Hi,

is it possible, at least theoretically, to categorize files based on a swish-e 
index?
What I'd like to do, is to calculate the "nearness" of documents.
The practical problem involves a large amount of support data about different 
products. The goal would be to get some idea, how many calls are actually 
about the same problems (despite different error description texts).

I couldn't find any description of the index format (I probably didn't look 
hard enough), so I have no idea if such a task would be doable.

Thanks for your input

Stephan