cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Eric SCHAEFFER <>
Subject RE: C2 Sitemap - Using a rule based engine ?
Date Fri, 01 Dec 2000 16:09:12 GMT
> -----Message d'origine-----
> De: Jason Reid []
> Date: vendredi 1 décembre 2000 15:57
> À: ''
> Objet: RE: C2 Sitemap - Using a rule based engine ?
> Just jumping in a bit on this thread...
> > On Fri, Dec 01, 2000 at 03:22:39PM +0100, Eric SCHAEFFER wrote:
> > > Using more powerfull rules, kind of "intelligence", being 
> > able to create
> > > personalized sites (using explicit and implicit 
> > personalization), etc...
> > 
> > Ahh, you mean so that, for example, the system could 'guess' what
> > areas of a site the user might be interested in? I agree that that
> > is an honourable goal, but I don't think this functionality belongs
> > in the sitemap. The sitemap is the central hub of Cocoon2, and needs
> > to run as fast as possible, which means keeping the complexity of it
> > relatively low. At present, the sitemap uses 'Matchers' and
> > 'Selectors' to choose which pipeline to generate a request with.
> > Although these are relatively simple tools, they provide an 
> incredibly
> > flexible architecture in which to manage a site. If you felt a need
> > to use AI to route requests to pipelines, there is no reason why you
> > couldn't write a selector using this technology to achieve it...
> > 
> Going as far as personalization with a rules engine is a much larger
> scope than the sitemap, and is something that could justify it's
> own OSS project.

Sure, it can go very deep. 

But personalization can also be personalized agregation of content (and
their position in the page) whitch overlap a bit the sitemap proposals...
(and what stylesheet to azpply, etc...)

That's why I thought about this.

> I haven't looked at the code for the Sitemap and how it applies
> matchers yet, but I have worked on a Rete-based rules engine.  The
> two concepts are very similar, in fact, I envisioned one use of
> the engine as routing requests to servlets.  The algorithms 
> developed for
> Rete and its successors might very well make the "Matching" process
> of the Sitemap more efficient (especially for larger, more complex
> sites).  Of course, it would add significantly to the work needed.

My impression was also that ReTe based engines were very quick...

> 	>	Jason Reid
> 		"Do not meddle in the affairs of programmers, 
> 		 for they are subtle and quick to anger."

View raw message