> Date: Fri, 11 Jun 1999 11:02:17 -0700 (PDT)
> Reply-To: ron@library.ucsf.edu
> Originator: swish-e@sunsite.berkeley.edu
> Sender: swish-e@sunsite.berkeley.edu
> Precedence: bulk
> From: Ron Klatchko <ron@library.ucsf.edu>
> X-Listprocessor-Version: 6.0c -- ListProcessor by Anastasios Kotsikonas
> 
> At 08:50 AM 6/11/99 -0700, Antonio Cisternino wrote:
> >I've extended the swishspider to store in a DB the abstracts of pages.
> >I've used the DB_File package of perl that uses the dbm database (hash).
> >In this way I can return an abstract of a web page using a simple
> >perl script.
> 
> Isn't an anstract a property of the document and not the method by which
> swish accessed the document?  Why are you putting the feature in request
> method specific code?
> 

For two main reasons:

- the first one is that it is much more simpler write an abstracting procedure
  in Perl rather than C.
- the second one is: this structure is much more modular because a document
  abstract is tipically required only for Web searching. Moreover the
  abstracting procedure can be different for different documents and
  I think that is much more simple to rewrite a perl subroutine rather
  than a C function.

In addition I need this feature quickly and 20 minutes of work have solved
the problem.

-- Antonio