Sun, 29 Oct 2000 19:35:14 +0000
Martin Gruner wrote:
> I am again coming up with a feature request.
> I want thesauri to be included (dictionaries for similar words) in sword.
> Why? Imagine a search for similar words. The user's search word list is
> expanded by substituting each word with itself and its synonyms.
> Might be very useful.
It is. My Java Bible program does this, the big problem is - where do you
get your thesaurus data from.
I have 2 solutions:
The official solution - wordnet, I'm sure you can find it in a search.
IIRC there is a C interface already. we had to write a Java one though.
However this is a large download 10Mb (english only) or something.
The cheat solution.
To find words meaning something like loving, you have a list of known
endings to words (s, ed, ing, etc) and you strip off any known endings
(leaving 'lov') and then you search for lov*. The results are actually
surprisingly good. I18N is a bit tricky with this though.