XLIFF Teleconference Minutes - Tuesday, 28 Jan
2003
1. Attendance:
Tony Jewtushenko, Oracle Corporation, TC
Chair
Yves Savourel, TC Editor
G�rard Cattin des Bois, Microsoft
Doug
Domeny
Mirek Driml, Moravia-IT
Milan Kar�sek, Moravia-IT
Christian
Lieske, SAP
Mat Lovatt, Oracle
Enda McDonnell
David Pooley, SDL
John
Reid, Novell
Bryan Schnabel, Tektronix
Shigemichi Yazawa
Apologies
Peter Reynolds, Bowne Global Solutions, TC Secretary
Mark Levins, IBM/Lotus
The following have lapsed TC membership
John Corrigan, Sun Microsystems
Ian Dunlop, Novell
Reinhard Sch�ler,
Localisation Research Centre
Enda moved to accept, David 2nd
1-- Issue 6 is the only open
issue: There are 6 options previously discussed. 1-4 based on Mat's
proposals, 5 Doug's proposal, 6 Yves's proposal to defer to a future
release. Option 5 was discussed. Shigemichi and G�rard discussed
compatibility issues for 1.0 parsers and the ramifications of the change: 1.0
only had "yes" or "no" with a default of "yes". Thus, the parser may only look
for "no" values. It would not discover those of "font". Mat then suggested that
the list of values for reformat be left open using the "x-" mechanism.
However, the opinion that any extended attribute from another
namespace wass that namespaces responsibility. Enda also pointed out that the
extended value would have the form "x-namespace:attribute", which is rather
cumbersome. Mat conceded the point. Tony called for a vote for the proposals
(Option 5 was only for the mechanism and not of the actual values), Mat 2nd.
Each member chose their preferred option in a role-call vote. The results are:
option 4, 1 vote; option 5, 11 votes; 1 abstain.
2- The values for reformat was then discussed. Doug read the values
proposed in his proposal (option 5). There was a discussion about adding color
but it was generally felt that this could be too problematic. There was a
discusson about css-style and if that didn't extend the values. It was
pointed out that since this only says the css-style attribute can be reformated
and does not specify what in that attribute can be reformatted, it was not an
extension of reformat values. Tony called the role-call vote and the values as
proposed by Doug was approved unanimously.
3- Mat introduced a new issue based on an earlier email: Since the
resname attribute can have whitespace it should not be a nmtoken as it is in the
schema/DTD. There was some discussion about whether this nneeded to be decided
by the TC. It was decided that this was a bug of the schema/DTD and should be
corrected without any action of the TC.
ACTION ITEM: Editor to modify the spec accordingly.
4/Schedule of remaining work
Release Candidate on Feb. 5
Allow two meetings
discussion of spec changes.
Feb 19 target date
5/ Carried over from last week:
Discussion and response to recent xliff-comment regarding resizing of RC type
data within XLIFF - http://lists.oasis-open.org/archives/xliff-comment/200212/msg00000.html . This query relates to Mat's "reformat" issue
(issue # 6). Do we need to communicate the
solution per Issue 6 with Mr. Tsepalov?
There was some discussion between Mr. Tsepalov and Yves. Since Mr.
Tsepalov has not responded recently it is assumed no action neds to be taken by
the TC.
6/Any New Business
1- John brought up the discussion between himself and Yves
concerning the origin, tu-state, and state attributes.
ACTION ITEM: John will propose a solution for this.
2- Yves brought up the incompatibility with 1.0 of forcing an
order on the elements, as was done in 1.1. John proposed that we rescind the
decision and recommend in the sppec the order specified. Yves 2nd. Role-call
vote was unanimous approval of the proposal.
ACTION ITEM: Editor to modify the spec accordingly.
3- Doug pointed out that the latest version of the schema has
an encoding problem.
ACTION ITEM: Yves will investigate.
7/Next meeting - Tues, 4
Feb 2003, 04:00PM Europe/London/Dublin / 08:00 AM
PST