L2/04-NNN

Date: May 26, 2004
Subject: Report on IRG #22
Source: John Jenkins


The 22nd meeting of the IRG was held in Chengdu, Sichuan, China from 24 
through 27 May 2004.  I attended for the US and for the UTC.  Hideki 
Hiura of Sun was unable to attend because of illness.  Representatives 
were present from the following IRG members:  China (PRC), Taiwan, Hong 
Kong SAR, Macao SAR, Japan, North Korea, South Korea, and Vietnam. 
Only Singapore had no representation at the meeting.  Michael Kung of 
Microsoft was also present as an individual contributor.

There are some important results from the meeting and some action items 
for the UTC and/or L2.  (This is done without the formal resolutions in 
front of me.  I'll submit them as a UTC document later.  I just wanted 
to get some of the juicier stuff out right away.)  The latter first:

1) Once again, the issue of the IRG maintaining a standing document 
giving its formal unification rules came up.  This has come up 
periodically, but nothing ever comes from it.  I think that we should 
ask WG2 to tell the IRG to create and maintain such a standing 
document.  It's ludicrous that we have no collective memory of 
unification precedents beyond what's in Annex S.

2) If we want to have a new CJK stroke block to house the stroke-like 
characters from HK SCS, we should have a document specifically about 
this issue and this issue alone.  Right now, it's buried within 
L2/04-161.

The IRG informally agreed that the stroke-like characters from HK SCS 
might best be encoded in a CJK Stroke block separate from the 
ideographs.  Members are to send feedback to Mr. Zhang in time for the 
upcoming WG2 meeting.  The IRG is willing to accept the removal of 
these characters from Extension C1.  There was also no objections 
voiced here with removing the actual ideographs from HK SCS and 
subjecting them to a "fast track" process, although if Extension C1 
will be ready by 2005 that may be moot.

The IRG also agreed to add two new characters from the most recent HK 
SCS.  Hong Kong issued the new HK SCS earlier this year, with sixteen 
new ideographs, fourteen of which can be mapped to existing ideographs. 
  The remaining two are now in the C1 bucket.

The position on Old Hanzi is pretty much as it was at IRG 21.  (Because 
there was only one of me, I wasn't able to attend all the break-out 
sessions and was therefore not involved in the Old Hanzi discussions.) 
The IRG continues to recommend that all the various Old Hanzi varieties 
be separately encoded.

IICore (the basic ideograph set) is now complete.  It's a bit slimmer; 
China removed 2000 characters, and it's down to about 9800.

The IRG is committed to having Extension C1 finished by the end of the 
year.

Mr. Zhang's latest term as rapporteur is ending.  The IRG recommends 
Dr. Lu Qin of Hong Kong as the new rapporteur.  (This decision was 
made, by the way, in a literally smoke-filled room.)  The clear 
preference was for Mr. Zhang to continue, but China was unwilling to 
support this.  (It's clear that there are some nasty internal Chinese 
politics going on here.) There was also the suggestion to ask SC2 for 
the creation of a new vice-rapporteurship to spread some of the burden 
out.  (The original intent was also to ease Dr. Lu into the world of 
WG2, but since Mr. Zhang had to withdraw as a candidate for rapporteur, 
that came not to pass.)  This didn't pass but will be discussed at the 
next IRG.

The discussion on who should be the next rapporteur was long and took 
up much of the last portions of the meeting, and because there was a 
group meal scheduled as soon as the meeting ended, things got 
enormously rushed.  As a result, I wasn't able to inform the IRG of the 
glyphs defect which we've found, nor to give my demo on variation 
selectors and Han.  Nor was I able to voice our objections to encoding 
seal forms as a separate script, since I wasn't able to get to the Old 
Hanzi sessions and the final resolutions were not individually 
approved.  I'll stay on top of these and submit documents for future 
IRG consideration.

The next IRG meeting will be in late November in Jeju, South Korea. 
The current provisional schedule is for meeting #24 to be hosted by 
Taiwan.  The IRG hopes that it will be possible for Unicode (or someone 
else) to host meeting #25 in the United States in the fall of 2005.

Dr. Lu, BTW, is working on creating Cantonese readings for everything 
in the BMP.  Where no authoritative source can be found, she'll be 
creating one (marked as such).  She's willing to donate this data to us 
for internal use, and for exposure in the on-line Unihan database.  She 
would rather not have it included in the text copy of the Unihan 
database, because she'd like to exert a bit more control over its use 
than that.  I've also talked over the problems of getting frequency 
data on HK SCS characters with Dr. Lu and other members of the HK 
delegation.  Dr. Lu has some data files she'll send me, and they had 
some people I can follow up with for further information.

========
John H. Jenkins
jenkins@apple.com
jhjenkins@mac.com
http://homepage.mac.com/jhjenkins/