Tuesday, February 1, 2005, 5:18:44 AM, you wrote:
> P.S. Feed Document may be somewhat misleading, because it's easy to > confuse it with Feed (which has connotations of the information > channel). I think "Feed Snapshot Document" or the like was once > proposed, but it was shot down. *shrug* In my RDF mapping <http://djpowell.net/atomrdf/0.1/> I modelled the thing with the id as a "Feed", and used "FeedInstance" as a term for the version of feed properties that is delivered in the document. Same for "Entry" and "EntryInstance". It is clear that feeds and entries are expected to change so I modelled the different instances as separate entities. This way you can use the same model whether you only care about the latest instance, or if you want to keep historic instances. [*] Anyway - would it be worth using the term "Entry Instance" in the spec? Would it make ids easier to explain? * You'd annotate each instance with something like atomrdf:receivedDate so that you can select the latest version. It works well from RDF - because you are no longer negating information every time something changes, instead you are just taking a selective view on the most recent version of what is available. -- Dave