ICU development team meeting
Time: 10am-11am PST, 1pm-2pm EST
Date: 2001-feb-07

Attendees:
    Mark Davis (IBM)
    Kouichi Izumi (Lotus)
    Steven Loomis (University of Malta)
    George Rhoten (IBM)
    Markus Scherer (IBM)
    Jim Snyder-Grant (Lotus)
    Tex Texin (Progress)
    Raghuram Viswanadha (IBM)
    Vladimir Weinstein (IBM)

Minutes taker: Markus Scherer

Agenda items:

* Lotus items

- Short & long-term strategies for dealing with iterator bugs.
  (move break file generation tools to ICU4C vs. having clients check in
  fixed .dat files generated from Java vs. other options, etc.)

  Mark: Short-term use Java tools to generate the .dat files;
    long-term goal to look into porting these to C.
    Less important than other features.
  Steven: Mid-term, we could make a tool (script or Java) to run
    the Java tools and then build the libraries.
  Alan has last worked on the Java break iterator builder.


- Make sure we understand the overwrite flag [in collation rules], again,
  and what Lotus is going to use instead of the overwrite flag, since
  it is pretty useless
  (and perhaps diagnose how Lotus' previous understanding was so wrong...)

  Problem: Sometimes Lotus may need different collations from ICU's.
  ICU has a flag that specifies if the default rules should be included or not.
  Lotus had thought that this would force a variant to take precedence over
  a non-variant locale's collation elements.

  Vladimir: This flag is obsolete in the new collation design.

  Lotus special collation: German alternate collation -
  regular ICU could have a "German Phonebook" variant.
  In order to have Lotus rules override "normal" ICU rules,
  there could be different locale files or Lotus variants.


- Did Jamo normalization change between 1.6 & 1.7?
  (We have a crash - not sure if it's ICU or our integration of it)

  Cupertino team: We believe that normalization did not change since 1.6.


- Standard for collation tailorings is IBM? or Excel/Word?
  Both, by providing variants within ICU? Is there a stated ICU policy on this?

  Mark: The current data is from Taligent and IBM Toronto with applied bug fixes.
    IBM/Lotus could push to have tailoring data be collected in
    the Unicode consortium to have it reviewed industry-wide.
  Tex: This is a good idea and would be a big asset to the industry.
  Mark & Tex will write a proposal for this.


- How to communicate jitterbug priorities to ICU, especially for triaging stuff
  that will or won't be fixed in 1.8.
  (I think we'd like to schedule a Lotus/ICU meeting to review some open jitterbugs
  Lotus cares about & make sure we agree on when they'll get fixed by,
  and who will fix them -- how about I prepare a list & a rough proposal &
  appropriate team reps can meet to talk it over)

  Mark: Look at all open bugs. Start sending a list.
  Jim: Yes, will start with bugs opened by Lotus.


* Progress items

- Tex told me he wanted to talk about how "open"-source ICU is
  and how we can make it more so.

  Tex: Do not feel strongly about this.
    Phone conferences tend to be technical instead of steering.
    We should discuss how we get more people involved?
    Web pages (team members etc.) are outdated.
    We should have more strategic discussions.
    Perhaps add applications that are built on top of ICU?
    This should help make ICU more successful.
    Progress gets out of ICU now what it needs.
  Mark: We could have more emphasis on icuapps.
  George: The resource bundle editor could be a candidate for this.
  Markus: ustdio and iostream libraries could be done productively by
    people outside the core group.
    Steven had suggested that the Perl ICU group has problems with their CVS
    and could be interested in being hosted by ICU's system.
  Mark: Make it clearer how people can contribute to ICU, call for contributions.
  Tex: Agenda for next time: ideas for good opportunities for contributions.
  Mark: Small steps now, will have more time for this after ICU 1.8.


- Mark: Collation design document will be updated later this week.
  Mark/Jim: Awaiting feedback for collation rules for Japanese.
  Mark: The locale explorer needs to get updated to use the latest collation code.


- Next meeting: 2001-Feb-21
  Mark: We could use Yahoo calendars or similar.