On Fri, May 03, 2002 at 12:17:10PM +0000, [EMAIL PROTECTED] said: > Recently I've been looking into a problem we've > been having with a system which indexes business > names using stopword lists(The list of common > words not to include) and porter-stemming (Which > reduces a word to it's common base, something you > seem to have done incidentally by change > 'Drinking' to 'Drink').
Sounds like what I want to do. Incidentally, what are you using for the stopword list and the porter-steeming? Existing code and lists or stuff you've done your self? > A real-world example of where this broke down for > us a little was when dealing with the business > name 'One To One'. Removing stopwords erased the > name entirely. Oops. Yes. I can see where that would be a problem :) Editor approval is definitely going to be required. Cheers Simon -- : omnipotence for dummies