Category: Ancient Web

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

Posted on by

Ancient Web: TEXTFILES.COM Kept the BBS World Readable

Before screenshots, social feeds and video ate most online conversation, an enormous amount of computer culture traveled as plain text. TEXTFILES.COM exists because Jason Scott kept copies.

Browse the TEXTFILES.COM directory

Scott started the site in 1998 with a specific mission: make available the BBS text files he had collected during his youth. These were the files that moved through dial-up bulletin board systems — stories, FAQs, technical notes, jokes, phreaking lore, game material, rants, newsletters and the miscellaneous debris people create whenever storage gets cheap enough to waste creatively.

The project did not stay small.

Scott later described the broader TEXTFILES.COM family as holding more than a terabyte of computer history. Separate collections expanded into CD-ROMs, shareware and other material that did not fit neatly into the original BBS archive.

Copy the damn thing

One of the best parts of the project is its attitude toward preservation.

Scott has repeatedly encouraged mirroring and bulk copying. In a 2003 essay about archiving, he wrote about finding entire copies of TEXTFILES.COM circulating on peer-to-peer networks and being delighted rather than offended. The important part was that the files had escaped dependence on one server.

That idea aged extremely well.

A preservation project that can only survive while one domain, one company and one administrator remain healthy is not very preserved.

The site was deliberately structured so people could take copies without being married to its HTML presentation. Directory description files made the collection easier to move and rebuild elsewhere.

For webmasters, there is a serious lesson hiding underneath the mountain of BBS weirdness: exportability is a feature. Simple files, descriptive metadata and stable paths survive organizational changes better than content trapped inside an application nobody can run ten years later.

A collection became history

When TEXTFILES.COM began, much of the material was recent enough to be remembered as junk from dial-up systems. Keep junk long enough and suddenly researchers want it.

Scott went on to make BBS: The Documentary, work on other computer-history projects and become an archivist at the Internet Archive. The old text-file collection sits near the beginning of that path.

It is also one of the 1,967 sites in CacheRat’s Ancient Web research list — which is admittedly a very on-brand place for it to end up.

Explore the full Ancient Web Domains research list

TEXTFILES.COM is still useful for the same reason an old shoebox of disks can be useful: nobody knew which apparently stupid file would become interesting later.

Go dig through TEXTFILES.COM

Posted on by

Ancient Web: Somebody Built a Giant 6502

The MOS 6502 is famous for fitting into some of the most important home computers and game systems ever made. The MOnSter 6502 project decided the obvious response was to make it enormous again.

Visit the MOnSter 6502 project

The site calls it a “dis-integrated circuit”: a complete working transistor-scale replica of the classic processor built from discrete components instead of hidden inside a tiny silicon package.

The board measures about 12 by 15 inches and contains 4,769 components. The transistors and LEDs make the internal activity visible in a way the original chip never could.

That is both ridiculous and genuinely useful.

A processor you can look at

The original 6502 family sat at the center of machines including the Apple II, Commodore PET, Atari 400 and 800, BBC Micro and many other systems. Related versions appeared in the Commodore 64, Atari 2600 and Nintendo Entertainment System.

Normally, all of the interesting switching happens invisibly inside the package. You can study a schematic or look at a microscope photograph of the die, but the processor remains abstract.

MOnSter 6502 takes the opposite approach. Its oversized implementation exposes the logic as physical parts spread across a board. LEDs make signals and state changes visible. The CPU becomes something closer to a working wall diagram.

The project site includes technical explanations, photographs, video and an FAQ that starts with the reasonable question: “Sixty-Five-Oh-What?”

This is what hobby websites are for

No normal product roadmap leads to a giant discrete 6502. It exists because somebody thought the idea was worth building.

That is one of the things the older independent Web was especially good at documenting. A person builds an unnecessarily large processor, creates a site explaining what it is, and now years later anybody interested in CPU design can wander in and see the result.

It contributes education, preservation, hardware art and a very strong answer to the question “but can you make the chip bigger?”

MOnSter 6502 is one of 1,967 sites in CacheRat’s larger Ancient Web research list.

Explore the 1,967 Ancient Web Domains research list

The integrated circuit won because making every transistor this large is a terrible manufacturing strategy. As a way to understand one, though, it is excellent.

Explore the MOnSter 6502

Posted on by

Ancient Web: U-boat Archive Is Still Growing

U-boat Archive is not a frozen website somebody forgot to unplug. In September 2026 it was still adding German submarine patrol records from the Second World War.

Visit U-boat Archive

The site’s stated purpose is straightforward: present documents and photographs from the battle between German U-boats and Allied anti-submarine forces during World War II.

The result is a specialist archive where individual boats, patrols and records matter more than broad summaries of the Battle of the Atlantic.

Patrols instead of bullet points

A major part of the collection is built around German KTB material — war diaries and patrol records. The site’s 2026 additions show the level of detail involved.

Records were added for boats operating in the Atlantic, Indian Ocean and Black Sea. One update covers U-435 recovering the crew of the manned weather station Knospe from Spitzbergen. Another follows U-354 picking up a weather troop from Hopen Island. The September additions included material from U-172’s third patrol during operations off Cape Town.

Those details are too small for most general histories and exactly what a researcher may be looking for.

The archive also preserves photographs and other wartime documentation related to both the submarines and the Allied forces trying to find them.

This is why specialist websites matter. A broad encyclopedia can tell you what a U-boat campaign was. A focused archive can help answer what a specific boat was doing on a specific patrol and what surviving paperwork says about it.

Still adding material

There is something satisfying about a site dedicated to an eighty-year-old conflict having a “Latest Additions” section that actually contains entries from this year.

Historical collections are never really finished. Documents are scanned, identified, translated, compared and connected to other records. A website gives one dedicated researcher or small group somewhere to keep doing that work without waiting for a publisher to decide the next detail is commercially important.

U-boat Archive appears in CacheRat’s larger collection of 1,967 Ancient Web domains for researchers and webmasters who want the full pile rather than one excavation at a time.

Explore the 1,967 Ancient Web Domains research list

If naval history is your thing, do not plan on opening this site for five minutes.

Dig through U-boat Archive

Posted on by

Ancient Web: The NES Color Problem

If you have ever argued that an NES emulator’s colors look wrong, congratulations: you have stepped into a problem large enough for somebody to spend years measuring, capturing, comparing and rebuilding palettes for it.

Visit FirebrandX’s NES Composite Palette Project

The annoying part is that there is no single tidy RGB palette hiding inside the original NTSC Nintendo Entertainment System. The NES PPU produces composite video rather than a native RGB color space, and the final picture depends on how a television decodes that signal.

Different CRTs can therefore make the same console look different. Once emulators, digital displays and RGB modifications enter the picture, the simple question “what are the real NES colors?” turns into a swamp.

FirebrandX walked into the swamp on purpose.

From eyeballing to captures

The project describes several generations of work: early palettes adjusted by eye, direct captures from composite output, hybrid approaches and reverse-engineered palettes.

One palette, “Composite Direct,” preserves an unmodified capture of the front-loading NTSC NES output. Another reproduces colors from the NES Classic. The “PVM Style D93” palette was built by comparing a Sony PVM CRT against a digital display and adjusting colors to make the two behave as similarly as possible.

FirebrandX says more than 30 hours went into that PVM palette alone, including sampling games across the NES library and repeatedly comparing synchronized game output on the CRT and LCD.

That is the kind of sentence that separates preservation work from “looks close enough to me.”

The page also includes downloadable palette files, firmware packages for NESRGB boards and practical notes about flashing hardware. It is research that can actually be used.

Why a page like this matters

Old game hardware is full of details that disappear when preservation stops at ROM files. Video output, display behavior, controllers, timing, audio circuitry and even the ugly characteristics of consumer televisions were part of the experience.

This page preserves one argument in unusual depth: what color was the NES actually showing us?

For emulator developers and retro-hardware people, the answer is useful. For everybody else, it is a nice example of how one apparently tiny compatibility problem can contain years of engineering work.

The NES palette project is part of the broader 1,967-site Ancient Web research list CacheRat is working through.

Explore the Ancient Web Domains research list

If nothing else, the page will permanently cure you of the idea that there is one obvious hex code for “Nintendo blue.”

Explore the NES Composite Palette Project

Posted on by

Ancient Web: Early Radio History Has Been Online Since 1996

Thomas H. White started putting early radio history on the Web on September 30, 1996. Thirty years later, the site is still doing the same job and its original articles were still receiving updates in 2026.

Visit United States Early Radio History

The site’s main focus is the United States from roughly 1897 through 1927, when “wireless” went from experimental signaling to broadcasting, amateur radio, commercial networks and government regulation.

That is already a useful niche. White went much deeper than a normal historical overview.

The strange details survived too

The site is organized into more than twenty major sections covering the electric telegraph, wireless telegraphy, vacuum tubes, early broadcasting, amateur operators, government regulation, radio networks and other pieces of the transition.

There is a section on the origins of words including radio, broadcast and ham. Another digs into United States callsign policies and the fascination with three-letter station calls. There are pages on pioneer broadcast stations, early telephone-based entertainment systems and the awkward process of figuring out how to regulate a technology that had spread faster than the rules around it.

And then there is section 21: “Fakes, Frauds, and Cranks.”

Every history archive should be honest enough to reserve shelf space for the cranks.

A large part of the value comes from period material. The site pulls together articles and extracts that let the old arguments, terminology and technical assumptions survive instead of flattening everything into a modern summary.

A long-running specialist site

White notes that the project originally lived on an ISP personal-page address before moving to its current domain in 2003. That little detail is part of Ancient Web history by itself. Many useful sites began as directories hanging off somebody else’s server and survived only because their owners cared enough to move them when the infrastructure changed.

For researchers, this is a deep reference. For webmasters, it is a case study in what happens when a narrow subject is documented consistently for decades: the site becomes difficult to replace.

United States Early Radio History is one of the 1,967 domains in CacheRat’s larger Ancient Web research list.

Browse the complete Ancient Web Domains research list

If you arrive looking for one callsign or one old radio term, there is a decent chance you will leave an hour later knowing why somebody in 1924 thought Mars might be listening.

Dig into United States Early Radio History

Posted on by

Ancient Web: The IBM 1401 Restoration Site

Some old-computer sites preserve photographs. The IBM 1401 Demo Lab and Restoration Project site preserves enough material to make you wonder whether the mainframe is slowly being backed up into HTML.

Visit the IBM 1401 restoration site

The project centers on two operational IBM 1401 systems at the Computer History Museum: one built in 1964 and brought from Hamm, Germany, and another built in 1961 from Darien, Connecticut.

The IBM 1401 was introduced in 1959 and became one of the defining business computers of the 1960s. IBM says that by the middle of the decade the 1401 family accounted for more than half of the world’s computers. It helped move businesses away from rooms full of specialized electromechanical card equipment toward stored-program computing.

That history is impressive. The website is impressive for a different reason: it shows the maintenance work.

Documentation all the way down

The site contains restoration status reports, component notes, manuals, Automated Logic Diagrams, maintenance procedures, photographs, software, bug logs, oral-history material, scanned IBM documents, card-reader information, printer work, tape-drive work and stories from the people who knew the machines when they were ordinary equipment instead of museum pieces.

A 2007 status report recorded roughly 9,000 volunteer hours by that point. The restoration team included retired IBM customer engineers and other volunteers who could still recognize what a half-century-old machine was trying to tell them when it stopped behaving.

The archive kept growing after the initial restoration. Its update pages show new manuals, technical notes and historical material appearing years later.

This is the part of preservation that gets missed when people think archiving means taking a photograph and putting an object behind glass. A working computer is a system of procedures, maintenance knowledge, software, peripherals, sounds, habits and people.

The website looks like the workbench

The main page is enormous. It is dense with links. Some sections have been added across years by different hands. There are smileys, odd spacing, old-school navigation and more technical material than any modern redesign consultant would allow onto one page without staging an intervention.

Good.

For researchers, that density is the value. For webmasters, it demonstrates something worth stealing: the site’s priority is keeping the knowledge reachable. It is not afraid to be a reference work.

This is one of the more substantial technical sites in CacheRat’s 1,967-domain Ancient Web research list.

Explore the full Ancient Web Domains research list

If you want the difference between preserving a machine and preserving the knowledge required to keep one alive, spend some time in this site.

Explore the IBM 1401 Demo Lab and Restoration Project

Posted on by

Ancient Web: The First Website Still Explains the Web

Before the Web became the place where a refrigerator can demand an account and a privacy policy, its first page had a simpler job: explain what the Web was and point people toward the software.

Visit the historic World Wide Web project page at CERN

CERN identifies info.cern.ch as the address of the world’s first website and web server. The first page address was /hypertext/WWW/TheProject.html, devoted to Tim Berners-Lee’s WorldWideWeb project.

The exact earliest page changed as the project developed, so what survives at the historic address should not be treated like a frozen byte-for-byte snapshot from one magical afternoon in 1990. What matters is that CERN has preserved the project and the address where the Web first explained itself.

A homepage before homepages had rules

The page is almost aggressively small.

It defines WorldWideWeb as a wide-area hypermedia information-retrieval initiative and then starts linking outward: project summaries, mailing lists, policy, software, technical documentation, people involved, project history, code and information about what was already available online.

There is even a section called “How can I help?”

Another link tells visitors how to get the code by anonymous FTP.

In 2026, that reads like archaeology. At the time, it was onboarding documentation.

By Christmas 1990, Berners-Lee had the first browser/editor and server running at CERN on a NeXT computer. CERN’s own history notes that the machine carried a handwritten warning telling people not to power it down because it was a server. The Web was global in ambition and still physically vulnerable to somebody reaching for the wrong switch.

The useful lesson for webmasters

The first website is not impressive because of its visual design. There essentially isn’t any.

It is useful because the information architecture is obvious. The page tells you what the project is, what exists, where to get software, how to learn more and who is involved. The links are the interface.

That is worth remembering after three decades of carousels, overlays, frameworks, cookie banners and buttons that animate because somebody had a quarterly objective.

CacheRat’s Ancient Web list contains 1,967 sites from later generations of that same medium: personal pages, hobby sites, technical archives and other things built before the Web settled into its current shape.

Explore the 1,967 Ancient Web Domains research list

The first page is still short enough to read in a minute. The consequences took a little longer.

Open the World Wide Web project page

Posted on by

Ancient Web: Multiplayer SimCity on X11

There was a multiplayer version of SimCity for X11 in the early 1990s. It ran on Unix workstations, could put several people into the same city over a network, and its surviving web page still has the product announcement to prove it.

Visit Don Hopkins’ SimCity archive

The page is maintained by Don Hopkins, who worked on Unix and multiplayer versions of SimCity. What survived is not a retrospective assembled thirty years later. It is a pile of primary material: screenshots, manuals, reviews, conference material, demo transcripts, interface discussions and the original announcement for Multi Player SimCity on X11.

That announcement is dated 1993 and reads like software distribution from another planet. A fully functional demo could be copied and played without a license, but there was a catch: the city melted every five minutes.

That is copy protection with personality.

SimCity escaped the beige box

The archive has screenshots of SimCity running on SGI Indigo hardware, Sun systems and NCD X terminals. There are transcripts from demos of both X11 SimCity and a HyperLook edition for the NeWS window system.

There is also a proposal for SimCityNet and material on user-interface design, including a summary of a Will Wright talk about interfaces for simulation games.

One line deserves to survive on its own: multiplayer SimCity went on tour with the Electric Carnival at Lollapalooza.

Today, networked games are ordinary enough that nobody thinks twice about the plumbing. These pages come from a period when putting a familiar simulation across networked Unix workstations was itself interesting enough to demonstrate at conferences and music festivals.

Why the page matters now

For game developers, this is source material. It shows not only what existed, but how it was described to users and other developers at the time. Product announcements list licensing models. Screenshots show actual interfaces. Transcripts preserve the way the software was demonstrated before streaming video became the default historical record.

For webmasters, the page is another good example of why boring HTML ages well. It is mostly a list of links with useful labels. Thirty years later, that is enough to reconstruct a little piece of game-development history.

This page came from the larger Ancient Web collection CacheRat is working through. The complete research list contains 1,967 sites for anyone who wants to do their own digging.

See the 1,967 Ancient Web Domains research list

The best part is that the old SimCity material is still sitting there waiting to be opened instead of summarized into oblivion.

Explore the SimCity archive and its original X11 material

Posted on by

Ancient Web: Professor Shoelace Built an Empire Around Knots

Ian Fieggen looked at the Internet and decided the world needed a definitive shoelace website.

He was correct.

Visit Ian Fieggen’s website

Fieggen, also known as “Professor Shoelace,” has kept the site online for more than two decades. The homepage says it has seen around 50 million visitors, and its biggest section is exactly what the nickname promises: lacing methods, shoelace information, knot instructions, and the “Ian Knot,” which Fieggen presents as the world’s fastest shoelace knot.

That would already be enough material for one extremely committed personal site.

Naturally, there is more.

A personal website is allowed to contain a person

The site also has sections for Fieggen’s old Windows software, computer graphics work, personal writing, family history, and a page about his father Chris Fieggen, an early computer programmer for ANZ Bank in the 1960s.

The software section preserves tiny Windows utilities such as JPGExtra and PNGExtra. The graphics section comes from decades of experimentation with computer graphics. The family section reaches into genealogy.

Nothing says these subjects belong under one modern content niche.

They belong together because Ian made the website.

That is one of the most useful things the old personal Web can remind a webmaster in 2026. A site does not always need to begin with a market category. Sometimes depth arrives because somebody keeps following their interests for twenty years.

And shoelaces turn out to contain plenty of depth.

There are practical tying problems, decorative lacing patterns, knot mechanics, common mistakes such as the crooked Granny Knot, and enough variations to justify an entire specialist reference.

The front page was updated in November 2025. It still looks and behaves like a personal reference site rather than a media property trying to capture a demographic.

This is one of the 1,967 domains gathered in CacheRat’s Ancient Web research list, where specialist sites like this are mixed with technical archives, museums, fan pages, historical resources, and personal rabbit holes.

Explore the 1,967 Ancient Web Domains research list

The funny part is that “a website about shoelaces” sounds trivial until you need to know something specific about shoelaces. Then suddenly you are very glad Professor Shoelace exists.

Explore Ian Fieggen’s website

Posted on by

Ancient Web: Sheldon Brown Built a Bicycle Encyclopedia by Hand

There are corporate bicycle sites with larger budgets than Sheldon Brown’s old personal pages and considerably less useful information about bicycles.

Visit Sheldon Brown’s site

Brown spent decades tinkering with bicycles and wrote extensively about them for magazines and the Web. His site grew into a dense technical reference covering brakes, frames, gears, drivetrains, wheels, fixed gear, tandems, touring, commuting, cyclecomputers, repair tips, old bikes, and a large bicycle glossary.

But the reason the site works is not just volume.

It has a person inside it.

Technical writing with fingerprints

Brown’s home page mixes bicycle expertise with photography, music, family stories, old mailing lists, Web notes, weird bicycles, personal beliefs, and links that made sense to him at the time.

Modern publishing advice would probably split all of that into six brands and a newsletter funnel.

The old Web let one person have a website.

That matters because the personality is attached to the expertise. Brown describes gathering discarded bicycle parts as a kid and building bikes to sell for pocket money. He later worked professionally around bicycles and wrote technical material under his own name and the pen name Christopher Joyce.

The bicycle articles are the durable part. Search for an obscure compatibility question on an older bike and Sheldon Brown’s name still has a habit of appearing.

The site also preserves an older webmaster attitude. Brown kept pages about browser behavior, HTML, spam filters, and learning to build websites. One line on the home page says you do not need books to learn to build Web sites because everything you need is on the Web.

That sentence belongs to its era, but the instinct is still good: explain the useful thing and link outward.

For webmasters, Sheldon Brown’s site is a lesson in accumulated authority. Hundreds of specific answers, written over years, become more valuable together than another polished homepage explaining that the organization is “passionate about cycling.”

This site was pulled from CacheRat’s wider Ancient Web corpus of 1,967 domains.

Explore the 1,967 Ancient Web Domains research list

If you own an old bicycle and are trying to understand what some strange component is doing, there are worse places to start than the archive Brown left behind.

Explore Sheldon Brown’s bicycle articles