mnot’s blog

Design depends largely on constraints.” — Charles Eames

Sunday, 22 June 2003


RSS History as state transfer

Mark Baker responded to my thoughts on RSS history a while back, and I’m finally getting around to responding (nothing like a hotel lobby to clear your thoughts…).

I agree very much with just about everthing Mark says; RSS feeds should be modeled as state transfer. The problem is that the approach he advocates implies the semantics of that state transfer, and RSS isn’t nearly robust enough of a format, nor its implementations interoperable enough, to count on implied semantics. That’s why I wanted to make them explicit (even if thereby redundant).

One Comment

Mark Baker said:

Yep, you’re right, there is a lot of inconsistency between RSS agents and specifications about the behaviour. I’d say that we really need to study the architecture of the RSS subsystem of the Web in order to get a better feel for what the best way forward is. I would expect it to be mostly URI/implicit based, but it’s possible, in cases where ambiguity is pervasive, that a mechanism such as yours could help out.

Monday, June 23 2003 at 7:37 AM