[Zope] - Finding recently changed pages

Kent Polk kent@eaenki.nde.swri.edu
Tue, 8 Dec 1998 20:02:54


Hi Jim (Jim Fulton), in <366DA7FE.BA753CC@digicool.com> on Dec 08 you wrote:

> Kent Polk wrote:
> > 
> > Hi A.M. (A.M. Kuchling), in <199812061624.LAA25965@207-172-39-232.s232.tnt10.ann.erols.com> on Dec 6 you wrote:
> > 
> > > At http://discuss.userland.com/msgReader$857 , Dave Winer suggests
> > > that Web sites should have a top-level siteChanges.xml file that lists
> > > URLs that have recently changed.  Search engines could then use this
> > > information to only crawl pages that have changed recently, which
> > > would let them keep up to date more easily.  How would one go about
> > > implementing something like this under Zope?
> 
> Andrew, 
> 
> I seem to have missplaced your original post, so 
> I'll reply here.  One problem we've run into whenever we've approached
> this is that it's really hard to *define* when a URL changes.  If
> A URL is computed fron n objects, when does a URL change?  When one of 
> the n objects have changed?  When they've all changed?  When m 
> (1 <= m <= n) have changed? When some particular subobjects have changed?
> I suspect that the definition could be application dependent.

Exactly. That's why I was wanting a tree-walking solution for having objects
contribute their modification date, or something like that.  Basically have 
a way to set what objects will respond to a give_me_your_modification_date 
method. Both standard objects as well as application objects could respond
as needed (such as query methods, etc.)

Kent