A historic note: When the API was added some where still skeptical of PDF, so everything that PDF needs to know about fonts is packaged up into the one API.  It seems perfectly reasonable to pull it apart into multiple requests; adding info to some existing APIs and creating some new ones.

A quick look through the current API...
- font name, type, max glyph id, total bbox, em size (there are multiple consumers em size) -- these can all be bundled together as a general info about a font, or pulled apart into separate calls.
- style info (flags, italic angle, ascent, descent, stemv, cap height)
- get advances (horizontal or vertical)
- guess unicode table
- get glyph names (type 1 only)

multimaster bit isn't really used.

Happy to talk more on this if useful.

Certainly one way to break things up is if the requested info is per-glyph or font-wide.

Assigning to current PDF owner.

why ice-box? I find that interface gigantic and hard-to-understand. Since its private (afaik), lets make it a priority to fix/reduce it.

If this should happen, I'm not the person to do it.  I think bungeman@ should own this.  He can provide any internal API on SkTypeface, as long as everything SkPDFFont needs is provided somehow.