Skip to main content.
home | support | download

Back to List Archive

Re: Re: SOIFs (fwd)

From: Simon Wilkinson <sxw(at)not-real.tardis.ed.ac.uk>
Date: Thu Oct 22 1998 - 13:27:11 GMT
> Recent versions of Harvest seem to come with support for Swish-e built
> in. This is done at the broker level (whereas the per-filetype
> summarising is I believe done by the gatherer component).

The support for SWISH is in the development engine, and is development
code ...

> The broker config file has a parameter that allows you to choose the
> indexing engine; it defaults to glimpse but can be set to use swish. I
> think(?) a version of swish is bundled but presumably could be replaced,
> much as the Glimpse component can (and should!) be replaced by a more
> recent version.

Well, at the moment the version of SWISH thats included in the development
code base is seriously hacked, mainly to allow it to sensibly recognise and
index SOIF files. I have been meaning to contribute these changes back to
the SWISH development effort, but as yet they're not really stable enough
to do so. As you mention, SWISH performance with large document sets is
fairly poor (although some changes can be made to improve it) - I've
yet to investigate SWISH-E, although it does look like it may be a good
solution.

> > Can Swish recognise file format types such
> > as Word, PDF, SGML etc ? Are there any good examples of Harvest and
> > Swish used together ?

If you are using Swish with Harvest you don't need to be able to recognise
any file format other than SOIF. The Harvest gatherer robot handles all of
the necessary file format conversion.

Cheers,

Simon.
Received on Thu Oct 22 06:37:27 1998