Based on non-online notes and the IRC log
1. Roll
2. Agenda review, and AOB See: http://lists.w3.org/Archives/Member/w3c-xml-protocol-wg/2005Apr/0013.html No additional AOB items 3. Approval of minutes: 30 March see: http://www.w3.org/2000/xp/Group/5/03/30-minutes.html Approved without objection 4. Review action items, see http://www.w3.org/2000/xp/Group/Admin/#pending. These action items are taken from the XMLP member page. 2005/03/30: Yves Update the home page to add a link to "Describing binary content in XML" spec Yves: Still Pending 2005/03/30: Yves Request publication of PER documents. Yves: Pending – once we resolve 31rec and 32rec, we can procede with the document transition 2005/03/30: Yves Close issue 29Rec with no action (and send closing email) Yves: Pending 2005/03/30: Yves Fix issue 28rec (add in errata) and fix also the broken rfc3023 anchor Yves: Pending 5. Status reports and misc XMLP/WSD Task Force - joint deliverable (Anish) Status on 'Describing Media Content of Binary Data in XML' – Anish Chair: Anish is not here; it seems that all of the changes for the media type doc have been done, and they're looking at a pub date of May 2nd. Anish: is there any objection from XMLP to publish? Chair: we probably need a week to review and provide comments. Anish: only comments we got from XMLP members have been resolved to their satisfaction; Jonathan said that he didn't think XMLP wanted a review. Chair: Vote - any pushback to publishing the document, given that we already had an action to review, and already provided comments? Yves: We have to look at REC32; it's about the namespace defined for the document (xmime). Anish: The only change is the prefix that was used; the URI is the same (xmlmime vs. xmime). Anish: We got so many comments back that we shouldn't use an 'xml' prefix, we changed it; but it's just a prefix. Yves: OK. I'd probably do the same in the SOAP primer, then. Chair: OK. Is there any other pushback? <<silence>> Chair: No objection; Working Group approves publication of WG Note about Media Types. RESOLUTION: Publication of Media Types WG Note approved. Chair: Any other work for the Task Force? Anish: After publication, I think it will end. 6. SOAP 1.2 PER specs & MTOM, XOP, RRSHB Recommendation maintenance Transition Status – Yves Yves: waiting to have clearance on REC31, also impact on examples in primer. Other than that, it's just a matter of finding time. 31rec: SOAP1.2 issue: http://lists.w3.org/Archives/Public/xmlp-comments/2005Apr/0000.html initial response: http://lists.w3.org/Archives/Member/w3c-xml-protocol-wg/2005Apr/0008.html Chair: Brought up by Mark Baker; Noah responded on the public list. Summary? Yves: http://www.w3.org/2000/xp/Group/xmlp-rec-issues.html#x31 Yves: http://lists.w3.org/Archives/Public/xml-dist-app/2005Apr/0002.html Noah: http://lists.w3.org/Archives/Public/xml-dist-app/2005Apr/0002.html (race condition...) that it cannot communicate the media type of the document encapsulated within the SOAP body. Mark Baker's comment is above This reminded me of discussions in the TAG; this is just my personal view (the TAG hasn't considered this specific issue). There are a few layers;
He refers to "the document" in the body; what I believe we have is an EII.
He implies that we have an obligation to be able to indicate that; we need to think about that obligation.
Media types refer to octet streams; we work in terms of infosets; the abstractions don't seem to line up, so it might not even be conceivable to give a media type to this thing.
My mail frames the decision tree as I understand it.
Yves: Possible relation to Compound Document Format WG. See: http://www.w3.org/2004/CDF/
Yves: They're in the process of defining how documents with multiple formats (e.g., HTML, SVG) work; they have to identify what it is. If they use mime type, we can base our solution on theirs.
Noah: I think you're going down one of the possible branches in the solution space.
Yves: I think that in part one, we say that SOAP is agnostic about the content of the body, so we don't have an obligation.
Noah: And mark is possibly saying that we have an obligation to meet an unspoken architectural requirement to define a standard, regular way of typing that subtree.
Noah: I'm not sure we do; one possible answer is "prove we have this obligation; users are free to use headers, and we choose not to standardise it."
Yves: Having Mark stating why he wants this would be a good thing.
Noah: He does propose a content-type header.
<env:Envelope xmlns="..."
<env:Header>
<foo:Content-Type mustUnderstand="true">application/xhtml+xml</foo:Content-Type>
</env:Header>
Yves: That's a solution, not a motivation.
Yves: For application/xhtml+xml, it will only be a fragment, not a complete document.
Noah: Even for a fragment, if I clipped a fragment of an octet stream that would be one thing. The SOAP body is not an XML document; it's an infoset or a sax stream, and we don't apply media types to a sax stream, even if it was a full document.
Anish: Does Mark intend for this SOAP header to be a binding/serialisation-specific sort of thing, or something that manifests in the Infoset?
Noah: I think he means "the soap envelope" which is a pre-binding thing.
Anish: What does the RFC for XHTML say? Is it always XML 1.0?
Noah: RFCs say that media types apply to octet streams.
Markn: I want to explore here.
Anish: In the media types note, the attribute can be only used on certain types.
Markn: http://www1.ietf.org/mail-archive/web/i-d-announce/current/msg04879.html
Markn: Better link: http://www.ietf.org/internet-drafts/draft-freed-media-type-reg-04.txt
Noah: It would be good to have a definitive word on this general issue.
Chair: Do we need to research more?
Markn: Yes.
ACTION: Mark N to research media type in XML issues.
32rec: XOP, Primer, DMCBDX
issue: http://lists.w3.org/Archives/Public/xmlp-comments/2005Apr/0001.html
http://www.w3.org/2000/xp/Group/xmlp-rec-issues.html#x32
Yves: I've just changed the examples in the primer.
Yves: If we agree to, I'll change the others as well.
Noah: The answer is "thank you, we accept, an errata will be published."
ACTION: Yves to close issue 32REC and respond to Mark.
7. AOB - None
Meeting Adjourned