[mpeg-OTspec] New "Composite Font Standard" document

[Ken Lunde]

Karsten,

You wrote:

> if I remember Mr Strebe's mail of 2.10.2009 which was the then final one 
> on the topic, and the earlier exchange, it could be GlyphTransformation 
> (outlines only) and MetricsTransformation (all kinds of metrics). With 
> help of the latter, (mark) attachment points could be "slanted" too if 
> the outlines get slanted and thus still provide correct attachment 
> points in this case. (With downside that creator needs to check 
> carefully what he's doing and with the possibility of odd layout 
> behavior when the transformation implies e.g. rotation.)
> Also please check these older mails (with "transformation" in the 
> subject line) yourself, I am a bit distracted with other things and do 
> not want to add mistakes to the document.

Other than the attribute name difference, what I proposed below, which was based on daan's input, should convey this. I will re-read the discussions.

GlyphTransformation can be conveyed as a single attribute, because X- and Y-axis sub-attributes (and other information) are encapsulated in the six-element array. MetricsTransformation must still be conveyed as two separate attributes, because a single scale factor cannot be assumed to apply to both axes.

> Cosmetics: Do I understand it correctly that, of the "Primary Actors", 
> the "creator" and "user" are human beings while "consumer" refers to 
> software? If so, and provided these are not common terms, could there be 
> a less "human" term instead of "consumer", a term which implies that 
> "comsumer" is a machine? But I do not have an idea for one ...

Leonardo's latest email cleared this up, I think.

-- Ken