ICU development team meeting
Time: 9am-10am PDT, 12am-1pm EDT
Date: 2000-oct-25

Attendees:
    Yves Arrouye (RealNames)
    Mark Davis (IBM)
    Steven Loomis (IBM)
    Markus Scherer (IBM)
    Helena Shih (IBM)
    Jim Snyder-Grant (Lotus)
    Raghuram Viswanadha (IBM)
    Vladimir Weinstein (IBM)

Minutes taker: Markus Scherer

Agenda items:

1. Users guide
  - Currently up for review, especially beginning from conversion chapters; ongoing work
  - Mark: pull contents from API doc into users guide, questions about what parts,
    how to sync api doc<->users guide, Mark: especially rules syntax,
    put links into api doc to users guide
  - Jim is personally happy about level of documentation :-)
  - Still need improvements
  - Mark: need real introduction chapter, use contents from homepage;
    Helena: has sent intro material to Roger about three weeks ago

2. Release milestones for ICU 1.7
  - On internal web site, will send out to public
  - Need to update after recent discussions with Lotus

3. IUC 18 ICU related abstract ideas
  - ICU collation implementation
  - Resource management, resource bundles (combine with ICU data item below?!)
  - Overview of ICU, Tutorial
  - ICU data: data packaging, locale data, charset data, aliases;
    public repositories & explorers/browsers
  - ICU from a system administrator's perspective;
    system & release management (part of the tutorial item above?!)
  - Comparison of ICU with other libraries
    (glibc (Unix98/POSIX), Qt, GlobalC, Win32/UniScribe, Rosette, ...)
    ->this should be a paper for now, maybe a talk for IUC 19

4. ICU workshop evaluation
  - Level too general/too detailed for some
  - Need more samples, more "hands-on" work & exercises
  - Most people would like to attend a second workshop when offered
  - Suggestion: "advanced-topic workshop" for specific aspects
  - Need to post material earlier & publish results on mailing list etc.
  - Prefer PDF & HTML format over PowerPoint file formats for presentations
  - Helena will post evaluation notes

5. Solicit help on verifying API javadoc and users guide accuracy
  - Send soliciting email to the list
  - Ask for comments to a certain email address (icu@jtcsv.com, multiple readers)

6. State of the collation implementation redesign, timing of changes
  - Preliminary version of design document, in internal review
  - Main goals: speed up performance & full compliance with UCA
  - Sort key generation partly rewritten in C,
    good performance improvement without normalization, little with normalization
  - Mark: need normalization if input not in NFD;
    rules could get a flag for the desired default of with/without normalization for these rules
  - Need improved normalizer implementation; performance! surrogates?
  - Lotus will measure performance improvements

7. Build issues
  - Changes in Unix builds, use lib folder instead of share for data
    (see proposal on mailing list), only really platform-independent files in share
  - Yves to write readme text about this