On 2006-09-06 02:25:47 -0400, Bob Wyman wrote: > Because of the nature of the tool being used here, a > hyperlink, we must accept that the binding between content and > license is a weak and fragile one. There is no guarantee that the > content of the license associated with a feed at one instance > will be the same as it is at some other instance. Also, there is > no mechanism provided to ensure that claims made about what the > license was at one instance can be proved at a later time.
I think this assessment mixes two aspects: The use of the rel=license link as an *identifier* for a specific license (which is pretty much what Creative Commons does), and the use of the link to retrieve license information. The value that this specification adds to Atom is mostly in the "link-as-identifier" field: It enables software to offer, e.g., retrieval of feeds (or entries) by license. Look at the Creative Commons powered aspects of flickr search for a related example. Trying to dilute the value of the license link out of fear that the link-as-retrieval-mechanism aspect is dangerous, however, impacts the value of the link-as-identifier aspect. As you point out, the weakness of the link between content and a specific license (in the sense of legal code) causes problems for both the copyright holder and the relying party. A corollary of this is that, in fact, both sides have an incentive to use licenses that are identified by a stable URI. Hence, I'd suggest that, in making balancing decisions for this specification, emphasis be put on using URIs as *identifiers* for licenses. It's fine to point out the lack of an enforceable binding on a technical level, but I don't think this spec is the place to discuss the legal implications that this might have. -- Thomas Roessler <[EMAIL PROTECTED]>