From IRC log
1. Roll 11/9 BEA Systems, Mark Nottingham Canon, Herve Ruellan IBM, John Ibbotson IBM, Noah Mendelsohn (scribe) IBM, David Fallside IONA Technologies, Suresh Kodichath Oracle, Anish Karmarkar SAP AG, Gerd Hoelzing SeeBeyond, Pete Wenzel Sun Microsystems, Marc Hadley W3C, Yves Lafon Excused BEA Systems, David Orchard Canon, Jean-Jacques Moreau Oracle, Jeff Mischkinsky SAP AG, Volker Wiechers Sun Microsystems, Tony Graham Regrets Microsoft Corporation, Jeff Schlimmer Microsoft Corporation, Martin Gudgin Nokia, Michael Mahan 2. Reviewed agenda. [scribenm] Agenda is at: http://lists.w3.org/Archives/Member/w3c-xml-protocol-wg/2004May/0004.html [scribenm] Note: in minutes to follow, DF = our chair 3. Approval of minutes [scribenm] DF: I haven't received last weeks draft minutes [scribenm] Volker: coming shortly 4. Review of action items (details not recorded here) 5. Status reports [scribenm] MNot = Mark Nottingham [scribenm] MNot reports media type submitted to IETF [scribenm] MNot reports media type info sent to editors. done [scribenm] Editors report having incorporated issue 464 resolution. [scribenm] DF: is the rep header spec published? [scribenm] Yves: not sure, anish sent some notes, didn't read them. [scribenm] Yves: yes, rep header published as WD during call last week. Was confused, thought you were talking about edits to editors' copy [scribenm] MNot: soap+xml media type submitted to IETF. Informal feedback looks good. 2 week waiting period to publish as RFC. [scribenm] MNot: on xop/soap media type: we publish our rec first, then as IANA to put in registry later. [scribenm] DF: Last call, CR or REC? [scribenm] Mnot: rec, I think. Will check. [scribenm] Start discussion MTOM/XOP/Xinclude FAQ [Yves] http://www.w3.org/2000/xp/Group/4/05/FAQXOPInclude.html [scribenm] DF: now available as doc. We have not heard back from core. DF spoke to Arnaud le Hors of core, who thinks (informally) we'd probably be ok to publish. [scribenm] DF: we've ping'd core twice, let's publish. [scribenm] Yves: editorial work to go from ed copy to WD is not bad. See link above. [scribenm] Anish: we're working on proposing schema annotations that would indicate media type of base64 item (I think we're discussing task force with WSDL on media typing...right?) [scribenm] DF: We'll come back to this when we discuss issue 443 6. Attachments [scribenm] DF: we had agreed on items i - iv in the agenda, with v and vi to go. Noah raised an issue in last couple of days suggesting different resolution. [scribenm] Consider Noah issues first? [scribenm] Noah: makes sense to me. [scribenm] Noah summarizes problem: we have a normative schema for SOAP, and there's no way to write schemas for XML 1.1. e.g. schema for not understood fault returns QName. [scribenm] DF: is Noah's question the right one? If so, what to do. [scribenm] M. Hadley: what if we just say we're 1.0 until schema goes 1.1 [scribenm] Yves; if schema does this soon, maybe we go to 1.1 now and do an erratum [scribenm] Noah: schema's not working on it yet, and my own opinion is that it will prove difficult. I'm not sure I'd gamble on schema supporting XML 1.1 soon. [scribenm] DF: what about Marc Hadley's proposal? [scribenm] Noah: let's do this at the infoset level. [scribenm] Noah: we'll keep application/soap+xml as XML 1.0 only, but other bindings can use XML 1.1 encodings to get just the newline handling. [scribenm] Yves: alternate proposal...allow 1.1 in SOAP envelopes as long as you don't try to validate. [scribenm] Yves: clarification in answer to questions: the schema for envelope would be normative for 1.0 envelopes, not for 1.1 envelopes. [scribenm] Extended discussion. We seem to have two net proposals: [scribenm] Modified MarcHadley/Noah: for now, SOAP envelopes are XML 1.0 at the infoset level. application/soap+xml remains 1.0 only and remains the only mandatory http media type. Other bindings or media types MAY use XML 1.1 serialization to get the NEL char, for example, but not to enable new element names, char content etc. (at this time) [scribenm] Modified MH/N continued: leave tracks in the rec resolution about the difficult choice we made, perhaps contact the tag, and indicate that we can reconsider when schema goes 1.1 [scribenm] Yves: allow 1.1 envelopes now, but make the schema for envelopes normative only for 1.0 content. I think there was some discussion of trying to operate "in the spirit" of the schema when using 1.1 content and/or upgrading the rec prose to be a superset of the constraints in the schema. [scribenm] DF: seems a lot of support for MH/N, any objections? [scribenm] Yves: yes DF: "lie in the road" Yves: "yes, I guess so" [scribenm] DF: other opinions [scribenm] Suresh: strongly in favor of Infoset [scribenm] Noah: strongly favor infoset. Can live with Yves if we rewrite the rec to include all the constraints form the schema. Would "lie" down against Yves proposal if we didn't do that. [scribenm] Mnot: Mnot: I'm reluctant to rewrite the spec that way, and don't like informal use of schema. Therefore support infoset. [scribenm] DF: we're missing a few people today who tend to have strong opinions, but then again last call is looming. [scribenm] DF: ok, let's try a straw poll [scribenm] Hadley/Noah infoset propsal: strongly in favor 6, in favor 1, opposed 1 [scribenm] Yves schema normative for 1.0 only, redunant text in prose: strongly in favor 1, in favor 0, weakly opposed 2 [scribenm] DF: we will put this discussion out on distApp. Be ready to decide next week. [scribenm] Noah: note schema WG meets next week [scribenm] ACTION: chair, ensure entire WG knows there is a 1.0/1.1 decision scheduled for Wed. telcon. [scribenm] Herve: don't we have to do the same for XOP and MTOM? [scribenm] Noah: Herve's right. XML 1.1 is valuable in XOP and easy to do. [scribenm] Herve: yes, there is a schema for XOP, but we could easily make it non-normative and/or duplicate its few constraints in prose. [scribenm] DF: please resummarize [scribenm] Herve: XOP: there is a schema for XOP but not normative for XML 1.0. Go to schema and make sure we get all the constraints in the prose. [scribenm] DF: why OK here and not SOAP? [scribenm] Herve: it's smaller and not yet a rec. [scribenm] Noah: agree completely. I would note that we'll need some care in referencing in prose types like xsd:anyURI. Wonder whether a URI/IRI in an XML 1.1 document might allow chars that weren't allowed by xsd:anyURI? [scribenm] Anish: I was for infoset for SOAP 1.2, but I agree the situation with XOP is different. [scribenm] DF: Hearing no objection, we conclude formally that we will for XOP make the schema non-normative and update the prose to ensure that all constraints that were in the schema are reflected in the prose. [scribenm] ACTION: Herve to draft XOP prose to make schema non-normative and incorporate all schema constraints into the prose. Due COB Europe time Friday May 7 [scribenm] DF: what about MTOM? [scribenm] Herve: we might need only small changes if we go with Infoset approach in SOAP, so we should wait for that? [scribenm] Noah: what changes? [scribenm] Herve: editorial only, you're right, no architectural changes needed to XOP if we go infoset in SOAP [scribenm] Noah: agreed. [scribenm] DF: agreed, we'll wait a week. [scribenm] ACTION: MTOM Editors, starting 5/12, reflect SOAP 1.2 XML 1.1 decision (infoset?) in MTOM. May be only editorial cleanup. [scribenm] ACTION: Noah to track down lexical form rules for xsd:anyURI. In particular, is it restricted to XML 1.0 characters? [scribenm] Time check? [scribenm] Begin quick discussion of issue 443: [scribenm] http://www.w3.org/2000/xp/Group/xmlp-issues.html#x443 [scribenm] DF: Anish, did our discussion of XML 1.0/XML 1.1 affect perspectives on issue 443? [scribenm] Anish: no, don't think so. [scribenm] Anish: issue raises three requirements, which task force has agreed on means of addressing. [scribenm] We could agree issue closed by media type document, and we can bring issue up again if media type changes and issue becomes unaddressed. I (Anish) supports this. [scribenm] Alternative: keep open until published as note. [scribenm] ACTION: Anish, (a) as editor of media type doc, check that media type attribute text in XOP is the same as corresponding text in media type doc (b) propose alternate text for XOP to appropriately reference media type doc. By COB May 7. [Zakim] WS_XMLP()11:30AM has ended [Zakim] Attendees were Yves, +1.503.830.aaaa, Herve, Mark_Nottingham, Marc, John_Ibbotson, Noah, +1.781.902.aabb, +1.503.830.aacc [RRSAgent] I see 5 open action items: [RRSAgent] ACTION: chair, ensure entire WG knows there is a 1.0/1.1 decision scheduled for Wed. telcon. [1] [RRSAgent] ACTION: Herve to draft XOP prose to make schema non-normative and incorporate all schema constraints into the prose. Due COB Europe time Friday May 7 [2] [RRSAgent] ACTION: MTOM Editors, starting 5/12, reflect SOAP 1.2 XML 1.1 decision (infoset?) in MTOM. May be only editorial cleanup. [3] [RRSAgent] ACTION: Noah to track down lexical form rules for xsd:anyURI. In particular, is it restricted to XML 1.0 characters? [4] [RRSAgent] ACTION: Anish, (a) as editor of media type doc, check that media type attribute text in XOP is the same as corresponding text in media type doc (b) propose alternate text for XOP to appropriately reference media type doc. By COB May 7. [6]