On 24 Mar 2012, at 17:36, James Leigh wrote: > The document makes it unclear how an agent should handle provenance for > NIR URI with a 200 response. > > If a :describedby relationship exists, should an agent assume that the > object of that relationship contains the same information the probe URI > response does?
The proposal purposefully doesn't state that, no; this is why the change proposal suggests creating a best practice guide for publishers and consumers, to flesh out the implications. I suspect that consumers won't want to make any assumptions and will just hoover up all the data that they can from wherever they can. I suspect that publishers will mostly want to provide just information about the probe URI at the probe URI, and more details about licensing/provenance at the description URIs. > Is it okay for a single agent to use the probe URI (with a 200 response) > as both an identifier for the documentation and an identifier for the > NIR it describes? That would not be a best practice, no. I think a best practice guide should tell consumers how to cope with the fact that publishers might. Cheers, Jeni -- Jeni Tennison http://www.jenitennison.com
