> > There's little to differentiate between these two formats on
> that level.
>
> All the RSS and Atom files seem to attempt to present the same
> data in slightly different dialects of the same language.

It's an arguable point, but I think I'd have to disagree - the differences
are such that although for the most part it may be the same data, there are
actually 3 very different languages: RSS 2.0, Atom, RDF. The first is a
classic stovepipe XML application language. The second is a more
sophisticated stovepipe app in that it is likely to offer generally
applicable extension mechanisms (likely to be mappable directly to a subset
of RDF). The third is a general purpose web-oriented language (only a subset
of the syntax is allowed in RSS 1.0, but RDF is the language).

> This nicely exposes some of the key issues in
> XML data interchange:
>
> 1) using XML simplifies token parsing, but doesn't help much
>    when processing grammars. As for being able to convey message
>    semantics ...

XML hardly gets a look-in with RSS - see recent discussions on liberal
parsing.

> 2) MovableType now includes 4 different templates for syndication:
>    is Stove-piping really the best way to version XML languages :-(

At this point in time, in this context, it may well be.

> 3) something which is quintessentially pub-sub ends up using
>    polling.

I'm not at all sure about that - polling is very much a feature of the
system.

> OTOH the massive success of RSS does demonstrate that simple
> approaches often work best. 'Good-enough' has been the story
> of the Web.

Rather it demonstrates that simple approaches with lots of hype can gain
widespread adoption. 'Good enough' begs the question, good enough for what?
If all you want is another kind of HTML browser, RSS is certainly good
enough.

Cheers,
Danny.