[XeTeX] Syriac abbreviations, and issues with polyglossia, fontspec and bidi

[David J. Perry]

----- Original Message ----- 
From: "Gareth Hughes" <gareth.hughes at garzo.co.uk>
> In Syriac, abbreviations and numerals can be marked by an overline. For
> abbreviations, the overline often extends over the last few letters of
> the abbreviation. This overline is supposed to be invoked by the Unicode
> character U+070F Syriac Abbreviation Mark = SAM, but I cannot get the
> Meltho fonts (specifically the Serto fonts are relevant here) to support
> it properly. I'm told that the Meltho fonts display the abbreviation
> mark correctly in MS Word.

Dear Gareth,

I don't think this is a font issue.  I don't know Syriac, but your question 
intrigued me since I am interested in OpenType and the various ways it 
handles language issues.  I downloaded Serto Jerusalem and it does indeed 
display the abbreviation marker correctly in Word (2007, running under Vista 
SP1).  I copied your example from the email and pasted into Word and also 
into Notepad -- that worked too.

I am not aware of any feature that could be put into a _font_ (at least an 
OT font--AAT is another matter) that would do what needs to be done in this 
case.  If I understand correctly, the convention is to print the line, with 
three dots, from the beginning of the abbreviation to the end of the word. 
Applications deal with things like finding the end of words (to determine 
line breaks, etc.); fonts do not, AFAIK.  There must be quite a bit of 
variability in terms of the length of the line.  That and the word end issue 
seem to me to call for the involvement of something at a higher level than 
the font.  The more I think about it, the more it seems to me that the ICU 
renderer that XeTeX uses may be at fault.  On Windows, Uniscribe handles 
such things; the fact that Syriac works in Notepad, which is a very basic 
editor with few capabilities of its own, suggests to me that it's the 
renderer that handles this.  Maybe someone more familiar with ICU than I can 
address this.

Just to be sure, I opened Serto Jerusalem in Fontlab and took a quick look 
around the OT tables.  I didn't see anything involving U+070F that looked 
like it would control the display of the abbreviation.  There is a high 
overline-ish charcter and a couple of different round bullets that might be 
used to construct an abbreviation marker, but that's all I saw.

Are you aware of anyone who has used the Syriac abbreviation marker 
successfully with XeTeX?  If so, what fonts have they used, and on Mac or 
Windows/Linux?

David