Re: [office-metadata] RDFa model and xml:id

[Elias Torres <eliast@us.ibm.com>]

Svante.Schubert@Sun.COM wrote on 12/13/2006 08:44:35 PM:

> Bruce D'Arcus wrote:
[snip]
>
> Let me summarize:
> You think there might be an issue for the xml:id approach as you think
> it might be not be possible to map to RDF (without giving a example).
> Further you think it is not straight-forward (without giving a example).

Maybe you missed my email in which I said:

[[[You are correct. If we only use xml:id or whichever "linking" mechanism
between content and meta.xml we don't have semantics, except whatever is
expressed in the meta.xml. However, we can go very far with the approach
because that's how RDF works today. For example, I can assert any
statements I want about any web page on the Internet through an out-of-band
RDF file hosted on my site.]]]

I agree that we can achieve a lot that way, but in mail exchanges with
Patrick today I was pointing out reasons, why I'd rather go with the
hybrid. Which by all means is xml:id + meta.xml files + meta:about,
meta:property, meta:resource, etc. I need clear objections to why hybrid
doesn't satisfy your needs. (For example, the "good design is modular" is
not a proper objection for many reasons). I hope we don't keep raising
those kinds of objections because I don't think they are productive. I
would like somebody to help us get us organized in our discussion, because
time is ticking. Patrick, would that be you?

The reasons are: in-context metadata (which in turns affects copy and
paste), content-duplication. I hope I helped clear up some of the technical
questions Patrick had.

Unfortunately, I can't say the same about your proposal because you have
not explicitly stated on the mailing list or wiki. You have made mentioned
of xml:id, xpointer-based stuff, xpath stuff, styles-based stuff and more
recently binding a la XForms stuff. Which one is it? Please understand that
in order for us to compare things scientifically we need to see a proposal
in an email or wiki addressing the use cases.

>
> How can we compare the approaches in a more scientific manner?
> >
> > * I'm actually more comfortable with using the style redirection that
> > Florian likes to indicate properties than the xml:id approach for this
> > very reason (keep the package RDF files standard and clean). But that
> > would still involve a meta:property and/or meta:class attribute on the
> > style definition, in which case it's effectively XML window dressing.
> >
> Do we all agree on the saying that good design is a modular design?
> If there are two parts like style and semantic, which sometimes might go
> together - as when a content of a certain semantic is formatted by a
> certain style - but usually are handled separately, a modular design
> would not interleave them, but handle them separately to avoid
> unnecessary dependencies.

I think you are stretching the analogy a tiny bit. You are saying that
because in general a modular design is well-received that it means that
content needs to be separated into two files. Is CSS a modular design? Are
microformats a modular design? Are XML namespaces a modular design? Is HTML
a modular design? Is RDFa a modular design? There I just pointed to five
well known and (one not so well-known) standards that *are* considered
modular yet the content is on a single file.

>
>
> Have a nice day,
> Svante