[mpeg-OTspec] RE: [OpenType] Kerning

Levantovsky, Vladimir vladimir.levantovsky at monotypeimaging.com
Thu Mar 25 22:24:14 CET 2010


Adding mpeg-OTspec at yahoogroups.com list back to this discussion in an effort to bring this run-away train back on tracks. When responding, please make sure that both MPEG-OTspec and OpenType lists are present in the "To:" field.


On Thursday, March 25, 2010 4:17 PM Behdad Esfahbod wrote:
> 
> On 03/25/2010 03:49 PM, John Hudson wrote:
> >
> > I can understand that if you encounter such fonts, then you want your
> > layout engine to do something with them, but I think what it does is
> > probably up to you and not something that the OpenType spec needs to
> > address because, as Adam suggests, this issue involves two different
> > layout mechanisms that have not been designed to be compatible. Can
> you
> > figure out a way to make them more rather than less compatible?
> > Probably, but you are never going to make them totally compatible --
> if
> > that were possible, there wouldn't have been a need for kerning in
> GPOS
> > --, so its a question of how much effort you think this deserves.
> 
> I think you are missing the point of the discussion.  This whole thread
> is
> about recommending what layout engines should do when both kern and
> some
> GSUB/GPOS is available.  Aren't you getting rid of the question instead
> of
> answering it?  If the recommendation is "no recommendation", sure,
> that's an
> option too.  But that's not what's currently on the table from my
> understanding.
> 
> behdad
> 

Yes, this is exactly right. We need to make a decision on whether we should recommend a particular order of processing when both 'kern' and GPOS are present (also taking into account that, as Sairus pointed out, GPOS may be present but do no kerning). If we decide to recommend a particular order, we need to reach a consensus on things. 

"No recommendation" is also a viable option if we believe that it doesn’t belong in the spec; currently, the proposed text says:
" If the number of kern feature lookups in the resolved language system in the GPOS table is zero, then the kern table should be applied, followed by any remaining GPOS features requested."

Regards,
Vladimir



More information about the mpeg-otspec mailing list