The Oldest Webpage

I love discovering things, which is why I'm into history. Of course, things can only be discovered once within popular memory, so when I find anything of the merest importance I'm very proud of the fact. One of the periods of history I'm most interested in is the early web, and I was therefore quite excited to read a news item on Slashdot at the moment called The Web Is 16 Today. It refers to the least recently modified web page known (Link.html from 13th November 1990), which is, in other words, the earliest HTML document on the web. This rang a bell with me because I'm pretty sure that I was the one who discovered it! So I went ploughing through some more recent history to find out whether I was really the first to find it or not.

The page that Slashdot cites is the W3C's History Timeline, which was created by the strangely accented synaesthetic early-web guru Robert Cailliau, but is apparently now maintained by Dan Connolly. I'm pretty sure, however, that TimBL has a hand in editing it, and I recall having mentioned Link.html to him some years ago. By searching the logs of Swhack, an IRC channel I set up with Aaron Swartz in 2001, I managed to find my original discovery, from the 12th December 2001. I also found the point where I mentioned it to TimBL and Dan Brickley on #rdfig, the Semantic Web Interest Group's old IRC channel, in March 2002. So was the W3C's History Timeline edited before or after I passed the Swhack discovery on?

At this point normally the Web Archive would come to the rescue, but sadly the W3C's robots.txt file disallows /History, meaning that it's impossible to find out when it was edited. There is, however, a copy of the page on CERN from 16th February 2001 (v.1.24), and the reference to Link.html evidently was not yet included therein. There's also a usenet message from 12th May 2003 which does include the reference, so it must've been added between these dates. Unfortunately I don't have a local copy archived, so I can't tell with any more accuracy.

Incidentally, I found Link.html by looking through hype.tar, an early tarball of the original website worked on by the CERN team. The first code, HyperText.m dates from 20th September 1990, as I wrote in my old essay an Early History of HTML. It also shows that HTML is descended from CERN's SGMLguid language, which is from Script GML and GML itself; the CERN SGMLguid language, which shows an amazing resemblance to HTML, is from at least 1986.

Note that though Proposal.html in the same historical directory as Link.html is intriguingly dated "12 November 1990" in its content, and though it may have been an HTML on that date, its headers sadly say that it was modified on the 7th October 1991 so we can't tell if it really is older or not.