On Tue, 1 Dec 1998, Rainer Scherg wrote:

> Doing such things via perl is not the way to go, because it's to slow 
> and if you are doing such things via perl, you could also do the complete
> search engine with find -name '*.htm" -print | grep "$what"...

Not slow at all.  Check out http://www.wested.org/werc/lessons.html - 
these are four swish++ indexes, each indexes about 1000 documents.  The 
Perl script pulls-out the first 300 non-html characters from 
each document that matches the query, and uses that as the desription.  
This happens interactively for each search result.
The software all resides on an old Sparc-10 with 48MB RAM (about as fast 
as a 486-66).  Not too shabby for a Perl script...




> 
> 
> Regards Rainer
>