Philippe A. scribbled on 9/24/06 1:53 PM:
>> Why not put the resumes into the database so you don't have this
>> linkage problem between the database and the resume?
>>
> 
> Because I almost have a complete working swish solution :-) It's true I
> could do that since I now know how to convert resumes to plain text. But I'd
> still need a piece of software that does the searches with booleans and
> such, calculate ranking, etc. Another thing I like from swish is its
> built-in ability to parse xml and html docs. I don't have tools to convert
> these formats at the moment. Good suggestion though. I may give it a try.
> 

check out SWISH::Prog::DBI on the CPAN: 
http://search.cpan.org/~karman/SWISH-Prog-0.02/

it's designed to give you full text swish search based on DB records.


> The one thing I have left to adress with swish is the incremental indexing.
> I would really like to avoid full reindexing when docs are
> added/changed/removed. But I'm not sure I want to use an experimental
> feature, as I may not be on site to assist my users if a problem occurs.
> 

I think what most of us do currently is set up a merging system to address this 
need. SWISH::Prog::Index has some methods to address this: merge() and add().

-- 
Peter Karman  .  http://peknet.com/  .  peter(at)not-real.peknet.com