From: Michael Everson (email@example.com)
Date: Fri Jul 08 2005 - 11:47:36 CDT
At 10:38 -0400 2005-07-08, Patrick Andries wrote:
>Michael Everson a écrit :
>>We're not going to do that. It would introduce
>>inconsistency in representation of Coptic text.
>Why do you say "we" when you express your
>opinion? How do you know what the WG2 will
>decide regarding a reference glyph change in the
>years to come?
Think what you like. I will stand with Stephen
Emmel and his Copticist colleagues. I do not
support the introduction of encoding ambiguity in
Coptic, and I will argue against it as I have
>>It may be the case that symbol is never found
>>without an abbreviation bar. That does not mean
>>that the abbreviation bar should be built into
>Well, you then expose yourself to some people
>encoding in their text the same symbol with and
>without the bar.
Not at all. If a person wishes to encode an
abbreviation for COIS, there are many to choose
from. He may choose to type C + O + BAR + S +
BAR, or he may choose to type CS + BAR, or he may
choose to type O + BAR + S + BAR. If he chooses
to type CS and leave the bar off, then he may not
be representing what is actually in his text.
It is foolishness -- yes, foolishness -- to
suggest that it makes sense to have *all*
abbreviations in Coptic which take abbreviation
bars to be encoded with a standalone letter plus
a BAR *except* for this one where the bar happens
to go over a standardized ligature symbol. We
have encoded that primitive symbol as what it is,
and we expect that in normal use it will be
followed by a BAR in the encoding. The encoding
model for Coptic does not pre-compose either
abbreviataion bars or the syllabic bars. And it
doesnt do so for the SHIMA SIMA SYMBOL either.
>I mean using two different Unicode character
>sequences (SHIMA SHIMA alone, and SHIMA SIMA
>with the standard contraction bar) to represent
>the same symbol.
We are not doing that. The SHIMA SIMA SYMBOL is
what it is. It may not be useful unless followed
by a BAR. In which case, the smart user will
follow it by a BAR.
>This is not a good idea, I would say, when this
>can simply be treated, as John suggested, as a
>glyph variant as is the case of the ordinal o
>(U+00BA) in Latin for example (may or may not
>have an underlining bar)
I have already explained why John's suggestion
should not be followed, but I will repeat it
again here. Coptic abbreviations using the
abbreviation mark are to be encoded by explicitly
using the abbreviation bar. If your font draws
the SYMBOL SHIMA SIMA with a precomposed bar,
then you will risk seeing two bars when you
receive *correctly* encoded Coptic text, say on a
Copticist's web site.
The analogy with U+00BA is false, as the underlining bar is purely decorative.
>>That isn't true. We knew perfectly well. SHIMA
>>SIMA may not be an obligatory ligature,
>So why even encode it : treat it as a normal
>contraction, the contraction bar can force the
>ligature depending on the font style chosen. I
>really thing this is where the problem stems
>from. But let bygones be bygones, the encoding
>cannot be undone.
It is not a "normal" contraction. It is a text
element. Coptic does not "normally" contract. We
encoded this text element the way we did *on
purpose* so that Coptic text could be represented
>>It may not "mean" anything without the bar. It
>>is nevertheless an element of the writing
>Well, by your construction. This ligature is in
>no way compulsory, not found anywhere else
>apparently but this common abbreviation and it
>is not a complete symbol : it lacks its bar on
>top as you admit yourself. When have we last
>encoded parts of symbols that need to be
>completed with an element found elsewhere ?
I don't know what theoretical question you are
chasing here. I worked with the Copticists, we
made decisions based on consistent representation
of Coptic text, and the result is what has been
encoded. If you wish to write COIS with an
overbar, write SHIMA SIMA SYMBOL and follow it
with an overbar, and you will get the result you
want. Every time.
What you three have proposed would introduce
inconsistency into the representation of Coptic
text, and that would be unwise.
-- Michael Everson * * Everson Typography * * http://www.evertype.com
This archive was generated by hypermail 2.1.5 : Fri Jul 08 2005 - 11:52:30 CDT