[MPEG-OTSPEC] lookup processing and 'chws', 'vchw'

[John Hudson]

On 20082020 12:43 am, Peter Constable wrote:
> I guess there’s another question I can ask: When using GPOS type 2 
> lookups for ‘kern’ (or ‘vkrn’, ‘chws’, ‘vchw’), do font developers / 
> font development tools usually make adjustments to the first glyph only?

Depends on text direction. LTR kern lookups typically adjust width of 
first glyph; RTL kern lookups adjust width and position of second glyph. 
If you don't handle RTL kerning this way, every-other pair in the run 
gets skipped during kern processing.

[Related to this: I have an open mental issue about incompatibility 
between HarfBuzz and Microsoft handling of directionality as it applies 
to bidi text. My understanding is that HarfBuzz processes e.g. all 
Arabic text as RTL, even embedded LTR character strings such as numbers, 
and hence doesn't segment glyph processing runs based on direction, but 
Microsoft puts bidi RTL and LTR sequences into different runs. Either 
model works, but I am concerned about what the inconsistency means for 
kerning between Arabic numerals, which I treat as LTR, but which 
Harfbuzz may be treating as RTL? This was discussed at one of the OTWG 
meetings a few years ago, but I never got it entirely sorted in my head.]

J.


-- 

John Hudson
Tiro Typeworks Ltd    www.tiro.com
Salish Sea, BC        tiro at tiro.com

NOTE: In the interests of productivity, I am currently
dealing with email on only two days per week, usually
Monday and Thursday unless this schedule is disrupted
by travel. If you need to contact me urgently, please
use some other method of communication. Thank you.