RE: Atom and RDF

["Danny Ayers" <danny666@xxxxxxxxxxx>]

> > Without knowing what "wibble" means in the abc
> > namespace, we still know from
> > the syntax that the entity 'something' has an
> > rdf:Property 'abc:wibble' with
> > the literal value 'Carrot'.
>
> This seems pretty meaningless to me and no different
> than what I'd get if I extracted unknown items with
> XPath/XQuery/XSLT.

You won't know that the entity 'something' has an rdf:Property 'abc:wibble'
with
the literal value 'Carrot'. This is only the simplest case, when you have an
RDF schema or even an OWL ontology as well you can reason with this stuff.
All you have with the XPath approach is some (possibly typed) data.

> > This is a huge head start for any processor. e.g. If
> > another system asks
> > "what properties has 'something'?", a system
> > ignorant of the meaning of
> > 'abc:wibble' can still selectively pass on the above
> > information.
>
> How is this any different from answering the question
> "what are the child nodes of 'something'?

The child nodes might have different relationships than abc:wibble.

> > > The link is
> > > http://www.kuro5hin.org/story/2003/8/8/122029/2205
> > but
> > > the site is currently down.
> >
> > I don't think the XSLT system you propose is 'that'
> > higher-order model
> > either - it might be handy for rearranging XML
> > documents, but doesn't
> > actually give you anything for the problem at hand.
> > XSLT probably passes
> > your Turing test already anyway - because it's
> > Turing-complete *, so the
> > system you describe, if possible, could be
> > constructed. (*Smart guy, that
> > Turing).
>
> I didn't propose any XSLT model. I asked that whatever
> technique you come up with for mixing namespace
> vocabularies should be able to solve the problem of
> automatically converting an XSLT processor to an EXSLT
> processor since this is "simply" a case of mixing
> namespace vocabularies. I specifically stated that I
> don't believe this problem can be feasibly solved and
> neither do I believe RDF is a stepping stone to
> solving this problem.
>
> For example, let's say my ATOM entries contain a
> <dare:stylesheet /> element that contains the text/css
> stylesheet that is supposed to be used to style the
> feed content

You'd probably want to use something like the DAML-S vocabularies to
associate the process with dare:stylesheet element. Not easy, but it
probably wouldn't be easy using any other approach either.

 or the author element has a
> employee:level that shows someone's level in the org
> chart which is is meant to be used to sort the posts.

That's easy, a pretty straight query would give you that.

> Now how does RDF bring my aggregator one step closer
> to doing the "right thing" in this case? Even if it
> does bring us "one step closer", is the increased
> complexity worth it given that there are probably nine
> hundred and ninety nine steps to go?

What increased complexity? Both of these examples would probably be
considerably easier to implement using RDF because you can reuse tools built
elsewhere - pretty much any of the RDF APIs would give you the org chart
ordering out of the box.

> My take is that RDF doesn't any practical benefit to
> mixing XML vocabularies in the general case nor does
> it even help the specific case of syndication feeds
> and news aggregators.

RDF/XML syntax provides a relational framework not found in XML.

> Now there may be other benefits of using RDF with
> Pie/Echo/Atom/Whatever but solving the mixing
> namespaced vocabularies problem is not one of them.

Edd Dumbill put it rather well : "failure-friendly interoperability".
http://usefulinc.com/edd/blog/2003/8/8#13:13

Cheers,
Danny.