See also: IRC log
pauld: minutes from our last meeting approved
pauld: asks people to raise new issues and complete AIs to help build agenda for next week
pauld: We have a number of people registered to observe, hopefully including some members of the XML Schema WG. Challenged on how to build an agenda around our weak pariticpation, with remote participants in Sri Lanka and West Coast, make best use of observers and deliver a working draft soon after the meeting!
pauld: several new issues, let's line them up!
ISSUE-13, ISSUE-4 surround PSVI. Are they the same issue? default value requires schema for processing.
ISSUE-15 null datatype, use-cases beyond ASN.1
jonc: is there a difference
between patterns on the wire? does Ed want us to anoint one of
... databinding tools may represent these differently
tony: Ed's use case is optimal saves sending a byte or three
ISSUE-16 will lead to the difficulty of quantifying how good a particular pattern is
ISSUE-17 duration type
Tony: can help by providing information on HL7 interval type
pauld: duration, echo of month
comes back as 30 days ..
... this may lead to enumerate types which work well, and not include duration!
jonc: does the issues list become something we reference ?
pauld: Yves has built a list of issues for databinding into our draft
tony: describes interval datatype .. posted to the list : http://lists.w3.org/Archives/Public/public-xsd-databinding/2006Feb/0027.html
<Yves> the document Paul referred to is http://www.w3.org/2002/ws/databinding/edcopy/basic/basic.html
<scribe> continues to introduce new issues
ISSUE-18 answered by Sandy on the list. However for databinding venetian does seem most natural. Should we Recommend only named 'types' for Basic patterns? That would negate a lot of public schemas.
ISSUE-19 all model group. Sandy answered this too. Schema 1.1 may have some answers, but we're concentrating on 1.0
ISSUE-20 may become an Advanced pattern given our experience with tools which bail on any.
tony: HL7 uses xs:any, but does
not appear in a 'final' document
... our reference information model 6 or so classes represent health care, different health groups specialises the anyType to their specialisation
pauld: use-case is for versioning and evolution
tony: describes HL7 datatype
versioning, adding types is usually safe, but modifying types
is less likely - we communicate breaking changes in
... we don't have a lot of experience with tools, since many choke on our valid complex schemas
... will contribute HL7 to working group members (not publicly available at present as between publishing cycles)
pauld: Zakim tells us we're out of time - see some of you next week F2F!