Meeting minutes
IFTB demo
Garret is presenting a demo he and skef worked on on the back of the Binned Incremental Font Transfer Demo
Demo demonstrates the IBFT mechanism to incrementally load Chinese text
Shows transfer sizes of IBFT side-by-side with Unicode range
skef: doesn’t know if it will ever be better than unicode
garret: pro is the size of the chunks that can be transferred, should allow us to be a bit tighter with what we transfer
garret: good to see this working, hasn’t shared publicly yet
vlad: consider showing examples that will make a case for IBFT as someone without context might wonder why to both with IBFT if unicode is consistently lower
Update on the latest state of shared dictionary transfer.
<Vlad> WICG/
garret: compression dictionary transport continued to evolve - changes to spec include: header names (such as use-as-dictionary)
garret: should we adopt these patterns?
garret: this will store dictionaries to disk and you can start incrementing against this; question is whether we want to continue to modify our spec to match the dictionary transport one
skef and vlad: agree it’s a good idea
garret: has sharable chrome change list to explain the mechanics behind it
skef: can we reuse chrome’s decompression code?
garret: not quite ready to have chrome integration for IFTB yet
garret: mark from http working group: consider integrating into range request framework as other-range. benefit: clearer semantically that you get a partial resource. Needs to understand better what the ramifications are.
garret: will probably work but might need to get font patch request on safe list
garret: will keep investigating that
TPAC planning & agenda
vlad: we’ll have one full day allocated for group meetings. Wants to plan what the best use of our time is. And plan what’s going to happen in the year ahead.
garret: progressing IBFT
garret: talk about patch subset range request in more detail
garret: what are the next steps on the patch subset side of the spec?
vlad: when spec work is complete and approved, job is not done. client performance testing might be more time consuming if we don’t plan this ahead of time.
garret: server side conformance tests in reasonably good shape, just have to be brought up-to-date
vlad: we’ll have to produce a large set of tests
vlad: expects this discussion to take quite a lot of time of the day
vlad: leaves us 1-2 hours to spare, anything else worth talking about?
skef: IFTB - use meeting to talk about organising the spec or garret, vlad will input on what goes where in spec to prep for the meeting
vlad: we can probably use range request doc as template
skef: link to spec work in presentation; review that and plan what else need to go into the spec
<skef> https://
garret: we can build onto skef’s slides, notes, spec draft to make it fit into split spec world
vlad: what are skef’s preferences for timing in the day since he’s joining remotely
skef: planning to attend all day, no preference for IFTB part of the agenda
skef: no easy way to link to material but everyone on the list should have received this as email attachment
garret: will block progress to next stage, keep separate until we have it where we want it to be
skef: agrees, patch subset doc to be editorially enhanced
<Garret> https://
skef: IFTB attaches terms to GIDs, he wants to describe that process in the doc
skef: rough and ready on language, clean-up to be shared responsibility of WG
vlad: list of agenda items: 1 - IFTB spec as separate draft, 2 - general spec development, 3 - conformance tests & timeline etc
vlad: social activities: dinner?
garret: rod, cosimo, jimmy (?) expected to join in person