[MPEG-OTSPEC] CSS WG liaison to SC29 on Open Font Format (from Feb. 2020)

Laurence Penney lorp at lorp.org
Fri Mar 15 16:27:03 CET 2024


In general it would be nice to have a method of knowing, given a designspace location, if any clamping has taken place. Currently the only way to do this would be for an app to load the fvar table and compare with the requested designspace location: not hard in principle, but often not possible (e.g. in a browser). Providing min/max/default for each axis would be the simplest way to achieve this.

- Laurence

> On 15 Mar 2024, at 15:07, John Hudson via mpeg-otspec <mpeg-otspec at lists.aau.at> wrote:
> 
> I presume the idea is that all the intelligence for substitutions and insertions resides in the feature variations GSUB for the JSTF axis, including contextually controlled kashida insertion logic, and the justification engine would only be responsible for figuring out how much to contract or expand lines to fit the measure?
> 
> For the benefit of the engine, we would need to define what should happen if the JSTF axis maxes out and the measure still has not been filled—probably allowing further expansion of inter-word spacing as per most existing justification engines.
> 
> As an axis that is intended to interact with common software functionality, this looks to me like something that might be a clear candidate for formal registration in the spec alongside weight, width, etc.
> 
> JH
> 
> 
> On 2024-03-15 4:29 am, Simon Cozens via mpeg-otspec wrote:
>> On 14/03/2024 15:02, John Hudson via mpeg-otspec wrote:
>>> ‘JSTF=0 - JSTF=1000’ implies that justification always means expanding-to-fill. Conceptually, the model could handle contraction also, I think. Has this been considered in the implementation?
>> 
>> Sure, starting the JSTF axis at zero was just an example to how different things could be done along different expansion values. It works mutatis mutandis the other way around: if you want a font that runs JSTF=-1000 to JSTF=1000, just arrange for the glyphs to contract or switch to condensed shapes at appropriate points along the negative part of the axis.
>> 
>> S



More information about the mpeg-otspec mailing list