|
Impressive work, Clemenseken. I have a question, however: would there be a possibility to change the dictionary file format from txt to something more effective from the point of searching (flat file db or something like that)? I put together a thesaurus for english, which has approx. 14MB (large, I know...) and it works very fine when I search for word from the beginning of the alphabet. But if I look for something starting with 'z', it tooks approx. 2-3 minutes to get there...
Maybe it would be benefitial to implement at least some kind of heuristic based on the fact that the dictionary file is ordered alphabeticaly (which is, I think, a reasonable prerequisition which can be fulfilled very simply by sort-ing the file before uploading) and based on that perform an interval-based split search. I don't know the inner workings of your program, but maybe there is a way how to speed-up the search...?
|