RSS 2.0 Extensiblity

[clintonG]

I'm wondering if any subscribers would care to offer brief comments
-- pro *and* con -- regarding the fact that RSS 2.0 is now extensible,
notably by using namespaces.

<%= Clinton Gallagher
NET (e-mail address removed)
URL http://www.metromilwaukee.com/clintongallagher/

 

[Andy Dingley]

I'm wondering if any subscribers would care to offer brief comments
-- pro *and* con -- regarding the fact that RSS 2.0 is now extensible,
notably by using namespaces.

pro:

Extensibility - it's a good thing.


con:

The failure of all these 0.9* RSS versions has been that of poorly
written specifications. RSS 2.0 is barely any better. This is still
causing significant problems in interpretation.

I don't understand your "RSS 2.0 is now extensible" coment. It always
has been. It's arguable that some early 0.9* versions were too, as
they didn't explicitly rule out namespacing. As for extensibility,
RSS 1.0 is still the only one that does it right.

And the major one; RSS 2.0 is _not_ extensible.

RSS 2.0 is not extensible by using namespaced elements. All this
allows us to do is to add elements to an XML serialisation of RSS 2.0,
without losing validity.

It's arguable (because RSS 2.0 doesn't define a data model or parsing
model) that like SGML, the data model for RSS 2.0 _is_ the XML
serialisation. In this case we can claim that a well-formed XML
document is also a "valid" RSS 2.0 data model. What we can't do is to
see how to _interpret_ these additional elements.

Maybe we can infer the use of these elements. But we can't do that
unambiguously, or particularly reliably. RSS (like any protocol) isn't
a problem in publishing data, it's a problem in -communicating_ data.
Unless you can transmit something to the receiver in a way that may be
_understood_ by it, then you may as well not bother!

RSS 1.0 is still the best RSS version, especially for extensibility.
This is because RSS 1.0 is an RDF application, not just XML. This
implies use of the RDF parsing model and construction of RDF's data
model, and that brings with it rules that allow unambiguous parsing of
unexpected elements. It can't on its own imply the semantics of these
any more than plain XML does (although it provides a mechanism to do
so, if you wish), but the automatic handling of new elements will at
least be unambiguous and consistent.

 

[clintonG]

Thank you. Very well said. I've learned RDF RSS 1.0 is the better
horse in this race myself but facts are RSS 2.0 has mind share and has
been made to appeal to the simpletons and those who support
undermining progress for the sake of manufacturing commercial viability.

It is in the latter regard that is important to me as I find myself
compelled to register copyright for namespaces I have been using.
So for me and others I must presume, as document and text processing
methodologies continue to evolve as fundamental criteria imperative
to contemporary computing methodologies the need to seek copyright
or other protection is going to become increasingly evident.

Unless I am wrong -- which I need to learn one way or the other --
the discussions regarding copyright and namespace registries
were common circa 1999 and shortly thereafter but those discussion
seem to have dissapated and I need to determine why.

<%= Clinton Gallagher