On 24/09/2009, at 7:37 PM, Ian Hickson wrote:
It would certainly be an improvement, though I'm still unconvinced that wewant different definitions for the same link type for different MIME types. Difference conformance classes, sure, but that's another issue.
Yes, that's the tricky part; there would be a tendency to define different meanings for different mime types, no matter how clearly it's stated that the additional information was restricted to the interpretation/implementation of the relations.
-- Mark Nottingham http://www.mnot.net/