Experiment

I was talking to mpt on IRC yesterday evening, and he mentioned that I probably wouldn't like his latest entry on the Serendipitous Web. I can't really say because I never bothered to learn what all this "Semantic Web" business really means. But his statement that validation wastes authors' time struck me as a bit short-sighted.

I explained what I was thinking at the time, but I also wanted to write a real reply to his blog entry. The proper way to do that would be to put a reply in my own blog--except I don't have one. There are a few reasons why (in order of increasing importance) - One: I'm lazy. Two: I don't have much to say that hasn't already been said more eloquently by others. Three: I tried to keep a diary once and failed miserably. (This was in elementary school, IIRC.) Four: I also tried to maintain a "personal" website for about two years (It was is about HTML.), but I didn't really update that much either. So, blogging just didn't seem like anything I should waste other people's time with.

Except now I needed one. It's a lot of overhead for what may well turn out to be three entries total. Still, although I'm a terrible correspondant, fantasai doesn't seem to have that problem, so we'll see. :)

BTW, before I get into what turned out to be a rather long writeup on validation, I'd like to point out to mpt that Google doesn't work because links "serendipitously" happen to make it work; it works because the designers leveraged some inherent characteristics of the system. What lots of people find related and relevant, other people are also likely to find related and relevant. On the Web, a large and diverse population writes about topics it finds interesting and/or useful and--because this is hypertext--these writers add machine-readable links to related pages. A standard search engine treats the web as a database of independent files. Google treats the web as what it was designed to be and is--a hypertext system. It's not an accident that Google works. It's engineering.

Speaking of search algorithms, here's something AltaVista brought up: Theme-based searching