I thought I'd get the ball rolling by providing some general comments (i.e. not specific to X-KRSS or X-KISS) that I have on the XKMS draft. i) The "pending" result code will be troublesome unless there is an indication for how the client should react to such a result code, e.g. Do they wait 5 mins, 10 mins, 24 hours, before sending in a new request? It seems likely that any follow-up request should refer to the original request as well. I believe that similar issues came up with a similar result code in CMP. I suppose that there are a few options - remove this result code - add a new request message the allows the user to "ping" the status of an expected response (similar to what is done for the proposed bulk operations) - rely on the underlying transport protocol (I'm least familiar with how this option would work but it is a workaround that is proposed for CMP) ii) There is no clear guidance as to which <Respond> elements (given in Section 3.1.5) are applicable to which operations. For example, "Multiple" doesn't seem to make sense for an X-KRSS request so that there should be some general guidance indicating that an X-KRSS server would ignore "Multiple" if included in the <Respond> element of an X-KRSS request. (If people agree this is important, I can write some text related to this.) In addition, there's no schema for the list of choices for <Respond>. Would it not make sense to define as a Choice element as opposed to just listing the possible values within the document itself? Cheers, MikeReceived on Wednesday, 31 October 2001 09:53:19 EST
This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 13:51:40 EDT