Meeting minutes
This meeting
Nigel: Today we have IMSC-HRM, DAPT and any other business
… In AOB we should mention the next meeting and gauge likely attendance.
… Any other other business?
no other business
IMSC-HRM
Pierre: We're done!
Nigel: Wow, I think I agree.
Pierre: There are two boxes in the implementation report that are labelled as n/a for the EBU-TT Live implementation
… Do you know what that's about?
Nigel: Yes. I updated the implementation report to add the second validating implementation,
… and I did put a note about the n/a entries.
… The reason is that they stress parts of the HRM that are not valid in EBU-TT-D.
… Specifically style attributes that form part of the glyph tuple, but are prohibited in EBU-TT-D.
… That means that the test documents cannot be loaded into the object model, because
… they give an instant validation failure, even before attempting to process against the HRM.
… There is provision for this in the IMSC-HRM spec, because something similar applies to IMSC 1.0.1
Pierre: Okay, the other thing is we need to remove the Image profile features before moving
… to the next stage. I wanted to get group views on that.
Cyril: What's the reason for that?
Pierre: We don't have implementations, users or implementation report entries,
… and it's been marked at risk.
… I asked if anyone has a library of image profile IMSC documents but nobody has come forward.
Cyril: At Netflix we do use image based captions and we have an IMSC-like manifest so potentially
… I could test against it. But I don't have any objection to removing it from the specification.
Pierre: imscHRM would need to add capabilities for it.
… Unless someone is using it for interchange and wants to see it in imscHrm implementation
… I am not excited to agonise over it.
Nigel: More importantly, should we wait, or move now with what we have?
Cyril: We can always add it back in later.
Pierre: Absolutely, it's being removed without prejudice.
Cyril: No objection from me.
Nigel: Nor me.
<atsushi> +1
Andreas: Fine with me.
Atsushi: No objection
Gary: Sounds good to me.
RESOLUTION: Remove the at-risk image profile feature from IMSC-HRM
Pierre: Good, I'll do that.
… Also, I will remove the editorial note pointing to the open issue.
… We will leave it open but not fix in this version of imsc-hrm
… It's w3c/imsc-hrm#5 - it has not been an issue so I did nothing about it.
Nigel: Yes, it felt like we were overestimating complexity but in practice everything has passed.
Pierre: We should keep it open but note that we are not addressing it in imsc-hrm v1.
Nigel: I think I will propose something different.
… I think we should close on the basis that we haven't demonstrated that there's a real world problem.
… If later someone says they have documents that they think should pass, but which don't,
… then at that stage we should investigate and open an issue based on the investigation, and there's
… a chance it could be this cause, so we could reopen or open a new issue. But right now
… this is not a demonstrated real world problem.
Pierre: I'm fine with that.
… We should note it as a comment, or record what we just discussed.
Nigel: We spent a lot of time on that issue, but I think it gets trumped by the implementation experience.
Gary: We could post this on the issue, wait a short while and then close if nobody objects.
Nigel: +1
… I will add a pointer to this conversation to the issue after the meeting and propose to close it with no change.
… I will also review issue 51, but I think we're fine there and can close without action.
… Then issue 47 is purely editorial to do with term definitions.
… Pierre, should we do anything about those issues?
Pierre: Not unless we're forced to.
… I'm going to concentrate on removing image profile, and the note referencing issue 5.
Nigel: OK, sounds good.
… In terms of requesting transition to CR, we need to see the updated document before proposing
… to move forward.
Pierre: In terms of that, I'd encourage a call for consensus whenever we have the document ready,
… rather than waiting for a meeting.
Nigel: I will wait until we have the document ready.
Pierre: I will try to get it done by tomorrow.
Nigel: Given the season, we should extend the CfC review period.
Pierre: Agree, that's fine.
… Nothing should be surprising here. The risk is limited.
Nigel: +1
… In terms of the tests we now have no open issues or pull requests.
Pierre: Thanks to you we now have working implementations for them all, with bugs fixed, so I think we're good.
Nigel: We even have an extra test that we didn't at first envisage.
Nigel: A nice point to get to at the end of the year.
… Any more on IMSC-HRM?
DAPT
Nigel: The only agenda topic in https://
… which we now have approvals for, barring some minor editorial tweaks.
… Thanks especially to Cyril and Andreas for reviewing so many iterations.
… Hopefully that's a process success rather than an editorial failure! I think it's a good team effort regardless.
Andreas: Thanks Nigel for processing so many comments.
Cyril: It's a big change and reviewing the diffs is very tricky - I had to have multiple tabs open.
… I don't know if there's a better way.
Andreas: I had the same setup, I did wonder if we should have a separate call to agree on resolutions
… that we then apply to the spec. I think you did a great job Nigel, but there were some issues where
… you proposed a solution and then applied it throughout the document. Sometimes you need to
… review the complete sections again.
Nigel: Yes, there were some changes introduced along the journey that I ended up deleting,
… big blocks of text that had no basis in the issue, which I didn't realise without the review comments.
… It's a shame that PR preview doesn't do the example inclusion properly.
Cyril: Yes, the examples make a big difference.
Nigel: One thing I use that could help with review is functionality that may be extensions to VS Code
… that allow me to review and add GitHub comments to the pull request, and preview the whole
… document locally including all the examples. Might help with future reviews.
… My plan for this now is to process the tiny editorial comments remaining, and go ahead and merge.
Next steps
Nigel: We have some new editorial issues, including those raised by DVW, thanks for that.
Cyril: Can external contributors comment in discussions?
Nigel: I can't tell
Atsushi: I don't know but I believe anyone can post there, like a normal issue.
Gary: That's what I would expect by default
Nigel: The reason we're asking is we had a report from someone who said they could not comment.
Atsushi: Let me try a test account later.
Cyril: I don't know if it's someone with a GitHub account, or someone who doesn't have a GitHub account
… and doesn't know how to use it.
Atsushi: Yes, GitHub account is mandatory. I thought the discussion is whether one needs a W3C account
… linked to a GitHub account or not, I'm not sure.
Cyril: I will ask a colleague if they can try to post a test message.
Gary: I'm looking at another repo and I can't see any access restrictions specific to discussions.
Nigel: If we can verify that anyone with a GitHub login can comment then we should add a "how to"
… explainer to the top of the discussion page.
… One issue we need to discuss maybe in the new year is w3c/dapt#110 which is marked as cr-must-have.
… It's about extensibility and backwards compatibility but I don't think we've discussed it or gone
… through the permutations properly.
… Then there are also #75 and #44 about script-type based restrictions and implementation types,
… which you might have new thoughts on Cyril?
Cyril: I'll look at those.
… I'm waiting for the languages PR to be merged before rebasing the other PRs otherwise there
… would be too many conflicts.
Nigel: Understood, that's why I didn't open any more too!
<gkatsev> github confirms that anyone with read for a repo can participate in discussions, see note at the top https://
Nigel: Thanks Gary, that's useful to know. So a GitHub login is all that is required.
… I have people asking me about when to implement, so I'm hoping we can transition to CR in
… maybe February now that the big things have been dealt with, like language and registries etc.
… Anything more on DAPT?
Cyril: nothing from me. It's also my goal to work on implementation.
… That would give me confidence that the spec has at least what I need.
AOB - next meeting
Nigel: The next meeting scheduled is on 4th Jan - I won't be able to make that.
… Not sure what people want to do?
Pierre: Next meeting on 18th is fine.
Cyril: Yes. 18th is better.
Gary: I doubt there'll be much to discuss between now and 4th Jan.
… I think we've skipped it in previous years for similar reasons.
Nigel: Ok, then I will cancel 4th Jan meeting and our next will be 18th Jan.
Meeting close
Nigel: I just want to say thanks to everyone for all your work this year.
… If you have a break, then enjoy it, and if you don't, also enjoy it!
… But thanks everyone, we keep pushing forward.
Gary: Thanks Nigel
Andreas: Yes, thanks a lot Nigel
Atsushi: Happy holidays and new year
Cyril: Happy everything to you all!
Nigel: [adjourns meeting]