Timed Text Working Group Teleconference

 

 

Test results

Sean: I'm working on an implementation, called TimedTextPad
... I cover all of the timing
... Paragraph002 is incorrect now, since body is a par
... so Timing and content are well covered
... and Glenn covers those pretty well as well
... issue is with Styling
... three major grouping of functionalities
... color: background, visibility, opcate, color
... I pass all of those. Glenn does. Philippe as well
... so we're fine
... then fonts stuff
... fonts is relatively easy to pass for me and should be straighforward
... both Glenn and Philippe pass those
... tests aren't very well organized
... some of those tests need to be slightly reworked
... the only problematic one at is the blur raduis on text outline
... Glenn and Philippe don't do

Philippe: willing to submit new tests for font family?

Sean: eventually, I'll figure out if every test is correct
... some of the tests cannot be told if they work or not
... biggest concern is writing mode and unicode bidi group
... vertical text and right to left. so, we don't have that many feature that don't have two implementations

Sean: there are some stress tests we could provide
... like lots of nested, multiple regions
... ie testing the implementation rather than the spec

Philippe: would be nice to have, but not a requirement

Sean: could put them into a separate bucket
... we need the animation tests and all the other in the spreadsheet

Philippe: will do

Sean: can you run Glen's implementation?

Philippe: not for the 27 unfortunately

Sean: I'll run tests on Glen then [David will send Glenn's implementation to Sean]

David: what about the metadata test?

Sean: our implementations don't do anything with those

Philippe: at least, we don't crash :)

Sean: If I get around the rest, I might do something

David: we do something with them, when converting to our own XML format
... the color code that we use is sophisticated
... using scene details as well
... I could send a pointer around, it's under license GPLv2

Sean: interested to the code but license issue :(.
... so, we're in a pretty good shape for the test results
... how about parameters: cell resolution? clock mode?

David: if you're dealing with region, it would be relatively easy

Sean: yes
... framerate is fine, no sure with multipliers. but that basic arithmetic

Philippe: I could deal with some of those as well but don't have time at the moment

Schedule

Sean: let's pick May 1. June 1 for end of last call
... July 1 done with comments
... PR by July 1

Philippe: that would be REC in August. not the best time for a press release

sean: then, let's say press release last week of August or first week of september

Philippe: if Glenn doesn't show up by April 1, we can look into finding an alternate editor

David: there was two or three required features (like the requiredFeatures attribute) what the authoring system is required to support. What about those?

Sean: I'll look at the features in appendix C. Since we don't have spec text for those, we might have to drop them.
... we should leave April for writing and testing implementations
... for example. for blur radius, we need to have that text provisional.

Philippe: we should make it clear that everything needs to be implemented no later than July 1. Everything that it is not is out. All issues should be identified by April 1. No more new things into the draft after May 1.

Next meeting

No meeting on March 20

Next meeting is March 27

Regrets from plh

[adjourned]