I've also begun updating the
feedvalidator to support
this. At the moment, I pass 28 of the 33 tests defined.
I plan to pass all of these, and keep up with changes to the spec
any additional tests as they are defined.
Based on comments from
Brent,
Dare,
Graham and perhaps others, it looks like the requirement for a
required unique identifier per item will be reinstated in
the next draft. Cool.
Another small tweak: nofollow links are now displayed with a preceding globe icon (just like MoinMoin does for external links). As this is accomplished with CSS selectors, it requires a standards compliant browser to be seen (sorry IE).
The RDF Validator requires that you check the box labeled "RDF is NOT enclosed in <RDF>...</RDF> tags" under the Advanced Features list in order to validate RSS 1.1 feeds and other rdf:RDF-less RDF/XML.
Micro-management of namespace declarations (must be declared on root). Not cool.
I would take all this RDF stuff more seriously if all semantically equivalent RDF/XML serialization were allowed (effectively forcing the use of an RDF processor or whatever is the correct term for Jena and the like).
Includes UTF-32 as one of the encodings. Not cool. Redundant. Impractical. XML processors only required to support UTF-8 and UTF-16. (No, C018 of Charmod does not require allowing UTF-32 on the wire.)
Tries to redefine XML rules when requiring explicit encoding declaration for non-ASCII. (UTF-8 and UTF-16 with BOM can go without as per XML.) Not cool.
Says “User-Agents MUST NOT continue to process feeds that are not well-formed XML” but goes on to contradict it: “User-Agents SHOULD continue to process feeds that are served with an incorrect encoding”. Not cool.
Explicitly bans “entity-encoded HTML”. Cool. The wording could be better.
HSF looks interesting, BTW. I still think one item per HTTP resource as XHTML would be even more Web-like.
As mentioned in few places including Sam Ruby, Miscoranda and Cafe con Leche. There is a draft version of RSS 1.1 available on a creative commons licence. The official specification is a bugfix rather than a real change. My own feelings are the...