Minutes for RELAX NG TC 2001-06-07

[Michael Fitzgerald <mike@wyeast.net>]

Minutes for RELAX NG TC 2001-06-07 at 8:30 AM EDT (UTC -05:00)

[Please post amendments and corrections to the list rather than just to
me. -Mike]

Attendees

James Clark (chair)
Norm Walsh
Josh Lubell
Kawaguchi-san
Murata-san
Mike Fitzgerald

Absent

Mike Smith
Fabio A.
David Webber

Next meeting: Thursday, June 14, 2001, 8:30 AM EDT (UTC -05:00)

1. QNames. Three solutions were proposed.

  1. Keep the status quo QNames for attribute and element names
  2. Use div element to declare a namespace plus special casing of xml:
  3. neither (Murata-san)

The TC voted in favor to retain the status quo and go with 1, but considered
#2 as a separate issue.

1a. Add div element (#2) in addition to using QNames.

An argument in favor of this was Norm's compelling use case of representing
parameter entity wrappers with a div element. It also allows for future
experimentation, etc.

TC voted in favor of the div element, but decided to take offline a
discussion of whether or how the div element will be treated as a child of
include.

2. Global attribute value.

James prefers not to have a boolean value for the global attribute to allow
for a third possible value. Murata-san asked, "Why do we need the global
attribute?" At Kawaguchi-san's suggestion, we deferred the issue to future
discussion so that we could work on issues we could decide now.

3. datatypeLibrary instead of datatypeNamespace, retain then ns/nsName
instead of namespace/namespaceName.

TC voted in favor datatypeLibrary and returning to ns/nsName (a reversal of
an earlier decision).

4. Other delimiters beside whitespace for list.

TC decided to defer on this issue [after 1.0?].

5. Should a redefinition without an original be an error?

TC reached consensus that such a redefinition should be an error, in spite
of a possible use case presented by Murata-san involving multiple includes.

6. Should multiple definitions in the same file be allowed?

TC decided to allow this. Murata-san's comment regarding combine="choice"
was that such multiple definitions would not then be construed to be
mistakes.

7. Should nested grammars be prohibited?

Murata-san withdrew his objection. TC consensus is that nested grammars are
OK.

8. key/keyRef and modularity: is anybody else concerned about this, or
should we forget about it for 1.0?

James has a gnawing feeling that this will become a problem later. TC
decided to publish a public draft and wait for comment on whether this will
be an issue that we must address in 1.0. Murata-san also explained that a
WebDB folks expressed a desire or need for multipart keys, which we don't
plan for in 1.0.

9. Publishing the tutorial. Can we agree to change the status section of the
tutorial and publish it as representing the current committee consensus?

All TC members were polled and agreed to publish the draft next week. Here
is the plan: James is to update the tutorial and post on Friday, June 8. TC
members may comment until close of business (US) Monday, June 11. James will
incorporate fixes to any showstopper bugs and publish a public draft of the
tutorial on Tuesday, June 12. [Do I assume correctly that only essential
changes will be made and that subjective edits may be set aside until
later?]

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