One of the best things about text on the web is the way you can link to other stuff. When I’m writing here I always try to link words and phrases to more in depth explaination or examples. Mainstream news sites like the BBC (see, there’s an example) don’t do that. The BBC often have a few external links to the right of the main story and they’ve just started linking to the same story from other (approved) news sites, but the body text is always untouched. Which is a shame.
At EuroFoo earlier this year I heard Stefan Magdalinski talking about TheyWorkForYou.com. After the presentation we had a brief discussion about some things they were thinking of adding. One topic that was discussed was the possibility of recognising proper nouns in text and automatically linking to a relevant web page. I mentioned Simon Cozens’ Perl module Lingua::EN::NamedEntity which would be a first step along this path.
And this morning, the blogsphere is full of reports of Stefan’s latest project. It’s a proxy front end to the BBC news site which searchs the text for proper nouns and makes them links to the relevant page on Wikipedia. For example, see this page about the Pitcairn Islands rape trial.
Stefan says he’s been learning PHP so it doesn’t look like he’s been using Simon’s module. It would, of course, have been easier in Perl :)