Re: Ontology-driven website

On Tue, 17 Jun 2008 00:03:26 +0200
Andreas wrote:

> do you mean an ontology-based "modeling framework" for modeling and  
> building websites based on an abstract model like OOHDM, etc.  
> (Daniel's pointers) or do you mean a library+API like php's symphony  
> or something like the Drupal core, which can also be extended with  
> modules based on callback hooks. There is sth. going on in the Drupal  
> community you should look at: http://groups.drupal.org/semantic-web

Unfortunately I took just a quick look at that page. Actually the linked resources recieved in your replies provide me a huge amount of data to be parsed ;) I should anyway thank you very much, this gives me something to think about, and I hope I'll have enough time to digg more into this.

Actually for what I've seen so far, a nice implementation problem, from the web-engineering side, is the linking between the model described and the code to be executed. I think that Schwabe's work is very well articulated on this and I was thinking about inserting an "abstraction level", an indirection level, moving part of the ontology inside the framework and hence use it to model the real site ontology. In other words, creating a set of small types bound to specific actions in the framework, that can be shared among all the instantiated websites and used as basic building blocks.
Replying to you, Andreas, this means both providing a sort of (semantic) library+API (like the drupal core) and even "modelling and building a website based on an abstract model".
Another slightly different approach is to let the user/designer model the "modeller" thus providing extensibility. The "extension modules" imported could be introduced coherently in the site/framework ontology.
I don't know if this pushes the implementation side too far from the actual needs, if it is totally futile or if I should shrink a little bit, thinking about the possibility to import new ontologies into the framework/site model and extend the framework with new vocabularies.

Anyway this is just a part of the problem, other problems, like where and how is modeled the presentational view are still to be discussed more carefully.



> I think it's best to re-use existing vocabs like FOAF, SIOC, DOAP, ...  
> and add your own vocabs when you need new concepts.

sure, that's another thing that what we are looking for.

I'm still reading the docs and several different approaches, with an eye on the implementation.
Again, if you have other comments or questions on what I wrote, feel free to ask me.

Thank you in advance,

-- 
Matteo 'Peach' Pescarin