[OTVar] OpenType minor update in preparation

We are preparing a minor update to the OpenType spec that will appear soon. For the most part, this will provide clarifications or correction of errors in describing the new table formats used for variations. Some of these may have impact for tool or platform implementations that are in progress if people didn't see past the errors or ambiguities to see the intended design.

There will be one noteworthy design change, though: The version 1.0 STAT table header will get an additional field at the end of the header, "elidedFallbackNameID" (a USHORT). As we worked on implementing support for the STAT table, we noticed that there would be cases, when using the STAT table to generate family and sumfamiy names mapped into legacy font models, that a subfamily string would have only elidable descriptors, leading to an empty string. The elidedFallbackName field allows font developers to specify a fallback subfamily name for use in such situations.

Comments

  • I was wondering about the "slant" parameter.  The spec says:
    Values can be interpreted as the angle, in counter-clockwise degrees, of oblique slant from whatever the designer considers to be upright for that font design. Values must be greater than -90 and less than +90.

    Slant angle is a non-linear quantity.  Should the values be interpolated between the extremes (0 and -90)?

  • edited December 2016
    Sorry for answering my own post, but I already found the answer.  For small angles (<10°) the difference is very small (=~ 4%), so interpolation should work fine.
  • Update: I said that there would be a version 1.0 STAT table update to add a new field. The thinking around extending the table header without bumping the version is that this is so new and tools are still in progress that this would be safe. But, after discussion among platform and tool partners, the conclusion was that it would be better to bump the version to 1.1. Note that use of the 1.0 STAT table will be strongly discouraged.
  • John HudsonJohn Hudson Posts: 1,096
    I suggest we formally deprecate the 1.0 STAT table.
Sign In or Register to comment.