See also: IRC log
<scribe> scribe: nigel
Nigel: Let's do what we can today
given the unusually low attendance.
... We have very little on the agenda today, just the following
that I'm aware of:
... TTML1 3rd Ed CR transition request
... The two open actions on tracker for TTML2.
... Anything else?
Glenn: I'd like to look at ttml2#696
Nigel: Referring back to the
resolution we made 2 weeks ago, to request transition of
... TTML1 to CR after merging the open pull requests, the
review period ends now, so the
... decision is final, and I will go ahead and make the
transition request based on the
... version on the TTML1-3ED-CR1 branch, viewable at https://rawgit.com/w3c/ttml1/TTML1-3ED-CR1-build/index.html
... In the absence of Thierry right now, I'll do that after the
meeting.
Action-443?
<trackbot> Action-443 -- Glenn Adams to Prepare a document showing mapping arib ruby extension features to ttml2 for use as a liaison document to arib. -- due 2016-08-26 -- OPEN
<trackbot> http://www.w3.org/AudioVideo/TT/tracker/actions/443
Glenn: I still don't think it
would be a bad idea to do that, I just don't have the
wherewithal
... to do it, so we could either close it and defer it for the
future or assign it to someone else.
... It may be that as TTML2 gets closer to Rec I might have
time to do it, but in terms of
... communicating the work that we've done to communicate the
work we've done to support
... Japanese to ARIB would be the proper thing to do.
Nigel: It would be timely to do
this now since we are in CR and interested in
implementation
... work, and ARIB members are possible implementers.
Glenn: OK, leave it with me.
ACTION-443: WG reviewed this and will attempt to satisfy it during CR1
<trackbot> Notes added to ACTION-443 Prepare a document showing mapping arib ruby extension features to ttml2 for use as a liaison document to arib..
Nigel: I've updated the due date to end of April.
Action-462?
<trackbot> Action-462 -- Glenn Adams to Create issue on ttml2 to add "mapping from other versions and profiles of ttml" -- due 2016-04-28 -- OPEN
<trackbot> http://www.w3.org/AudioVideo/TT/tracker/actions/462
Nigel: The question in my mind here is: has this been overtaken by events?
Glenn: I think putting this kind
of information is putting the cart before the horse. The
... mapping of profiles should be put into updates to the
profiles.
... We have listed substantive changes since TTML1 already.
Nigel: What about mapping to make a TTML2 document instance from a TTML1 document instance?
Glenn: That's easy, it's a no-op.
Nigel: Because we haven't prohibited anything from TTML1, right?
Glenn: Yes, we've deprecated a
few things, that's about it.
... If we obsolete anything then we should do something
here.
Nigel: Looking back on the
minutes from when we originally raised this action, there
was
... a query about mapping from smpte:backgroundImage to the
TTML2 image element.
... In fact that is already noted in TTML2 and more detail is
in IMSC 1.1. So I think there is
... no remaining action needed here.
Glenn: For me the potential
action would be for SMPTE to update their spec to refer
to
... IMSC 1.1 or paraphrase it. Not for TTML to go back and
reference SMPTE-TT.
ACTION-462: Discussed in WG meeting today, overtaken by events and no longer needed.
<trackbot> Notes added to ACTION-462 Create issue on ttml2 to add "mapping from other versions and profiles of ttml".
close action-462
<trackbot> Closed action-462.
Glenn: After this is merged I'm
willing to create a release task, and make a release
containing
... the artefacts, pointing to gh-pages. I still want to upload
the tar file, which is consistent
... with what we did in the past in both TTML2 and TTML1. I
haven't looked at gh-pages to
... see if there are other files.
Nigel: This is a technical point - we could achieve the same goal by tagging the gh-pages branch with a release.
Glenn: I'm suggesting merging
this pull request and then I can also add a release based
on
... the gh-pages branch.
Nigel: Ok, I'll approve.
Glenn: Thank you, I'll go ahead and create an artefacts release when the build has completed.
Glenn: Since one was necessary to put into the CR SOTD I created a template as a starting point
Glenn: I've added the new
features and have started putting the implementations in.
... I expect the entries to change as I verify each one over
the next month.
... I'm also discussing this with Netflix and they will be
adding their entries as well to new
... columns, so I'll coordinate with Cyril on that.
Nigel: I'll add what we have
especially for the audio features from BBC.
... I think I'll ask Thierry to style this to look more like
what an implementation report
... normally looks like.
Glenn: I'd like to discuss it before we do that.
Nigel: For an example of another
one, see
https://www.w3.org/wiki/TimedText/IMSC1_0_1_Implementation_Report
... We can't really claim anything in the implementation report
until we have tests.
Glenn: I don't think so, we used
the features as a list and solicited input from
implementers
... on whether they implemented it or not. We did not tie
specific tests to specific features,
... at least during that process.
<glenn> https://www.w3.org/2006/DFXP/DFXP-implementation-result.html
Nigel: That link above is what
you get to when you click on Implementation Report in the
... SOTD from TTML1 first edition PR https://www.w3.org/TR/2010/PR-ttaf1-dfxp-20100914/
Glenn: I don't recall the
discussion of this in the group at the time. I think we threw
it over
... to Thierry to create whatever documents were needed to meet
the Process.
Nigel: I've not seen any W3C CRs whose implementation reports did not refer to tests.
Glenn: We will be making a lot of
tests available. We use features to identify
implementable
... support so we need to tie the features to tests as you
suggest.
Nigel: One feature to one or more tests, yes.
Glenn: Either we can integrate
those into the same table, for example with tests
corresponding
... to each feature. I'm concerned about the need to correlate
tests to feature designations.
Nigel: I agree that we would need to do that.
Glenn: As long as we can do that,
I'm happy with that. The current proposed table does
... not include that information so we need to add it
somehow.
Nigel: I agree.
Glenn: One way would be to have a
second table that maps features to tests, and then we
... could also have pass/fail type entries for the
implementations in there. Then the feature
... based one would be a summary of the test one.
Nigel: Seems too complex to me. They can just go in one table.
Nigel: The IMSC 1 one does that
to some extent, though again like the TTML1 implementation
report,
... it does not explicitly list the feature designators.
... It does include heading rows linking back to a part of the
spec.
Glenn: What I want to get back to
is a claim by the implementor that specific features are
... supported.
... This IMSC 1 table doesn't map the tests to features so you
might hypothetically pass the test
... but not fully support a feature, and then abort processing
as a result.
... In the TTV test suite I have a test called
ttml2-valid-all-styles that has every style that is
... defined in TTML2 that is new. If I were to list that in a
table like that here, then it would
... map to 20-30 features.
Nigel: Agreed, it isn't useful in
this context, though it may be useful in its own right.
... You might need to split it up into one style per
feature.
Glenn: Making a judgement about which features are needed for each test might be tricky.
Nigel: Agreed, each test is
dependent on a number of features.
... If you had a single test that checks for support for all
the mandatory features, that could
... be a good basis to begin with.
Andreas: It could be overkill to
list all the features for every test. If a test were to fail,
how
... would you discover exactly which feature was not supported
properly. There's an endless
... list of features that are depended on - XML itself for
example!
Glenn: It has never been the
explicit requirement that testing is needed to
demonstrate
... interoperability to transition to PR - the implementation
is there to demonstrate
... implementability.
Nigel: That's true in the Process:
Glenn: I would like to follow our
history and demonstrate implementability as opposed to
... interoperability, but we need to have everyone on the same
page, since you and maybe Pierre
... had some different views.
Nigel: That's true. It would be
crazy to attempt to test every possible permutation of
combinations
... of features. Right now I'm expecting 2 tests per feature,
one that is valid and demonstrates
... a useful behaviour related to the feature, and one that is
invalid and causes a validating
... processor to issue an error. This is from our exit
criteria.
Glenn: I think I agree with that.
We will submit three categories of test for adoption and
use.
... I had asked a question to the group about where
ESH_additions came from in TTML1
Nigel: I don't know. What is in Mercurial too?
Glenn: I couldn't find a record
in the history from Mercurial to indicate where it came
from.
... Inside the files in that directory, some were labelled W3C
test material, others designated
... as copyright Microsoft. There's one that was completely
bogus, and had microsoft labelling
... in it, which makes me nervous, which has a "test" on
profile, with a features element
... that used a syntactic style similar to how we do
tts:styles, with a bunch of tts attributes
... the value of which was broken or optional. It was broken
compared to the actual syntax
... and the label on it made me worried that someone had
actually implemented something like that.
Nigel: Do we need to worry about it?
Glenn: We need to remove the
directory because people might think it is part of the
TTML1
... test suite. I'd like you Nigel maybe to at least
preliminarily take an action to ask if we
... should remove it and find out what the provenance of it
is.
Nigel: Ok I'll do that.
Nigel: We're out of agenda so I'll close the meeting now. Thank you! [adjourns meeting]