The spec does not say whether or not clients can edit entries via the URI provided in the Location header.
GData currently uses the Location header to provide a read-only URI for the entry. IBM's impl and Joe Gregorio's impl, I believe, both return the edit URI in the Location header. Which approach is correct? Does the Location header have to return the Edit URI or not? I would rather that it does, as that means I don't have to parse through the returned entry to get the edit URI. However, the spec would need to be updated to reflect this. Opinions? - James
