Category: Ancient Web

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

Posted on by

Ancient Web: Lena Games Still Sells Software the Old-Fashioned Way

Lena Games still looks like a software developer’s website rather than a software company’s marketing department, which is increasingly a distinction worth preserving.

Visit Lena Games

The site belongs to Lena Pankratova and centers on a long catalog of puzzle and casual software: jigsaw programs, solitaire collections, Taipei-style tile games, checkers, Sokoban, mosaics, timers and small utilities.

The presentation is straight out of the independent shareware tradition. The front page introduces the developer, gives an email address, announces updates, then gets on with listing the software.

No cinematic hero section. No pricing funnel that waits three clicks before admitting what the product does.

The flagship material is heavily puzzle-oriented. BigJig supports a large collection of downloadable jigsaws and includes a utility for creating your own. The Astra Jigsaw line has accumulated themed collections covering paintings, countries, landmarks and regions. BigPatience collects dozens of solitaire games. Smaller programs include FineCheckers, MineFinder, SokoMan, BowlingTile and NewLines.

A software catalog that kept moving

What makes the site especially interesting is that it does not appear to have survived by becoming a museum.

The front page still carries current-looking update notes. Programs have gained Mac support, Android synchronization and support for newer operating-system releases while the surrounding site retained the compact catalog structure of an older software business.

That combination is rarer than it should be.

A lot of independent Windows software disappeared because its distribution depended on download portals, expired domains or companies that were effectively one person. Lena Games shows the opposite case: one developer’s catalog remained findable because the developer kept the site and the products attached to each other.

There is useful web-design archaeology here too. Each program gets a short explanation, platform indicators and a direct path toward more information. The site assumes visitors can read a list and decide what interests them. It does not attempt to personalize the experience or trap the user in a carousel.

For anyone who remembers downloading small Windows programs from individual authors, the structure is immediately familiar. More importantly, it still works.

The lesson is not that every modern software site should look twenty years old. It is that a durable catalog needs stable pages, clear product names and enough plain text for a human or search engine to understand what is actually being offered.

Lena Games has a lot of that accumulated surface area now.

Browse the Lena Games software catalog

Posted on by

Ancient Web: The W3C Still Hosts the XHTML2 Working Group’s Dead Future

The W3C still hosts the home page of a Web standards group whose future did not happen.

Visit the archived XHTML2 Working Group page

The page belonged to the XHTML2 Working Group, chartered in 2007 and closed in 2010. The W3C now labels the page as a historical archive and points current visitors toward HTML5 material, but the old working-group site itself remains intact.

That means the abandoned path is still documented in its own words.

The page contains news posts, specifications, public drafts, issue links, test suites, tutorials, slides, validation material, translations, meeting information and the group’s charter. It explains XHTML as an XML-based successor to HTML and lays out the modular approach the group expected to support everything from mobile devices to forms, mathematics, vector graphics and richer structured data.

The branch that lost

The most interesting line is dated July 2, 2009. The W3C announced that the XHTML2 Working Group charter would not be renewed and that resources would instead be increased for HTML5.

You can read that decision today while standing inside the site of the technology being wound down.

The archive preserves a lot more than XHTML2 itself. It links XHTML 1.0, XHTML Basic, XHTML 1.1, XHTML Modularization, XML Events, RDFa work and earlier HTML standards. There is even a compact history of HTML 4.0, HTML 3.2 and HTML 2.0.

For developers who arrived after HTML5 became the obvious center of gravity, this page is useful because standards history looks much cleaner in hindsight than it felt while it was happening. There were competing assumptions about how strict markup should become, how XML should fit into the Web, how devices with tiny memory budgets should consume documents and how future languages ought to be assembled from modules.

The site also demonstrates something the W3C has generally done well: old standards material remains linkable. A specification can be obsolete without being erased.

That matters. Technical history becomes much harder to reconstruct when the only surviving explanation of a failed idea is somebody else’s retrospective summary.

Here the drafts, announcements and vocabulary are still sitting on the original server.

The future changed direction. The directory stayed put.

Explore the W3C’s XHTML2 Working Group archive

Posted on by

Ancient Web: Nintendo Still Hosts an Old F-Zero Product Page

Nintendo still serves a small official page for the original F-Zero, and it feels less like a modern product page than a file somebody forgot to remove from a very old directory.

Visit Nintendo’s surviving F-Zero page

The page identifies the game simply as F-ZERO and gives its Japanese release date: November 21, 1990. Around that are small images for the package, title screen, gameplay and characters.

That is basically the page.

Modern Nintendo product sites arrive with video backgrounds, responsive layouts, legal overlays, analytics and enough JavaScript to qualify as light industrial machinery. This one delivers a few graphics, a short description and links. Even better, some modern browsers can display the old Japanese text with broken character encoding unless they guess the page correctly.

That mojibake is not useful, exactly, but it is authentic archaeological dirt.

A product page from another Web

The text describes the game’s futuristic racing setup, its machines and circuits, and the basic goal of learning how each craft behaves while chasing faster times. The screenshots do most of the sales work.

What makes the page interesting now is not that it contains secret F-Zero information. It does not. The value is that Nintendo kept the original-style page reachable under its own domain while the company, the Web and the hardware around it changed completely.

Official corporate sites are usually terrible at this. Old product material gets migrated, redesigned, summarized, redirected or deleted. A current catalog page may tell you that a game existed, but it rarely preserves the way the company presented it during an earlier generation of the Web.

This one survived with its tiny images and old layout intact.

For web researchers, the URL structure is worth noting too. The page lives deep under Nintendo’s old n02/shvc path, a directory naming scheme tied to the Super Famicom era rather than a modern content-management system. Stable ugly URLs win again.

There is almost nothing here, and that is precisely why it is such a clean artifact. No retrospective. No remaster pitch. No “legacy” branding.

Just F-Zero sitting where Nintendo left it.

Open Nintendo’s old F-Zero page

Posted on by

Ancient Web: Jon Skeet Still Has a Worms Fan Page

Before Jon Skeet became one of the most recognizable names in programming Q&A, he had a Worms fan page.

Visit Boggy B’s “Armed to the Teeth” Page

The surviving page is wonderfully direct. It identifies itself as unofficial, notes that the views are Skeet’s own rather than the University of Cambridge’s, acknowledges Team17’s trademarks, and then gets to the important business: Worms.

The page offers personal tactics, cheats and Challenge mode codes, a FAQ, PC utilities, favorite landscapes, animation screenshots and links to other Worms sites on the Web.

That list captures what a serious game fan page looked like before a wiki absorbed every category into one database. The author did not merely write about liking the game. The site collected the things players actually needed.

Information for addicts

The page’s own description says it is not intended to advertise Worms, only to be a source of information “for addicts like me.”

That sentence could serve as a mission statement for half the useful old Web.

Fan sites were often built by somebody who knew a subject well enough to realize that the official material was not going to answer every question. So they made their own FAQ. They saved codes. They wrote utilities. They linked to everybody else doing similar work.

The title, Boggy B’s “Armed to the Teeth” Page, is itself a period Worms reference. Boggy B appears in the original Wormsong and became one of the recurring names associated with the series.

The page also preserves the old expectation that the Web was conversational. Skeet explicitly invites visitors to email him with material to add and even provides an address for people whose browsers do not have built-in mail support.

That line dates the page more effectively than any copyright footer could.

There is no account system, no voting mechanism and no comment platform. If you knew a trick the page should contain, you mailed the webmaster.

Modern community databases are much better at scale. What they often lose is the visible presence of one person deciding what belongs, explaining it in their own voice and wiring the useful pieces together by hand.

The Worms page is small, but that is part of why it survives so cleanly. A couple dozen lines of HTML can carry a surprising amount of cultural residue.

Browse Jon Skeet’s old Worms page

Posted on by

Ancient Web: Yoda’s Datapad Turned Star Wars Continuity Into a Spreadsheet

Somebody got tired of wondering how old everyone in Star Wars was, so Yoda’s Datapad turned the problem into a giant table.

Visit the Star Wars Character Age Table

The page lines up characters against the movies and works out their ages at each point in the timeline. Luke, Leia, Han, Obi-Wan, Yoda, Chewbacca, Palpatine, Rey, Kylo Ren and a long list of supporting characters all get rows.

The result is one of those fan references that looks simple until you notice the continuity problem hiding underneath it.

Star Wars has two large bodies of chronology to reconcile: the older Expanded Universe material now called Legends, and the post-2014 canon. Yoda’s Datapad does not quietly mash them together. Where the newer canon and older sources disagree, the page gives separate entries.

Anakin Skywalker is a good example. The table preserves different age calculations from pre-2014 material and newer canon sources, then explains why older references could make him appear a year older in Attack of the Clones. Han Solo also gets separate canon and Legends rows.

A fan page built like a reference book

The table uses the Battle of Yavin as its chronological zero point. Birth dates are expressed relative to Episode IV, then translated into ages for the major films. That makes the page useful even when the answer is not especially intuitive.

Chewbacca is listed as roughly 200 years old around the original film. Yoda reaches 900 by Return of the Jedi. Maz Kanata crosses the thousand-year mark by the sequel trilogy. C-3PO gets the appropriately messy treatment of being “built” at one point from components that may themselves be much older.

The notes matter as much as the numbers. The author identifies approximations, contradictions and places where a source only establishes a minimum age. That is better reference work than pretending fictional chronology is cleaner than it is.

Yoda’s Datapad also maintains a Star Wars book timeline, which fits the same basic philosophy: make a useful map through a franchise that accumulated decades of overlapping material.

This is one of the things independent fan sites did especially well. Instead of writing another article about the same famous characters, somebody built the odd reference table they personally wanted and kept maintaining it.

Search engines can answer “How old was Luke in A New Hope?” now. The underlying fan work is still more interesting because it shows where the answer came from, where the sources disagree and where the universe refuses to behave like a clean database.

Explore Yoda’s Datapad character age table

Posted on by

Ancient Web: David Wonn’s Nintendo 64 Glitch Archive Is Still Ridiculous

David Wonn’s Nintendo 64 glitch page is what happens when enough players spend enough time poking old games in places the developers never expected anybody to reach.

Visit David Wonn’s Nintendo 64 Glitches archive

The archive is huge. The first page alone runs from A through K, with separate pages continuing through the rest of the alphabet. Entries are organized by game, and many preserve the name of the person who submitted the discovery.

That makes the site more than a cheat list. It is a record of a particular kind of player research from the cartridge era: people repeatedly trying jumps, camera angles, collision seams and button combinations, then emailing somebody when the game did something stupid.

Banjo-Kazooie is especially well represented. The page documents ways to slip through polygon seams, stand on the bottom of a lake, fall forever without actually falling, pass through a snowman, bypass windows and abuse the Beak Buster move for temporary invulnerability.

Some instructions are almost archaeological field notes.

A player notices two sand polygons meeting at a line, follows the seam into the ocean, presses Banjo’s head into the geometry and eventually gets the engine to stop drawing part of the island. Another entry describes a floating tree in Aidyn Chronicles that is visibly disconnected from the ground and hollow underneath.

Bugs were social objects

Before YouTube walkthroughs made every exploit instantly visible, a glitch could circulate as text. Somebody discovered it. Somebody else reproduced it. Somebody wrote down the sequence carefully enough that strangers could try it on their own hardware.

The tone reflects that. Instructions include warnings about sharks, comments about attempts that took “about 100 tries,” and occasional admissions that a trick might not technically be a glitch at all.

There are also hard crashes. One Banjo-Tooie entry explains how releasing a battery in the wrong place can freeze the game while the music keeps playing, requiring a reset. That is exactly the sort of detail a polished strategy guide would normally omit and exactly the sort of thing players remember.

For game researchers, these pages preserve behavior that can disappear when software is patched, emulated differently or simply stops being played on original hardware. They also preserve the informal language players used to document technical behavior before “speedrun tech” became a much more formal discipline.

The page design is basic because it only needs to do one thing: hold an absurd quantity of very specific information.

It succeeds.

Dig through the Nintendo 64 glitch archive

Posted on by

Ancient Web: Oklahoma State Put a 1998 Geometry Class on the Web and Left It There

In fall 1998, students in an Oklahoma State University geometry class made displays about border patterns. Somebody put the work on the Web. It is still there.

Visit the GeoSET Border Pattern Gallery

The material came from MATH 3403, Geometric Structures, and was part of Oklahoma State’s GeoSET project: Geometric Structures for Elementary Teachers. The larger project was designed around teaching geometry to future elementary-school teachers and helping them carry that material into classrooms.

The border gallery is wonderfully specific. Students were asked to imagine an educational conference where seventh- and eighth-grade students and their teachers would visit campus. Each group had to build a display explaining border patterns in a way those visitors could use.

The surviving project list includes titles such as All About Borders, Can You Identify the 7 Symmetries, Border Safari, Different Types of Symmetry, Borders, Borders, Borders, Under the Sea Borders and a Christmas-themed border project.

Seven ways to repeat a strip

A border pattern is a design that repeats along one direction. What looks like decoration becomes a compact lesson in symmetry: translation, reflection, glide reflection and rotation combine into exactly seven mathematical classes of strip patterns.

The GeoSET pages turn that fact into something visual instead of burying it in notation. The site also preserves supporting material intended for teachers, including a downloadable border archive with PDF, Word, Adobe Illustrator, EPS and BMP versions of designs.

That file-format list is a small time capsule by itself.

The archive documentation explains that the Illustrator files were the highest-quality versions, while individual patterns were also supplied as 1,500-pixel-wide BMP files and 360-point EPS files. The stated purpose was practical: teachers could cut, paste and reuse the patterns when preparing exams or other curricular material.

This is the kind of university Web material that ages unusually well because the underlying subject did not expire. Browser fashion changed. The seven border symmetries did not.

There is also something worth stealing from the site’s structure. The project does not make visitors click through six layers of institutional branding before reaching the work. The gallery names the projects, links the student material and provides the files. The useful thing is close to the surface.

A class assignment from 1998 became a public teaching resource simply because somebody bothered to leave the directory online.

Explore Oklahoma State’s surviving Border Pattern Gallery

Posted on by

Ancient Web: A 1999 Stanford Page Still Explains How Bunyaviruses Replicate

A Stanford student page from winter 1999 is still online explaining viral replication with a few paragraphs, a couple of images and no attempt to turn itself into a modern publishing platform.

Visit the 1999 Stanford Bunyaviridae replication page

The page was created by Arthi Chakravarthy for Stanford’s Human Biology 115A course, taught by Robert Siegel. Its subject is the replication of the Bunyaviridae family inside a host cell.

The material starts with the segmented viral genome. The L, M and S segments are described separately: the L segment encodes proteins used for transcription and replication, M produces a polyprotein associated with envelope glycoproteins, and S encodes the nucleocapsid protein.

Then the page walks through the infection cycle in order.

Attachment, copying and exit

Surface proteins bind to a host cell. Entry depends on acidic conditions. The page then explains that bunyaviruses carry their own polymerase machinery and transcribe RNA from the genome’s three segments.

One of the more interesting period details is the discussion of “cap stealing.” The page compares the process to influenza and notes that transcription takes place in the cytoplasm rather than the nucleus. The explanation reflects what was being taught and understood in that course in 1999, which is exactly why the page is useful as a historical academic artifact as well as a biology reference.

Translation gets its own section. L and S segments are described as being translated by free ribosomes while M uses membrane-bound ribosomes. Assembly occurs in the Golgi apparatus, where viral particles acquire membrane and glycoproteins before leaving the cell through exocytosis.

The page ends with an electron micrograph and a link back to the larger Bunyaviridae project.

This was a common form of early academic Web publishing: students built actual public pages around course material. They were small, linkable and independent enough that individual sections could survive long after the class ended.

A modern learning-management system might hide the same assignment behind authentication, course expiration and a JavaScript application that dies when the vendor changes something. This page is just HTML. Twenty-seven years later, the lesson still opens.

That does not make every scientific statement on a 1999 student page current medical guidance. It makes the page a clean snapshot of how the subject was being explained at Stanford at the end of the 1990s.

Read the surviving Stanford page

Posted on by

Ancient Web: The Cunningham Project Is Still Factoring Huge Numbers

There is a Purdue web page where enormous unfinished integer factorizations are still maintained like an active wanted board.

Visit the Cunningham Project

The Cunningham Project is devoted to factoring numbers of the form b^n ± 1 for bases 2, 3, 5, 6, 7, 10, 11 and 12. That sounds narrow until you open the tables and discover how quickly “narrow” turns into thousands of giant integers, partial factors, composite cofactors and unfinished business.

The project is associated with the Cunningham tables published in Factorizations of b^n ± 1, by John Brillhart, D. H. Lehmer, J. L. Selfridge, Bryant Tuckerman and S. S. Wagstaff Jr. The first two editions existed as paper books. The third edition, published in 2002, became an electronic book with machine-readable tables.

That transition is visible in the surviving site. The front page is mostly links and text files. There is no app pretending to be a document. If you want the main tables, you get the tables.

A mathematical job board

One of the best surviving pieces is the project’s “wanted” list. It names composite cofactors still worth attacking, including entries hundreds of digits long. The notation is terse because the intended readers already know what they are looking at.

The main tables are much larger. A table for factors of 2^n - 1, for example, marches through exponent after exponent with known prime factors and unresolved pieces. Some rows collapse neatly. Others end with labels such as P57 or larger unresolved cofactors.

The home page was still being updated in January 2026. It links to current wanted lists, champion results, the latest main tables, appendices, old pages and a page showing who is factoring which number.

That last link says a lot about the culture behind the project. This is not a frozen exhibit. It is a long-running cooperative calculation problem that happened to acquire a Web front end.

For modern webmasters, the page is also a useful reminder that a specialized site does not need much machinery when the information itself is the product. Plain links, plain text, durable filenames and stable URLs have kept a highly technical research resource usable for years.

The Cunningham Project is not visually impressive. It does not need to be. The interesting part is that a mathematical effort old enough to have originated in printed tables is still being extended through one of the simplest interfaces the Web can provide.

Explore the Cunningham Project and its current tables

Posted on by

Ancient Web: A Boston University Page Still Teaches How Transistors Become Logic

A computer can feel impossibly abstract until somebody shows where the ones and zeroes physically come from.

Visit Boston University’s From Transistors to Functions module

This old Boston University computer-science lesson starts low in the stack: a transistor has terminals, a control input and two useful states. From there it builds upward until those switches become logic gates and those gates become functions.

The page’s teaching trick is a faucet.

A transistor is compared to water controlled by a knob. Turn the gate on and current can flow. Turn it off and the path closes. Complementary CMOS transistors reverse that behavior, giving the two kinds of switches needed to construct digital logic efficiently.

It is simplified, but it gives beginners something physical to hold onto before the diagrams multiply.

Two transistors, then a computer

The lesson next connects transistors into an inverter, or NOT gate, and works through both possible inputs. A truth table reduces the circuit to its behavior: zero becomes one, one becomes zero.

Then come NAND and NOR.

The page does not simply present their symbols. It walks through which transistors are conducting for each input combination and derives the output. NAND can then feed a NOT gate to produce AND; NOR can feed a NOT gate to produce OR.

By the time the familiar logic symbols appear, the reader has already seen what they are hiding.

The module continues into larger functions and eventually constructs XOR from combinations of simpler gates. From there the conceptual ladder is clear: transistors make gates, gates make larger logic functions, and enough logic functions make the machinery inside a processor.

The page includes period markers almost accidentally. It refers to roughly one million transistors per square millimeter as circa 2006 technology and to an Intel processor with more than 1.5 billion transistors as circa 2007.

The underlying lesson is older still. The material says it was adapted from lecture notes by Azer Bestavros first developed for CS-101 in 1995.

That gives the page an interesting durability.

Processor counts have exploded. The fundamental bridge from switching devices to Boolean logic has not.

A modern course could rebuild the lesson with animations, interactive widgets and a JavaScript simulator. Those might help. But the old page already has the essential pieces: diagrams, truth tables, a patient sequence and enough explanation to make the abstraction collapse one layer at a time.

The Web is good at preserving tutorials whose examples become dated while their reasoning remains useful.

The transistor count changed.

NOT is still NOT.

Read the full From Transistors to Functions lesson