[mpeg-OTspec] RE: [OpenType] Kerning

[Sairus Patel]

Indeed, the kern table format is too limited to be able to do things like skip marks -- at most it would be able to kern two bases in a base-base-mark sequence. You are right that it would not be able to kern the bases in a base-mark-base sequence.

Sairus

-----Original Message-----
From: Behdad Esfahbod [mailto:behdad.esfahbod at gmail.com] On Behalf Of Behdad Esfahbod
Sent: Wednesday, March 24, 2010 7:36 PM
To: Sairus Patel
Cc: opentype-migration-list at indx.co.uk; mpeg-OTspec at yahoogroups.com
Subject: Re: [mpeg-OTspec] RE: [OpenType] Kerning

On 03/24/2010 08:01 PM, Sairus Patel wrote:
> 
> However, I don't believe it follows from this that the GPOS needs to be
> applied before the kern table in the situation described below. Bidi
> level partitioning needs to happen before either the kern table or GPOS
> table is applied, it's just that the kern table needs to be looked up
> with glyphs in visual order and the GPOS needs to be looked up with
> glyphs in "logical" order. Applying either table requires knowledge of
> the bidi level of the sub-run.

Fair enough.  However, are you also implying that the kern table matching
should skip marks?  Otherwise I don't see how a couple glyphs with a mark in
between can be kerned with the kern table regardless of when it's applied.

And if you do imply that it should skip marks, then how can be kern table used
to position marks?

behdad