On Mon, Oct 27, 2008 at 17:32, Alan Runyan <runyaga@runyaga.com> wrote:
I agree with Tres. A lot more can be done with Indexes and Catalog without caching.
The most exiciting development in Catalog optimizations comes out Jarn. Helge Tesdal (iirc) did a buncha work at a RDBMS company when he was in college. He has a protoype of a query plan for ZCatalog.
http://www.jarn.com/blog/catalog-query-plan
I would like to ask Roche and others to look at the Query Plan.
Caching is a total PITA because invalidation machinery becomes overwhelming complex and unwieldly quickly in production.
I don't know very much about searching, but this definitely sounds like a good idea. Also, especially when doing free text searching that has large result sets, incremental searching is very beneficial. I know Dieter Maurer has made a Zope2 implementation of this. I would be interested in seeing a bunch of Gurus sit down at some sprint and trying to come up with a catalog engine that is incremental and uses query plans. There is no reason that would not be stupidly fast. :) We can then make a new catalog that uses this engine but has the same API as the old one, to ship with some future version of Zope, say 2.12. -- Lennart Regebro: Zope and Plone consulting. http://www.colliberty.com/ +33 661 58 14 64