From: Chris Pratley (chrispr@exchange.microsoft.com)
Date: Wed Apr 16 2003 - 01:11:06 EDT
Peter,
We generally try not to have options that equate to "fail to correctly
interpret data I might receive from another user". If we have that, then
part of the world starts creating docs that other parts of the world
can't render or layout correctly, and we've spent the last 10 years
trying to get away from that.
I'd be interested to hear if people trying to use the PUA in Word have
actually had any issues with this Asian assumption in Word2002, or is
this mainly a problem in Word97/2000? I think in Word2002 we made a
change that allowed the semantics to be switched by formatting the PUA
text with non-Asian fonts.
Chris
-----Original Message-----
From: unicode-bounce@unicode.org [mailto:unicode-bounce@unicode.org] On
Behalf Of Peter_Constable@sil.org
Sent: Monday, April 14, 2003 10:00 PM
To: Michael (michka) Kaplan
Cc: unicode@unicode.org; unicode-bounce@unicode.org
Subject: Re: Variant Glyph Display
Michael Kaplan wrote on 04/14/2003 11:11:11 AM:
> Its always fun when people who were unaware of an implementation talk
about
> how a fundamental change in it would "not be difficult". :-)
Granted. I should have said, "I wouldn't think it should be
difficult..."
> Not speaking for Microsoft here, but offhand this strikes me as a
positively
> overingtonian idea that I would never recommend. Please forward
extensive
> discussion on this point to a PUA mailing list as it has nothing to do
with
> Unicode....
I think this is quite unlike overingtonian ideas of PUA usage: such
ideas
propose complex semantics for PUA characters and, more to the point,
suggest that there should be common understanding of those semantics, or
that there should be mechanisms (usually of a sort that doesn't use
commonly implemented protocols like XML) for interchanging information
about those semantics. What is happening here is that a particular
widely-used product assumes a semantics for certain PUA codepoints that
serve the needs of a specific (albeit significant) regional market, and
that I have suggested that it would be helpful to users in other regions
if
those assumptions could be overridden.
I continue to suspect that this would not be difficult to implement.
Clearly, the UI part of it is not difficult -- it's probably just a
checkbox in an options dialog. The issue is the algorithms for
controlling
various aspects of text processing -- aspects that are clearly sensitive
to
character ranges. At some point in those processes, a decision has to be
made as to whether these codepoints will be considered "Asian" (and so
have
the Asian generalisations applied to them) or not. It cannot be that
difficult to add one other condition to the tests in that decision
process:
if blnTreatEUDCPUAAsNonAsian and CharInEUDCPUARange then
CharBehaviourSlot = NONASIAN_NONCOMPLEX
else
(apply usual tests)
endif
(or something equivalent)
Also, you appear to be saying that this is relevant for some PUA mailing
list but is off-topic for this list. On the contrary, this is not a
discussion of individual suggestions for PUA characters -- the kind of
thing that is often discouraged from extensive discussion here -- but
rather is about general approaches in software implementations to
handling
the PUA range. That is certainly relevant here.
> Since this is an issue that you had never ever noticed until now, I
can't
> imagine that it has severely impacted you? :-)
Not yet, but given that the users I support are only beginning to make a
transition from their hundreds of custom legacy encodings to Unicode,
the
possibility of it impacting me (and these users) is still mostly a
future
one. The fact that I haven't been impacted severely up to now signifies
nothing; when I started reading in this thread about Asian-semantic
assumptions in Word for portions of the PUA, I became sincerely
concerned.
- Peter
------------------------------------------------------------------------
--- Peter Constable Non-Roman Script Initiative, SIL International 7500 W. Camp Wisdom Rd., Dallas, TX 75236, USA Tel: +1 972 708 7485
This archive was generated by hypermail 2.1.5 : Wed Apr 16 2003 - 02:05:29 EDT