> > I'm not sure I understand your point here. RSS 1.0 certainly also has a
> > URI identifying the entry (or 'item') -- the rdf:about attribute.  So
> > this could be written as:
>
> Would you use the rdf:about attribute as a GUID? I sure wouldn't. It
> doesn't identify the statements themselves.

For an <item> that's exactly what it's intended to provide.

> > The RSS 1.0 spec doesn't mandate that an item's rdf:about be identical
> > to it's <link>, just that it should be if possible.
>
> Someone should compile a list of the recommendations in that spec that
> are OK to ignore.

That's not what's involved.  One has to appreciate the reasons RSS has been
used, over time, to grasp why the link and item about URL have been
interchangeable.  Since tools existed that expected a link to be present it
wasn't a good idea to take it out.  Many feeds, at the time, used the item
to say something about what the link pointed toward.  In those feeds there
was no expectation that the item would have it's own web page.  This was a
debate that raged for quite a while.  The emergence of easy weblogging tools
changed that.  But then half-assed RSS spec hijacking screwed up unique
identifiers.

-Bill Kearney
Syndic8.com

Reply via email to