FW: [mpeg-OTspec] RE: Proposed changes to the OFF specification

David Lemon lemon at adobe.com
Mon Jul 25 22:28:34 CEST 2011

Vlad et al.,
The parsing engine would need heuristics (e.g. "if font type is one 
of the following values...") to know what version of the PANOSE 
specification is supported in the font. There are multiple engines 
that do not include such heuristics, and thus make the implicit 
assumption that value n has the same meaning regardless of value 1. 
(It's not just GDI; Adobe has a couple too.) Consequently the result 
is that in the case of Latin-Handwritten or any non-Latin, the values 
are misinterpreted.

Greg is right, of course; the engines are conformant with the spec, 
and the fonts are not. (I'll observe that it would be possible to 
make the spec a lot clearer on this point, which would be a good 
thing.) This is an unfortunate mismatch. The good news is that it's 
not a calamity. I don't know about GDI, but Adobe's engines look at 
only one or two values (depending on the engine), and to the best of 
my knowledge haven't triggered any complaints to date.

- thanks,
  David L

At 8:33 AM -0700 7/25/11, Levantovsky, Vladimir wrote:
>Rather than simply identifying (or even worth, trying to retrofit) 
>the 'offending' fonts that may contain OS/2 'Panose' values outside 
>of "Latin Text" family kind, I think it would be interesting to see 
>how the existing implementations behave when those font are 
>encountered, and what the outcome is. Knowing whether the "Panose 
>1.5" values are ignored, result in being misinterpreted or if they 
>actually cause problems would help to get a better understanding of 
>the problem.
>This is not to say that the limitations imposed by the previous 
>versions of the spec should always define the outcome of the future 
>changes - I think it is perfectly normal and reasonable to expect 
>that at some point the new features may be introduced to improve the 
>standard, and that implementations would have to be updated to 
>handle them. In case with Panose field - it seems reasonable to 
>extend the existing functionality from supporting a "Latin Text" 
>only kind to a full Panose spec, and I can definitely see a benefit 
>of extending font matching capabilities to support non-Latin fonts 
>as well.
>Thank you,

David Lemon
Sr Manager, Type Development
Adobe Systems, Inc.

408 536 4152
lemon at adobe.com

More information about the mpeg-otspec mailing list