I'd be interested to know what actual benefit you get from this. "Single reliable data point" doesn't really mean anything to me. Could you come up with a real world use case that shows how a single reliable data point is essential, or at least immensely useful?
I can certainly see the problems it's going to cause me. When you give me a ref that's claiming to be an atom:id when it's really not an atom:id then I waste my time searching through my database looking for a parent message that's not going to be there. This is not the end of the world, but if I have to put up with that kind of pain I'd like to know that it's for a good cause. ;)
I'm sorry if this has all already been explained before. I kind of zoned out this thread when the arguments about extension attributes started again.
Regards James James M Snell wrote:
A concern was raised about having a single reliable data point available in all cases. The actual text in draft -11 says: If the resource being responded to does not have a persistent, universally unique identifier, an IRI used to retrieve a representation of the resource SHOULD be used so long as that IRI can also be used as a valid atom:id value and so long as the "href" attribute is also specified.
