As recently as the end of last month, Dave Winer publicly reasserted what he considers to be the RSS 2.0 spec [wikipedia edits]. Dave also recently completed a restoration of that specification.
Meanwhile the RSS Advisory Board promotes an alternate URI, and are in the midst of a protracted effort which intends to replace the content of that specification with this one.
The new specification draft purges the roadmap, and makes a number of changes, such as allowing skipHours to be zero to 24 (a change with some basis), and permitting namespaced attributes to be added to existing RSS 2.0 elements (a change without basis).
The latter was originally brought up by Randy Charles Morin, discussed (“strong argument” and “The spec, on the other hand, if read literally, is clear on the matter”), added to the draft, clarified, and hastily reversed.
Namespaces served as the basis for Dave Winer’s original fork, the basis for a nasty smear campaign, and the basis for one of Dave’s fears.
If you look closely, IE7 transforms feeds in exactly the way that Dave feared. It adds a completely undocumented cf:type attribute to description elements, something that Dave explicitly rejected, as he felt it would “add confusion”.
Meanwhile, the Advisory Board selectively promotes the use of Dublin Core.
As I pointed out last year, the reality is that just like there are two distinct RSS 0.91 specs (UserLand and NetScape (Advisory Board archive)), there are now two distinct RSS 2.0 specs (Harvard and Advisory Board).
P.S. In October of 2006, I asked that all references to my name be removed from this effort, and Rogers agreed. Over six months later, my name still remains in a document that was just published yesterday.