Submitted by: Bruce Lewis
We are sorely lacking a search service for web.mit.edu pages.
Harvest seems to be our only choice, as described in
for making MIT's web pages searchable. Version 1.3 worked well for
several weeks, but then dumped core, probably from a bug that's been
fixed in 1.4. There are also other improvements in 1.4, including
improved handling of files in AFS.
One day getting the new version into CVS and reintegrating changes
done to 1.3. (This will be much less effort next time due to
CVS's ability to track 3rd-party sources.) One day to get installation
working right. A couple day's effort testing, spread out over
I think Harvest will initially need a lot of attention because
we need to keep an eye on performance/load. Also, other MIT
web maintainers will want to integrate with us using advanced
Harvest features; we will probably want this too.
Few users will use any of Harvest's features beyond what's
intuitive. Consultants can quickly figure out the advanced features.
Last modified: Wed Jan 31 15:36:11 EST 1996