OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

relax-ng message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]

Subject: Minutes of RELAX NG TC 2001-08-23

Minutes for the RELAX NG TC telcon on Thursday, August 23, 2001 at 10:30 am
EDT (UTC -04:00)





Before the meeting, Josh remarked that Makoto's presentation on RELAX NG at
Extreme Languages (August 17) went very well and that Makoto was a bit
modest about it. Makoto feels that XML Schema is favored by those working
with relational databases.

1. Do we want to use datatypes instead of an annotation to handle
ID/IDREF/IDREFS? (Following issues assume a "yes" answer to this.)

Makoto: I prefer datatypes.
James: Does anyone disagree?
All: [silence]

2. Do we use the XML Schema Part 2 datatype library or do we have our own
compatibility library? If the latter, what about xsd:ID/IDREF/IDREFS?

Norm: users will expect a certain behavior...
James: if we go with xsd:ID/IDREF/IDREFS then the implementation should have
the same functionality as XML Schema Part 2 (XSP2)...
Makoto: I'd like to see the same semantics as DTD. Do we really need to
force them to implement XML Schema Part 2 just to get ID/IDREF/IDREFS
James: I have two concerns: (1) my gut feeling is that a lack of modularity
will not be clean; and (2) it is hard to describe crisply what the
compatability spec is doing.
Makoto: 2 sets of ID/IDREF/IDREFS?
James: one could alias the other...
Makoto: they wouldn't have to do all XSP2 if they don't like it...
Josh: this is a naive question, but who doesn't like XSP2?
James: all kinds of people...they got decimal wrong, etc. I prefer to avoid
ties to XSP2.
Norm: the compelling argument would be that this would allow them to have a
few more datatypes than ID/IDREF/IDREFS if they wanted...
Mike: to clarify, we are dropping a:attrbuteType altogether?
James: Yes.

The decision was to create a spec for our own compatability library.

3. Does this go in our annotations spec?  If so, what do we call the
annotations spec?

Mike: I prefer that we just drop 'annotations'
Makoto: So do I.
James: Two namespaces then? http://relaxng.org/ns/compatibility/annotations
Others: [yes]

Agreed to call it the DTD compatability spec, and drop the 'annotations'
from the title.

4. What should "ID-compatibility" involve (ie what constraints should the
compatibility spec impose on the use of ID/IDREF datatypes)?

James: we should keep it simple
[James' explanation]
James: I'll write something up
Mike: maybe we should give it a week
James: we have to stop giving things a week
Makoto: I like the separation of ID-soundness and validation
James: should we have ID or other things? infoset modification comformance
Makoto: I don't see a lot of benefits to this.
James: I use cross-references with XSLT in DocBook...you have to declare the
IDs to make it work. I think the benefits of ID are substantial.
Mike: I don't think we can avoid doing this.

Agreed that we impose a conformance level constraint for ID-compatibility.

5. What about infoset modification?  Do we want a conformance level in the
compatibility spec that involves changing the infoset?

6. If yes to 5, do we also want all the other XML 1.0 datatypes along with
ID/IDREF/IDREFS? If so, what constraints does the compatibility spec impose
on their use in the schema?

Norm: I think its all or nothing...11 levels of conformance?
Josh: What about #FIXED?
James: It's a weird beast.
James: If we do all 11 levels...some people won't care about some of them
[NMTOKEN, for example]...we could provide aliases...people expect type
assignment in XSP2 anyway...
Mike: Maybe we ought to just note it in the spec?
James: Okay, we'll put it in the spec and ask people for feedback

James: here are my marching orders:
- we will have a compatibility spec
- we will use two different URIs [see 3]
- conformance levels, including one for infoset modification
- a XSP2 implementation spec, ID/IDREF/IDREFS
- aliases

James: all in favor, say aay...
All: aay

>From James, notes after I left at about 20 past the hour:

- Murata-san said that Lauren Wood had invited himself and James to lead a
RELAX NG group in a schema comparison panel to be held in conjunction with
XML 2001 (other groups being W3C XML Schema, Schematron and DTDs)

- James said he was working on a program to convert DTDs into RELAX NG
(retaining parameter entities)

Wy'east Communications     http://www.wyeast.net     mailto:mike@wyeast.net

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]

Powered by eList eXpress LLC