PMC Meeting Minutes
===================
2003-Oct-01 - 10:10am PT

Attendees
=========
Mark Davis           (IBM)
Deborah Goldsmith    (Apple)
Andy Heninger        (IBM)
Steven R Loomis      (IBM) - notestaker
Syn Wee Quek         (IBM)
George Rhoten        (IBM) - notestaker
Markus Scherer       (IBM)
Raghuram Viswanadha  (IBM)
Vladimir Weinstein   (IBM)
Tex Texin            (XenCraft) @ 10:25a
Doug Felt            (IBM) @ 10:30a

Agenda
======
Meeting started at 10:10am
ICU 2.6.1 is going well
Apple will use a future version of ICU. Probably 2.8 or 3.0.

The collation chapter wasn't reviewed. It needs to be reviewed.

Apple is not concerned about library loading and unloading.

Onto Apple topics:
Apple is interested adding more date/time formats to ICU
Mark says that OpenOffice has similar data, and we have no objection on adding it in the future.

Plugging in language support:
When it's safe to call u_cleanup, you can plug in new resource bundles.
	This will cleanup any caches.
	During logout, process reinitialization will also allow plugging in/changing language support.

Break iteration:
Mark states that service registration would be helpful for Apple break iteration.
The service stuff isn't persistent. So you would have to reregister it on load time.
On a similar note, the collation service registration API only works on the C++ API.
It is unsure if the collation service C++ API affects the C API.

Break iteration performance:
There are some performance issues.
Mark and Andy stated that ICU optimizes forward iteration. Backwards iteration is slow.
	Mark requested that Deborah find out what APIs were being used.

Thai break iteration:
Some discussion involving Apple and ICU Thai break iteration occurred.
ICU's Thai break iteration doesn't do full grammar analysis.
Mark states that the Thai break iteration may go into the base iteration.
	When certain characters are encountered.

Mark states that most of this stuff is already planned except the extra format strings.

Deborah states that iCal uses many formats (up to around 20).
Mark states that Open Office has a similar model.

Message formatting
Mark states that some people are complaining about quoting in message format. e.g. ''
Apple has no interest in seeing this quoting fixed.

Mark will give Deborah a copy of the list of features that will be going into ICU.
3.0 won't have a lot of features. It's mostly about bug fixing and performance.
3.0 will be a limited release. It will be on a shorter time scale.
3.0 might be released around May or April.
Mark states that the ICU core team can help with the design,
	and Apple can help with performance work and some implementation.
Andy will be working on "multiple tag values" in break iteration.
Mark would like to schedule another meeting with Apple about scheduling.

Mark summarizes meeting so far.

Deborah asked about proposing the script specific collation to the UTC.
Mark says that it sounds like a good plan, but there is a concern about
	synchronization with the collation ISO-14651 committee (JTC1/WG20).
There is some talk about moving the collation out of the WG20 committee
	and into the WG2 committee.
Mark suggests waiting

A Technical Discussion was had, about script code, locale ID and displayname. It is found here: 
   http://oss.software.ibm.com/cvs/icu/~checkout~/icuhtml/design/locale/20031001scriptcode.txt