Identifying RSS-Like Formats

Thursday, 12 June 2003

I’m surprised by Dave Winer’s continuing reluctance to identify RSS 2.0 with a namespace, given how strongly he feels about interoperability and respecting format definitions.

If RSS 2.0 had a namespace, the conformance and semantics of a particular RSS document would be unambiguous - if you use the namespace, you follow the spec.

Versioning doesn’t buy you this, because there isn’t any management of the version namespace - anybody can say “that’s RSS 2.0”, and only people that recognize Dave as the owner of 2.0 will acknowledge that his is the spec to follow.

Of course, it would be even better if people didn’t use the term “RSS” to identify multiple formats. That situation brings problems with humans, but it doesn’t directly cause interoperability problems between computers.

As it is, RSS is a conceptual term, like “Wiki” is. When I go to a Wiki, I don’t expect it to necessarily have the same markup conventions as my own; it might be completely different technically, but it’s still a Wiki. In the same manner, RSS is a shared concept; it’s definately not a shared format.

2 Comments

Dave Winer said:

I made a proposal about that, apparently it didn’t go anywhere, but that’s not something you can blame me for Mark.

Monday, June 16 2003 at 4:08 AM

Danny said:

Mark - I’m not sure I entirely agree. There can be conformance and shared semantics without proper namespacing, in fact I’d guess that an awfully large proportion of data around the web falls into this category. RSS 2.0 is by general human (volunteered) consent stuff that’s like that of Dave’s spec. Things like this just aren’t as machine-readable/processable without the disambiguation. I do agree with your general thrust though - if it needs humans to do a lot of the work, it does defeat the point of computers somewhat…

Dave - if a person who writes, maintains and evangelizes a spec isn’t responsible for it, who is?

Thursday, June 19 2003 at 8:57 AM

mark nottingham