Comments on Guidance - Partial Update

1. CLOSED Reconsider using SPARQL Update as the patch document and even if we don't choose this, consider providing SPARQL update equivalents for each OSLC Patch Document example.
   • Response using SPARQL update for now, if something better pops up we'll reopen
2. CLOSED partial update by command: what are the motivations? (IanGreen, 03/03/2010)
   • Proposed: I tried to explain the motivations up front
   • Response -Accepted- Closed by IanGreen 29-april-2010
3. CLOSED When adding a property, it recommends using POST. Is it expected to respond with a URI in the Location: header? I would expect could be a URI that builds the right URL to GET the resource with properties. I'm not certain we need to say anything about Location header, just wanted to raise it. (SteveSpeicher, 03/31/2010)
   • Response We're not doing creation, so I don't think we need to mention the HTTP Location header (DaveJohnson 05/21/2010)
4. CLOSED I am concerned about this use of QNames in http headers. If we want to POST a command object, I think the body of the request should be XML/JSON/etc and should describe everything about the command. (IanGreen, 29-april-2010).
   • Response: this is no longer in the spec (DaveJohnson 05/21/2010)
5. CLOSED Addressing the triple by suppling the subject, predicate and object requires that the client can take those elements from the representation and shift them to a different media type. Objects is the trickiest (for example, taking an RDF Literal from RDF/XML and putting it into the body of an http request. Encoding and escaping (for example, in the case of strings) needs to be accounted for, since the body is not an XML document (or at least, I don't read it as being an XML document). Subject and predicate pose fewer problems. (IanGreen, 29-april-2010).
   • Response: this is no longer in the spec (DaveJohnson 05/21/2010)