Henri Sivonen wrote:
Are you suggesting Mozilla is wrong in supporting xml:base in a generic way (including in XHTML)?
I don't care about Mozilla (right now). We have xml:base, and we know it's not going to work quite right, all of the time. Either we warn people, or we don't. This may signal a deeper problem with xml:base, or XML itself.
If Atom Processors are to take the Feedparser approach, they shouldn't be used as endpoints for the protocol, because you don't want the server changing your nice, short relative references into full URIs.
I'm not sure any spec language will help. Maybe 'use absolute URIs for maximum data integrity', but I suppose that is always true.
Robert Sayre
