Meeting minutes
<chris> oh come on
garret explaining the two changes to sparse bit sets to reduce size.
<Vlad> Discussion of the experimental results documented at https://
Vlad: if you had to pick one branch factor (so no variable) which one?
Garret: 4 is probably the best, set sizes are usually going to be the smallest
Garret: if we use variable the encoder could use a heuristic or test the different factors
Vlad: picking between the two will take some time, probably not justified by the savings.
Garret: ok, so sounds like we will stick with just 4
Garret: likewise for the intervals lets stick with the current method and not introduce a second mechanism for encoding intervals.
codepoint injection
<Vlad> See https://
Next topic, should developers be able to manually request extra codepoints?
Main use case I see is when developer may know what content is coming in the future.
Vlad: my concern is you start by sending a request for a specific set of codepoints. The process stays the same regardless of where they came from. For the incremental loading we don't care.
Vlad: so seems like an orthogonal issue.
Garret: my feelings is this doesn't need to be part of the spec, but could be added later to either the font loading api and/or css specs.
Chris: will we be leaving it up to the browser to decide what to request.
Garret: current plan is to continue working on spec, who wants to review.
Vlad: I can.
Vlad: call next week?
Garret: probably not.
Garret: will send email.