RDF and (old) Atom discussion [was Re: Media Enclosures and RSS 1.0 / RSS 1.1]

(subject line changed as I rambled off topic)

On 7/23/05, Michael Zeltner <mzeltner@netalleynetworks.com> wrote:

I only saw the enclosures in RSS 1.0 used like
> your example from http://dannyayers.com/archives/2004/11/01/rss-10-
> enclosures/ hence my complaint.

Ah, right (*blush*). I've now added a link to mod_enclosure from that post.

> Sadly. Maybe we can get the process to accelerate a bit through just
> using RSS 1.0 enclosures in our feeds, while trying to get them into
> iTunes. Once iTunes supports it, everyone will probably going to
> implement it too. Hm.

I'd rather it were that way, but I'm not optimistic. 

> > But it's not all bad news, I think 'simple' RSS in the wild has been
> > getting progressively cleaner. I imagine the desire for people to get
> > their material on iTunes and to aggregate Yahoo! Media RSS will act as
> > fairly powerful validation processes for RSS+Media XML. If the XML is
> > half-decent, RDF is but a stylesheet away.
> 
> I'm probably too idealistic with wanting it to work without hassle :)

Heh, well quite. 

Frankly I was surprised at the level of resistance to RDF around Atom,
despite a significant proportion of developers knowing very well what
it's about. One or two RDF advocates lost interest after encountering
resistance early on, and at the start there was a lot more noisy
flaming, but fortunately that declined as the months and years wore
on...

I suspect what would have made a difference would have been one or two
high-profile applications in the domain with an RDF base, with
corresponding practical demonstrations of the benefits (I think the
nearest thing to that was BlogEd). A lot of weight was given to the
requirements of existing apps and there isn't really a good
counter-argument to that.

The "what matters is the bits over the wire" kind of argument
(spurious, IMHO) was also given a lot of weight, which made
suggestions of a model behind the format rather tricky. Despite the
fact that exactly such a model was being developed in the spec prose.

What I felt mattered most in relation to such a model was how
extensions were handled. As it happened there were others who
recognised the issue (namespaces offer separation but nothing more),
and despite heavy reluctance from some quarters a simple mechanism did
make it into the spec. The reluctance was understandable, there
weren't really any non-RDF, non-trivial extensions around while
discussion was going on.

Bit of irony there. Soon after consensus had been reached, in a flood
out came Yahoo! Media RSS, iTunes and Microsoft's 'lists' extension,
all for RSS 2.0. It remains to be seen how helpful Atom's extension
mechanism will actually be for interop, but at least now there's some
post factum justification for trying.

I've no idea where things will go from here around syndication. Oddly
enough the only fixture seems to be the RDF model, as a wallflower at
the dance. I continually find it hard to believe no commercial
syndication company has appeared to take it for a spin.

In practical terms there's certainly more work to do on getting a
usable representation of the media extensions in RDF, and mapping Atom
to RDF/OWL. Where and when that gets done is yet another question...

Cheers,
Danny.

-- 

http://dannyayers.com

Received on Saturday, 23 July 2005 15:30:43 UTC