RE: editoral comments on rdf:text

> 
> My elevator pitch for rdf:text, and what I expected to see in the
> intro
> is: RDF has three types of literals (plain without language tag,
> plain
> with language tag, and typed), and sometimes when you're designing
> systems layered on RDF, this gives you three times the
> complexity. rdf:text lets you treat RDF as having just typed
> literals,
> so it's sometimes good for simplifying things.

Somehow I don't see it this way. The "plain" and "typed" flavors are side effects of the fact that text in RDF didn't deal with language identification in the first place. Compare, for example svg:text (and friends). The use of xml:lang to identify the natural language of string data is important--even critical--to processing natural language text.

There exist, of course, content with no language tags--simple strings--and these are valid and need support. But natural language text is at least as important in RDF.

> 
> (In retrospect, after all this, Boris, are you having second
> thoughts about using rdf:text at all in OWL 2?  :-)

Actually, the lack of support for natural language identification (via xml:lang or via other mechanisms like rdf:text) in RDF from the outset is an oversight that I18N has been requesting be addressed for some years now.

> > 2 Preliminaries
> 
> The link "Char" is http://www.w3.org/TR/xml11/#NT-Char

> I think you now mean http://www.w3.org/TR/xml/#NT-Char


It should.

> 
> I loved the first example.   (it's odd, but cool.)

The "foo-bar" language example?? I *hated* it. Invalid data is always uncool. 

Addison

Addison Phillips
Globalization Architect -- Lab126

Internationalization is not a feature.
It is an architecture.

Received on Tuesday, 7 April 2009 04:41:58 UTC