Meeting minutes
Review open issues
Vlad: there are 4 lingering from 2022. One we want to keep, other 3 I'm not sure
Skef: Not sure if Discoverability is relevant
Garret: I will reach out to Yoav to see if it is still relevant
Garret: Shared Brotli ID issue (101) I think is no longer relevant and can likely be closed (will check with Chris)
Skef: Issue 103 (a-la-carte IFT) is relevant, but possibly related to a different spec
Garret: Yes, I think possibly extending CSS font laoding
Skef: Issue 192 - might want to load only a chunk of a font (like certain locales in Noto). Being able to specify code point sets could be helpful
Garret: You mentioned a sort of subtraction of one set from another so you could then get 'everything else' - I think this could be implemented pretty easily
Garret: the second approach seems much more complex
Garret: if you have 3 subsets: A, B, and 'everything else' (all minus A and B)
Garret: I think are covered by existing methods, but if we had a real example we could test it out
Skef: I can provide a sample
Vlad: let's go back to the agenda
Hosted format discussion
John Hudson: I think we covered it mostly on our last call but we need to be able to explain it to font foundries what is required for font formats
John Hudson: It's mostly an education issue. Foundries need to make a decision about whether or not existing subsetting language covers this use case
John Hudson: We need to clear up what format is required (WOFF2 or TTF, etc)
Garret: The spec is designed to work with WOFF2 as well, and there are only a small number of things you can't do
Skef: this is more of an issue for the previous versions of the spec, but for the latest iteration is likely not an issue at all
Skef: In general the main licensing issue is to clarify that it's acceptable to serve fonts in this new format
John Hudson: I'm sure it's not a problem with intent, it's going to be language specifics. The question will be the language details of how subsetting and the format will determine if current EULAs will allow this kind of subsetting
John Hudson: Adding tables in the font files to identify the other segments is technically a modification, so that has to be acknowledged
Skef: We have language in the spec that says 'If you follow the "shoulds" in the recommendation, and the client does so as well, the resulting font should be identical and unaltered
Garret: It's possible to also simply ship fonts in the 'IFT' format already, so it would not technically be an alteration
Garret: WOFF2s wil work, IFT will simply be a bit more optimized
Skef: If you provide a font with no segmenting instructions, the encoder should still work based on reasonable use and it should be fine
Garret: I haven't had time to prototype this, but we might be able to create an initial patch that is just a patch map that could be encoded in a data URI within the CSS to avoid the initial round-trip
Garret: It would be functionally similar to Unicode subsetting
Garret: the encoder could spit out some CSS to implement
Stakeholder support/opposition data gathering
Vlad: We need to all use our contacts to popularize IFT and if possible get statements of support
Garret: Dave said he could reach out to some CJK foundries
John Hudson: When Open Type Variations were announced, Peter Constable centered discussion on Type Drawers
Vlad: We discussed using Github
John Hudson: It may not be the best platform to discuss licensing questions. Not everyone has a Github accounts
Skef: At Adobe we already have licensing that covers our own subsetting capability that adds tables. We feel this covers the new format and intend to try it out
Vlad: It's mostly educational outreach that will help foundries understand and adopt this
Garret: If not Github, where?
John Hudson: Type Drawers might work. It's in a good place now, but for a time there were some very toxic folks there and a lot of people left.
John Hudson: Alphabettes came about in part as a response to that
Bianca Berning: I can help spread the word, but not sure there is a single place anymore
Garret: We need an explainer
bberning: I wonder if Adobe could share the changes they made in a sort of case study
Skef: I'm not sure - our licensing sort of goes in the other direction
bberning: OK - maybe it just needs to be part of the explainer. It needs to have info to help foundries and also end users (does their license cover this use?)
Skef: If we want to do this we might need to have a decent encoder prototype to have as part of the intro/explanation
John Hudson: I agree, and you'll probably end up getting more feedback from foundries about what they'd like the encoder to do
Garret: Yes, I haven't gotten to that yet but it's certainly possible
Garret: I have been working actively on the encoder lately, which is where a lot of these proposed changes come from
TPAC updates
Vlad: some changes in format this year. Registration is open (early bird has expired). Good news for Invited Experts: no charge to attend!
Vlad: Location is in Anaheim, September 23-27 (our meeting is on the 24th)
Vlad: there's another reason to try and go if you can. There will be a big event on Wednesday to celebrate the 30th anniversary of the W3C!
Vlad: heads up that the Zoom will be different for that meeting
Vlad: next call on August 20