On Nov 10, 2004, at 3:55 PM, Robert Sayre wrote:
http://www.intertwingly.net/wiki/pie/PaceExtendingAtom
I'm having a hard time parsing the requirements on markup appearing somewhere not mentioned by the spec. For example, markup from future versions of Atom could appear in atom:feed, but would be indistinguishable from foreign markup to an Atom 1.0 processor. The spec doesn't say anything about processor behavior when foreign markup is encountered in an "illegal" location. Could we say something like the following?
"When foreign markup is encountered in another location, software behavior is undefined."
I'm OK with that, but do note that the proposed section 9.2 says that unrecognized atom stuff may appear anywhere. But if I encounter a <link> element that's weirdly non-empty and contains markup from some other namespace, that's the kind of situation you're talking about. I think it would be OK to leave behavior undefined as you say.
It's ok to halt and catch fire, ok to do something with it, ok to ignore it, and ok to preserve it.
Also, would a signed feed require a DSig element in atom:feed?
Yep. But I think we already covered that, right? -T
