From: Joan_Wardell@sil.org
Date: Tue Jul 29 2003 - 16:10:12 EDT
----- Forwarded by Joan Wardell/IntlAdmin/WCT on 07/29/2003 03:08 PM -----
John Hudson
<tiro@tiro.com> To: Joan_Wardell@sil.org
cc:
07/29/2003 02:09 Subject: Re: Your name came up!
PM
At 09:32 AM 7/29/2003, Joan_Wardell@sil.org wrote:
>Your name came up on the Unicode discussion list. Is this true? You would
>be happy
>with Hebrew as it stands with the current canonical order if we only added
>one meteg
>(presumably the right).
Ken quoted me out of context, but perhaps I was unclear. At one point, I
said that I didn't think a medial meteg character was necessary for
rendering, because the ligation can be handled with the left meteg.
Earlier, we were discussing various options for solving the re-ordering
problem and I said that *if* this problem could be resolved without
changing the combining classes (which I really don't think is going to
happen), a right meteg might still be desirable.
>You would be able to handle the examples below?
>shin < hataf < dagesh < regular meteg < shindot
Yes, this works fine. The font itself does not require strict ordering in
all cases; in fact, the font can be remarkably flexible in this regard
because of the different anchor attachment points defined (note that this
is true for OT; other font technoligies might requires stricter ordering).
There is, however, a subset of cases in which strict mark ordering is
required for correct display. These have not all been documented yet
>shin < hataf < dagesh < shindot <new right meteg
Obviously I have not made this yet, but it would work in the same way as
above.
>shin < hataf < dagesh < regular meteg < shindot <
>-where do I put the ZWNJ or the CGJ to get a left meteg on hataf?
Okay, that one is ugly. ZWNJ breaks the dagesh and/or shindot positioning,
depending on whether it is positioned immediately after the hataf vowel or
the dagesh. A painted CGJ character might not cause the same problem or,
rather, the problem might be resolved in glyph space with some fancy
footwork. Generally speaking, there isn't a lot that can't be done once
glyphs are painted, but obviously one doesn't want to over-complicate the
lookups in a font: there are significant performance costs to smart font
rendering, and it doesn't encourage font development for a script if things
get too complicated for the average type designer to wrap his head around.
Feel free to forward this to the Unicode list, if you think it will help
clarify things. The third case of normalised text you identify is a serious
problem for rendering, and having more of these sort of examples before the
eyes of the UTC members and other interested parties will help refine
possible solutions.
Regards, John
Tiro Typeworks www.tiro.com
Vancouver, BC tiro@tiro.com
The sight of James Cox from the BBC's World at One,
interviewing Robin Oakley, CNN's man in Europe,
surrounded by a scrum of furiously scribbling print
journalists will stand for some time as the apogee of
media cannibalism.
- Emma Brockes, at the EU summit
This archive was generated by hypermail 2.1.5 : Tue Jul 29 2003 - 16:57:44 EDT