- From: Jacek Kopecky <jacek@idoox.com>
- Date: Wed, 3 Oct 2001 15:41:02 +0200 (CEST)
- To: "Williams, Stuart" <skw@hplb.hpl.hp.com>
- cc: David Fallside <fallside@us.ibm.com>, <xml-dist-app@w3.org>
Yes, my text was more a primer-speak (I think), your text is more spec-speak. 8-) The information about Actor URIs seems to be scattered now, both sections 2.2 and 4.2.2 contain some of it, maybe some consolidation is necessary. Other than that, I tend to prefer putting the actor-choosing discussion note in the primer and not in the spec, as it basically says "you can do anything you will". I don't think such a section belongs to the spec which is mant to set rules. Let's see what the telcon brings. 8-) Jacek Kopecky Idoox http://www.idoox.com/ On Wed, 3 Oct 2001, Williams, Stuart wrote: > Hi Jacek, > > > -----Original Message----- > > From: Jacek Kopecky [mailto:jacek@idoox.com] > > Sent: 03 October 2001 10:55 > > To: Williams, Stuart > > Cc: David Fallside; xml-dist-app@w3.org > > Subject: RE: Issue 140 bogus? > > > > > > Stuart, > > so it seems that to resolve your issue #140 you'd like to see > > some informative discussion on the bases for determining the > ^^^^^basis > > Actor URI set of the node for this message, right? > > Just about right! I'm less interested in enumerating the set of Actor URIs > than I am in discussion of the possible basis upon which a SOAP Node decides > that it performs the role of a particular actor with respect to a given > message. Enumerating the set and testing for set membership certainly would > do. > > > I may try to propose a first draft of such a discussion. Below > > is what I would say taking into account my SOAP building > > experience: > > > > ------- begin > > The set of Actor URIs that the node assumes for processing a > > message can come from various sources: > > > > - the Specification: ".../next" is always in the set > > - static configuration for a combination of the endpoint URL, > > SOAPAction URI (when applicable), even first Body child's qname > > or an other part of the message. > > - dynamic configuration based on some (as yet unknown) extension > > whose SOAP block would carry the necessary information. > > > > This set could include the empty Actor URI which would mean that > > this node is the final receiver of the message. > > ------- end > > > > This is a very first rough draft of what I think might satisfy > > issue #140. Stuart, others, is this a good proposal? 8-) > > I've a slightly different suggestion, but I think the spirit is the same. > Something like the following at the end of Section 4.2.2 in Part 2 would > work for me. This may need a little work by the editors, the first two items > tersely restate what is in the 3rd to last and 2nd to last para of the > current 4.2.2. The last item is the informative item which I think would > cover what I think is missing. Stylistically the MAY may not be the right > way to 'tack' this on to the list... > > ---being > In determing whether a SOAP Node performs the role of a particular actor > with respect to SOAP message that is being processed, a SOAP Node: > > - MUST always performs the role of the ".../next" actor. > - MUST never perform the role of the "../none" actor. > - MAY make a determination based upon such factors as: > local configuration information; > the receiving transport endpoint address; > the message content (covers dynamic content and 1st body > child); > any other implemenation dependent factors; > ---end > > > > > As for where to put it, I think that as a non-normative > > discussion it could fit very well into the primer. 8-) > > The primer might want to expand on it by example. Personally I continue to > think that the spec should offer something. Part 4.2.2 seems like thre right > place to me, but I'll go with the flow. > > > > > It's on the agenda today, so we can propose some draft resolution > > during the telcon. 8-) > > > > Jacek Kopecky > > > > Idoox > > http://www.idoox.com/ > > Regards > > Stuart >
Received on Wednesday, 3 October 2001 09:41:07 UTC