Re: Updated Fresnel Ontologies and Examples - where to stop

From: Chris Bizer <>
Date: Sat, 12 Mar 2005 11:58:08 -0500

Hi Emmanuel,

yes, "where to stop" is an interesting question. The CSS guys answered it
with a 600 page spec ;-)

So question to everybody: What should we do?

I have the feeling that we should provide vocab for usefull stuff like
missing values, but put it definitively into the extended vocabs. I guess
that some browsers will implement these terms and others won't, but it is
still good to have them in order to prevent that every browser comes up with
his own term.

But generally you are, right: Where to stop?

> > # If several lenses are used together (fresnel:sublens) then a value box
> > contain another complete Fresnel box model.
> Here I have a slight concern with the whole box model. I understand that
> we can indeed repeat the full box hierarchy starting from what you
> called a container. But for me the container box that would appear here
> (1. container box) is slightly different from the top-level root
> container associated with the page (which is closer to the body element
> of an HTML page, as opposed to other container boxes that would be more
> like div elements). Just a question of vocabulary though... and I don;t
> really have any offer to make right now.

Hmm, yeah. I wouldn't understand the outer countainer to be attached to the
BODY element, more to the most outer DIV element.

I think the model is pretty easy when you are having only on set of styles
for the box models on all levels (standard case), but is also pretty
flexible when people want to do more complicated stuff.

I guess that the Haystack people are most experienced with this kind of
stuff. What do you think?

> > 2. I also added the possibility to style classes, which is important for
> > mixed content. Example:
> >
> > ex:gifImage rdf:type rdfs:Class ;
> > fresnel:displayAs fresnel:image ;
> > fresnel:styleGroup :groupOne .
> >
> > What do you think?
> This seems necessary indeed. Though I hope we won't keep adding new
> vocabulary items at the same pace.
> > 3. Going through the CSS Spec, I also added some more selector
> > pseudo-classes
> > like :link and :hover and some of their output media types like :screen
> > :print as style purposes.
> >
> > What do you think?
> I would have hoped we could delegate these :link/:hover/:visited
> constructs to the CSS layer, but as these are actually part of the
> selectors in CSS (which we don't have here), I do not see any other
> possibility than what you propose.
> As for the purpose controlloed vocab items, that's fine with me. They
> should probably not be direct values of purpose though (we might want
> something more structured as purpose values).

Don't know. As long as we don't have the terms for this something, I think
putting the purpose right there is OK. Dublin Core also got pretty far
without structure ;-)

Hope, that you are having better weather in France than currently here in


> Emmanuel
> --
> Emmanuel Pietriga
> INRIA Futurs - Projet In Situ tel : +33 1 69 15 34 66
> Bat 490, Université Paris-Sud fax : +33 1 69 15 65 86
> 91405 ORSAY Cedex FRANCE
Received on Sat Mar 12 2005 - 16:57:05 EST

This archive was generated by hypermail 2.3.0 : Thu Aug 09 2012 - 16:39:17 EDT