reading it—a tiny personalization feature that produced years of prank messages.

The site expanded into buddy icons, message articles, status-message collections, and related social-web material. Its footer dates the project to 2001, placing it squarely inside the period when an instant-messenger client might sit open for hours on the family computer.

The interesting part now is not whether any particular joke is funny.

It is that thousands of tiny social scripts were considered worth collecting at all.

CacheRat’s 1,967 Ancient Web Domains research list includes plenty of technical history, but pages like this preserve behavior—how people signaled absence, personality, mood, and availability before modern social apps standardized those things.

An away message was only a sentence or two.

For a while, that sentence was your front door.

Return to AwayMessages.com

Posted on by

Ancient Web: Lifeprint Teaches ASL Handshapes the Way People Actually Use Them

An alphabet chart can show you a handshape. It cannot show you what happens when somebody uses that handshape several letters per second.

Visit Lifeprint’s ASL handshape reference

Lifeprint’s page on American Sign Language handshapes is useful precisely because it refuses to pretend that every letter has one frozen textbook form.

The page works through the alphabet, but it also explains palm orientation, classifier handshapes, regional variants, efficient movement, and how neighboring letters affect what a hand actually looks like during fingerspelling.

Real language keeps moving

The entry for the letter B notes that books often depict the thumb folded tightly across the palm, while everyday signing may use a more natural position.

The M, N, and T entries make the same point. Skilled signers do not pause to sculpt each letter for a diagram. Fingerspelling is movement, and the hand adapts to what comes before and after.

The page also distinguishes alphabet letters from handshapes used grammatically.

A flat B hand can describe flat surfaces or rectangular objects. A C hand can indicate round or cylindrical shapes. A bent V can represent crouching people or small animals. These are not merely letters recycled as gestures; they participate in the language’s system of classifiers and spatial description.

That makes the page valuable to beginners because it addresses a predictable failure mode: learning from static charts and then assuming natural variation is incorrect signing.

Lifeprint, created by ASL educator Bill Vicars, grew into a much larger instructional resource with lessons, videos, dictionaries, discussions, and archived student questions. The handshape page retains the conversational feel of an instructor answering exactly the questions learners get stuck on.

CacheRat’s 1,967 Ancient Web Domains research list includes many teaching pages whose usefulness comes from accumulated explanation rather than polished presentation.

This one teaches the gap between a diagram and a living language.

Return to Lifeprint’s handshape reference

Posted on by

Ancient Web: The Valve Museum Indexes 8,256 Vacuum Tubes and Keeps Counting

A vacuum tube with no readable markings is not garbage if somebody built an 8,000-entry museum to help identify it.

Visit The Valve Museum

Allan Wyatt’s Valve Museum is a sprawling reference to thermionic valves—the components Americans usually call vacuum tubes.

As of July 2026, the site reported 8,256 valve exhibits and was still being updated.

That number matters because tube history is mostly a history of variation.

The obscure part number is the point

The homepage leads into a classified index, a long all-valves list, a type finder for unmarked examples, technical articles, valve equivalents, references, vintage advertisements, original boxes, historical timelines, recollections, and specialist collections.

The site covers receiving valves, transmitting types, rectifiers, display tubes, specialized industrial devices, and generations of components made by firms such as Mullard, Brimar, BTH, Philips, and others.

For restorers, the useful layer is not just a photograph.

A valve may need to be identified by electrode structure, base, envelope, manufacturer markings, electrical characteristics, date, or known equivalent numbers. One obscure tube can decide whether an old radio, amplifier, instrument, or piece of test equipment can be understood at all.

The museum also preserves context around the parts. Old advertisements show how manufacturers explained new tube families. Packaging documents branding and replacement markets. Technical articles include classic amplifier material such as the Mullard 5-10. Recollections and timelines keep the components attached to the industries that used them.

The result is much closer to infrastructure than a gallery.

CacheRat’s 1,967 Ancient Web Domains research list repeatedly turns up these enormous specialist indexes: sites that look visually modest because nearly all the effort went into classification.

The Valve Museum does not need a cinematic interface.

It needs the right tube number to exist when somebody searches for it twenty years later.

Return to The Valve Museum

Posted on by

Ancient Web: Doctor Macro Scanned Classic Hollywood One Glossy at a Time

Before image search made every actor look infinitely available, finding a clean full-size studio photograph could be an actual research problem.

Visit Doctor Macro

Doctor Macro’s High Quality Movie Scans was built to solve that problem the slow way: acquire old 8-by-10 glossy stills and publicity photographs, scan them carefully, clean the damage, identify the people and films, and publish the results at useful resolution.

The site focuses heavily on movie stars from the 1940s and earlier, though the collection ranges beyond that period.

A scan archive with receipts

The homepage explains that preparing one image can take several hours.

The old news pages show what that means in practice. Updates identify contributors, films, actors, corrections, and restoration work. Scratches and physical damage are mentioned. Misidentified stills get corrected. Guest-submitted material lives in a separate Annex, while lower-quality items are preserved rather than simply discarded.

In May 2009 the site marked its sixth anniversary and published scan number 1,000.

The project also grew sideways.

Beyond actor galleries, Doctor Macro includes movie summaries, poster artwork, full-length radio shows, movie fan magazines, old movie books, songs, film-noir posters, and years of update logs that preserve the rhythm of the collection itself.

That makes it useful for more than wallpaper hunting.

Researchers can identify costumes, actors, publicity poses, production stills, minor performers, magazine layouts, and visual details that may never have received formal catalog records elsewhere.

There is also something worth noticing about the site’s stated design goals: high-quality images, frequent updates, easy navigation, accuracy, and no ads, pop-ups, banners, or cookies. That list reads like a small manifesto from a different phase of the web.

CacheRat’s 1,967 Ancient Web Domains research list contains many specialist collections whose real value is not rarity but careful human preparation.

Doctor Macro is one of them. Someone did the scanning work.

Return to Doctor Macro

Posted on by

Ancient Web: The Tale of Genji Maps a Thousand-Year-Old Novel Back Onto Japan

Most literary websites stop at the text. The Tale of Genji site keeps walking until the novel turns back into geography.

Visit The Tale of Genji

The site is devoted to Murasaki Shikibu’s eleventh-century work, but its structure is part reading guide and part field guide to the physical world behind the story.

It includes chapter summaries, principal characters, maps, background material, related books, pages about Murasaki Shikibu, and a large geographic section organized around Kyoto and other locations associated with the novel.

Literature with coordinates

The geographic index is where the site becomes unusual.

Instead of treating Heian-period place names as footnotes, it follows them outward into temples, shrines, rivers, cemeteries, neighborhoods, hills, and surviving landscapes.

Pages cover Kiyomizu-dera, the Kamo River, Daikaku-ji, Nonomiya Shrine, Lake Biwa, Ishiyama-dera, Mount Hiei, Uji, Sumiyoshi-taisha, Suma, Akashi, and dozens of other sites. Many entries include photographs so a reader can compare the literary setting with the modern place.

That makes the site useful in several directions at once.

A reader can use it to keep the 54-chapter novel straight. A traveler can build an itinerary around Genji-related places. A student can understand how court culture, religion, geography, and travel shaped the story. A historian can use the location pages as leads for deeper research.

The site also illustrates something the older web did particularly well: one person or small group could build a subject guide that ignored institutional boundaries. Literature, tourism, biography, architecture, religious sites, and photography all live together because separating them would make the subject less useful.

CacheRat’s 1,967 Ancient Web Domains research list repeatedly turns up pages like this—projects organized around curiosity rather than content categories.

A thousand-year-old novel becomes easier to enter when the web page can point at a river and say: this is the river.

Return to The Tale of Genji

Posted on by

Ancient Web: John Wolff’s Web Museum Preserves Music Machines That Play Paper

A century before streaming music, people were already buying songs encoded on rolls, cylinders, and perforated paper.

Visit John Wolff’s Mechanical Music museum

John Wolff’s pages document the machines that played those formats and the work required to keep both the instruments and their music alive.

His collection includes 14-note American Organettes from the late 1870s, 20-note Celestinas, Gem roller organs, a homemade crank organ, and the wonderfully strange Rolmonica: a pocket-sized 1920s instrument that automatically plays a harmonica from a paper roll.

Preserving the song means preserving the mechanism

Mechanical music is awkward archival material because the composition is inseparable from the machine that reads it.

A piano roll can survive physically and still become effectively mute if nobody understands the tracker bar, scale, hole spacing, roll speed, pneumatic system, or catalog numbering that made it work.

Wolff’s site attacks the problem from both sides.

The instrument pages describe mechanisms and restoration. The music section digitizes old rolls into MIDI, catalogs hundreds of Rolmonica titles, preserves 20-note organette music, and documents dealer labels and Australian Broadway pianola rolls.

One section covers his efforts to preserve original organette rolls and tune sheets dating from the 1880s through the 1930s. Another explains how to make replacement roll boxes. That sort of practical detail is what turns a collection into a working archive.

The site is also unusually candid about how hobbies expand. Wolff says he originally moved toward smaller mechanical instruments because children consumed the time, space, and money once available for reproducing pianos. Four decades later, the “smaller” hobby had grown into restoration, transcription, instrument building, arranging, and public performance.

The page itself dates to 2003 and was still being updated in 2026.

CacheRat’s 1,967 Ancient Web Domains research list is full of sites where the real asset is accumulated specialist labor. This is a textbook example.

The machines are antiques. The documentation is infrastructure.

Return to John Wolff’s Mechanical Music museum

Posted on by

Ancient Web: ZMatrix Turned the Matrix Code Rain Into a Windows Desktop

There was a point in computing history when making your Windows desktop look like The Matrix counted as a legitimate software project.

Visit ZMatrix

ZMatrix is a desktop enhancement that renders streams of characters behind ordinary desktop icons in the visual style popularized by the film. The project page is still sitting on SourceForge with its original screenshots, downloads, change log, bug reports, support links, and development notes.

The surviving page captures the software culture around Windows 98, Me, 2000, and XP unusually well.

Desktop customization used to require code

ZMatrix was not simply an animated GIF stretched over a wallpaper.

The program rendered text, blended it with existing backgrounds, tried not to interfere with desktop icons, included configuration controls, and eventually gained multi-monitor support, screensaver operation, Winamp visualization support, and special text streams.

The change log is the best part.

It documents taskbar-position bugs, disappearing icons, Active Desktop compatibility, installer privileges, password-protected screensavers, font spacing problems, Windows 9x support, and the sort of obscure rendering failures that only appear after strangers install your program on machines you have never seen.

In March 2002 the author released source code, initially under a BSD-style license and later under the GNU GPL. The page says a full build required Visual Studio 6.0 and Borland C++ Builder 6.0, which dates the project almost as efficiently as the screenshots do.

By February 2002, the author noted that the program had been linked from enough places to earn a USA Today “Pick of the Day.”

The site also preserves a miniature history of freeware distribution: dial-up-conscious installer sizes, SourceForge bug trackers, download portals, donation links, and irritation at commercial download sites charging developers for faster listing updates.

That is why pages like this belong in CacheRat’s 1,967 Ancient Web Domains research list. The software itself is interesting, but the surrounding page records how independent Windows utilities were actually made and distributed.

Return to ZMatrix

Posted on by

Ancient Web: Tilde.club Put Personal Home Pages Back on a Shared Unix Box

Tilde.club is what happens when somebody remembers that a web page can just be a file in your home directory.

Visit Tilde.club

The service gives users accounts on a shared Unix machine where they can log in, write files, publish from ~/public_html, use shell tools, send mail, and generally treat the Internet like a computer instead of a content platform.

Paul Ford started the project in late September 2014. His own surviving page describes it as something he “set up accidentally,” then watched fill with hundreds of people who wanted a place to make low-stakes personal pages again.

That accidental quality is important.

A server instead of a platform

Tilde.club does not begin with profiles, feeds, engagement metrics, or monetization.

It begins with a Unix account.

The public site lists member homepages directly, highlights recently updated pages, publishes SSH fingerprints, and still exposes the community as a collection of people with directories rather than accounts inside a social product.

The model is old. Universities, ISPs, shell providers, and early hosting services routinely gave users web space under URLs containing a tilde before hosted site builders made that structure look primitive.

Tilde.club deliberately brought the primitive structure back.

The project also became part of a larger “tildeverse” of shared Unix communities. Its code and administration documentation remain public, and the machine is still being maintained; by 2026 the service had completed an upgrade to Fedora 43.

That combination makes the site more than retro styling. It is a functioning demonstration of a different social architecture for the web.

CacheRat’s 1,967 Ancient Web Domains research list contains many pages built before the web became synonymous with centralized platforms. Tilde.club is interesting because it consciously recreated some of those conditions after they had already become unfashionable.

You get a shell, a directory, some disk space, and a URL.

Then you decide what the page is for.

Return to Tilde.club

Posted on by

Ancient Web: The University of Illinois Put Concrete Into a Web Course

Concrete is so ordinary that it takes a materials-science website to remind you how weird it actually is.

Visit the University of Illinois concrete module

The module was developed and revised during Materials Technology Workshops at the University of Illinois at Urbana-Champaign from 1993 through 1995. University professors, high-school science teachers and undergraduate students worked on it as a curriculum aid for teaching materials science and technology.

The resulting site is much more than a definition of concrete.

It contains historical timelines, scientific principles, laboratory activities, a glossary, equipment grids, quizzes, references, videos, and classroom exercises. Teachers were explicitly encouraged to use the whole module or lift individual experiments from it.

A construction material becomes a chemistry lesson

One section starts by asking students what concrete actually contains and why it hardens.

The answer leads directly into hydration: water reacts with cement so crystals form and interlock, binding the aggregate into a solid mass. From there the module asks why concrete is stronger in compression than tension, how setting can be accelerated or slowed, and what materials belong in the mix.

Then it gets gloriously specific.

Students are asked to brainstorm concrete applications ranging from roads, bridges, dams and sewer pipes to canoes, caskets, missile silos, nuclear-waste containment, furniture and even proposed lunar-base applications.

The module’s age is visible in some infrastructure statistics and projections, but that does not erase the educational design. The valuable part is the sequence: start with a familiar material, break it into components, explain the chemistry, test it in the lab, then connect the science back to things students physically encounter.

The project also preserves a particular moment in early educational publishing. This was classroom material translated onto the web before online courses became software platforms of their own.

CacheRat’s 1,967 Ancient Web Domains research list contains many university pages like this where the original teaching structure remains more interesting than a modern redesign would be.

The concrete module still teaches because the material science underneath it has not become obsolete.

A sidewalk remains chemistry with gravel in it.

Return to the concrete module