[ZDP] Some remarks about the portal design(s)

Christian Scholz cs@comlounge.net
Thu, 11 May 2000 22:05:15 +0200


Hi!

> > 2. move the archives to a Zope based solution. This is really low
> > priority though. It's a pain to make changes to them, but they do work
> > on the whole. They're based on proven high-volume archiving and
> > searching technology (Lotus Notes) and, although it would be fun to try,
> > I haven't got the time or confidence to move it all into a ZODB and use
> > a ZCatalog. Hopefully someone at DC is going to send me some copies of
> > the CDs which were done with the list archives on them and then maybe
> > I'll see...
> 
> Do you have an idea how you can integrate searching technology like
> htdig in Zope ? Has anybody tried/succeeded at it? I guess we need a
> new Subject: "Search Engine Integration" in the "Developers" portal :-)
> 
> 
> [rh]Hm. By wrapping it, supposedly. There are usable tools for this in
> python already, however. There is an ftp mirror script in the standard
> distribution. There is also a sophisticated python linkchequer tool called
> linbot (<url: http://starship.python.net/crew/marduk/linbot/> modelled after
> linkbot I believe, but I do not know that). Licence is GPL.

Hm, I am wondering how such an integration should look like? Basically I guess
that one would use some Zope stuff instead of the needed CGI. Thus access
to the htdig database from Zope is needed.. Unfortunately I don't know how 
htdig stores it's data. With udmsearch on the other hand it should be quite
easy as everything is stored in a mysql database and thus accessible through
Z SQL methods (though I would like some more abstract product or whatever which
has some more knowledge about the structure of a search engine).

I guess this is not solveable in a generic form then.. 

cheers,
  Christian

PS: I don't get the relationship to linbot right now.. This is not a search engine
    but a link checker as said above.. huh? ;-) ?!?