Meeting minutes
ray-schwartz is observing via Zoom
as are a few others. Present plus if you want, folks!
beginning recording for APA and COGA Structure v2
APA and COGA Structure v2
janina1 Introducing Rain Michaels for content usable from COGA
Roy - staff contact for APA
Rain Let me clarify this is not the proposal for structure, it is where we currently are in developing it.
… we've been doing a lot of user research to get the structure figured out. We've done quantitative research on how people are using making content usable.
… that data led us to a technique, card sort. Describe the technique as in throwing a lot of ideas out, then have folks group them, then do quantitative analysis on where people are grouping the items.
… we did two different card sorts: one with a broad community not part of W3C who use it.
… then another with the COGA and other W3C.
… then compared the findings.
… That was the pre-work. Now, I'll present a slide deck on the groups we ended up with.
… we then took that into qualitative testing.
… now I'll walk through the high-level groupings and how that is presented, get to feedback and questions.
… we broke it into five objectives: design, interactions, functionality, content, and help.
… [reads through the content on the slides linked]
<Rain> Slide deck that shows the groupings of patterns: https://
Rain continues reading through the linked slides: https://
Slideset: https://
janina1 The word easy shows up often. It's a result, can you control results?
… you can control quality creation.
Rain Thank you for calling that out, we'll look into that.
janina1 Another thing that jumped out was a single-sign-on.
janina1 Perhaps biometric, everyone has a face.
JenStrickland: what is meaningful to me here is that included research...
… sought to speak with people outside and within W3C to inform direction...
… notable. Feel very often that more of that needs to be done, instead of just within the insular community.
<Angela> +1
ray-schwartz I'm a newish member to the ARIA group, so I just wanted to observe.
… I want to comment on "easy" -- yeah, easy vs simple, what is easy for one is not for another, but perhaps simple as the opposite of complex or complicated might be a good option.
… easy is more subjective, where simple is more objective.
janina1 Finding the most important, because what's important at any time may not be at another.
… complexity, we tend to put everything on a page. For APA we make our own wiki that points to the calendar, because it's easier / simpler. What's most important depends on what I went there to do.
… that takes me to a thought about help. We've talked about that a bit in our Adapt. There's an attempt to be contextually helpful, within a process you initiated. AI could be helpful for this.
… Or, you're in a copyright page, table of contents, introduction, and preface… that's maybe not as helpful, especially in a neurodiverse situation.
Rain this may be a good time to cover what we learned in testing.
… showing the Making Content Usable subgroup prototype for TR group layout
… one of the things that made making content usable difficult was the objectives were repeated in multiple places.
<shawn7> [ Shawn asks Janina and/or Roy to please connect with Tamsin about what Janina was saying you do separaretly because the calendar is too hard to use ]
Link to slides: https://
Rain So, a user might be looking for a specific bit of info that was important to them, but couldn't figure out which part of the page where the objective was spelled out was important to them. We found people were overwhelmed by the sheer amount of content on the page. So the "know your user", personas, and research stuff… then creating easy
way finding to get to those, then focus eat TR down to the design patterns themselves.
… Roy built a version of the prototype where we have 3 key sections: 1st is how to use the documents, then breaks down to user role, here's what you may want to know, here's how to learn more about the group… here's links to learn more.
… then the objectives and patterns under objectives, honing in on the pages with the five sections listed out, design, interactions, content, etc.
… so you can see the list of patterns in one moment and you can choose to expand them.
… this tested really well, but the TR note didn't test as well, so we're going to work on another version.
… please note that this multipage prototype is a bit broken, the headers don't work for screenreader as we wanted, so please understand that.
… we certainly learned a lot about how these bits of functionality work in terms of making it easier for them to get to the content that tis most important to them.
… There's a sticky item at the top of the page that allows one to show the navigation, and then it is possible to expand or collapse all things. I stole this from QuickRef, which I have depended on it. It seems like an effective way to give people to get to the thing they need most.
… Another thing that made a big difference is a Note style, a big purple block. The styling helps people.
… What we may be able to explore is creating another couple styles of notes. People wanted a clear visual style, that this piece of content is a warning, or this says look at this other requirement in WCAG.
… Really helping people make sure they understand the relationship between the recommendation and the way it could be implemented that could harm another.
… I'm showing the five principles -- objective or principle, neither has worked well in our research.
… Principles people think of P.O.U.R. and objectives people don't quite understand.
… So here are the high-level principles, then people can drill into them. This worked really well. If folks had a little bit of time, they could work on one, then find where they were and pick it up again.
… Universally, people wanted a little more than just the name, they also wanted a bit of summary.
… I need to leave in 2 minutes, so will stop and address feedback.
<Roy> https://
Multipage prototype: https://
janina1 Anyone have questions / feedback?
… I'm fascinated with taxonomical identification and categorization.
… We had a similar discussion with ebook. We really just reinvented the scrolls like they had in a library in Alexandria, in Egypt… but made it fare more powerful if we organize it well.
… Expand / collapse all is useful, if we do it well.
… I encourage that.
… Theoretically, we have it in HTML5.
… if we're having problems we should identify that.
Preliminary findings (work in progress): https://
@rain Yes, plus 1 to testing. I put a link to a Google doc that has preliminary findings on the work in progress. It will grow as we continue.
janina1 Anyone else with thoughts to share? We are still recording.
APA recharter
@janina we're starting to talk about what needs to change in our existing charter.
… the existing charter is likely the best starting point for the conversation.
… we see things that can stay possibly with some small tweaks, some that are gonna absolutely need to change. This is a common situation, and we certainly are no exception.
… And, one more comment, we'll be asking our various task forces -- and APA does have several -- to look at deliverables, and look at recorganizing for the next charter.
… Now, asking if there are things you would like to see done differently in APA.
… shawn7 are you in a position to comment?
<Roy> Previous conversation on charter
<Roy> Charter of APA
<Roy> Charter of APA
janina1 Roy I need to go to a meeting that starts at the top of the hour.
matatk: *asked about reference to 'previous' charter* Roy pointed out this is correct, as recharter extends a charter.
janina: The other day, we talked about asking the TFs to report on their deliverables.
Roy: What about well-known destinations? Will this be REC-track?
matatk: I think from what we've learnt this week, some of the things we had expected to be TR may actually be achievable without making new specifications - so we would probably use Note track for those, until and unless we need to standardize parts of them in future.
janina: Seems reasonable.
janina: How much breadcrumbing do we need to carry forward into the next charter - noting that e.g. the Adapt TF was formally the Personalization TF, and documenting all the exclusions. This extra text makes it hard to read. Can't we drop some of it?
Roy: Maybe we can add an extra section to document the history? I agree the documents, particularly for Adapt, have changed a lot. I will confirm with W3C process.
janina: I don't mind keeping the data, but the current presentation is drowning out info on deliverables.
Roy: Janina mentioned about FAST - we should refine HR process now we have TAG questionnaire and TAG checklist. In the next 2yrs we should update our process for accessibility review and integrate the self checklist into the HR process.
matatk: Spoke with Sangwhan (former TAG member who championed the TAG questionnaire): the goal there is to see if further review is needed, and if so then groups would be directed to the FAST checklist. It's there to catch things earlier, making sure groups are aware earlier in their work.
matatk: Do we need all of the related/liaison groups?
Roy: re session on HR, maybe we can review some of their specs?
janina: ITU makes sense - maybe for immersive captions.
… We/our members were active there in the past.
matatk: Charter looks good overall; nice! Doesn't seem like major objections. Noted that we need TF input.
janina: agree
Roy: agree - I'll make the template for our October 9th call.
matatk: thanks Roy
MEIG, Timed Text, and APA
https://
Matthew: We have 5 topics, all proposed by APA. Anything else?
Nigel: We had an APA review of IMSC some time ago, need to talk about actions we deferred
… And give a status update on DAPT
cpn: More from Media WG, would like advice on development of media features for the web,
… and how to start with the horizontal reviews
DAPT
<Roy> w3c/
DAPT = Dubbing and Audio description Profiles of TTML2
Nigel: It's a file format specification to support the production of localised an accessible versions of media
… It supports workflows including transcribing the original content, producing a text equivalent what is in the audio and video image, translation, then a re-recording process to generate dubbing, captions
… and mixing instructions for an audio-described version of the video
… It will end up in players at some point. It allows you to expose the text of the descriptions
… An ARIA live region for assistive tech
… We discussed a year ago, and we thought we'd get to CR in a few months. Now we really are about ready for CR
… Some changes made, how to support a re-review?
… WG set to agree in the next few weeks
… It's currently a WD, going to CR
<Roy> https://
Roy: We have reviewed this already, maybe haven't sent to you?
… Some comments in the email, from last year
Matthew: Comments about audio mixing
Nigel: I don't think we have any preference expressed at the moment on that
Matthew: We can file an issue to follow up, if needed
Nigel: Action on us to prepare the list of substantive changes
… On expressing preference, is that informative or normative requirement?
Matthew: I think a Note, might need to wider platform changes but those would be handled elsewhere
Matthew: Thank you for your patience with the review
Nigel: Thank you too, some comments needed detailed thought
<kaz> a11y-request GitHub issue 66 for DAPT review
Kaz: The GitHub issue title says 2021?
Nigel: Looks like a mistake, should be 2023
Matthew: So we'll await your review requet
IMSC review
Nigel: This is the profile of Timed Text Markup Language (TTML), known as Internet Media Subtitles and Captions (IMSC)
… There are a few versions, as Recommendations
… They support delivery of subtitles and captions that meet global layout requirements
… Last version published is 1.2
… There's a set of backlog issues, that are worth working on for a 1.3 version
https://
Nigel: Pierre shared those in TT WG this morning, including some APA feedback
<pal> w3c/
Nigel: This is about semantic layers, forced content
… The purpose of the feature is to allow, within a single document, translation and xx subtitles to be present together
… And the player shows the translation subtitles which are forced, and optionally show those that are not forced, the hard of hearing subtitles
<pal> https://
Nigel: The comment is that the user might want to see speech only, subtitles, or some combination
… Missing comment might be sound effects, not speech, that might be in subtitles
… We deferred this in 2020. Pierre, any questions on the feedback?
Pierre: Not really, it would be good to have something actionable, so we can decide what to do
Matthew: Sounds like what was requested of us was use cases, so the groups could discuss what could be done to address them
… Does that sound reasonable?
Nigel: Yes
Nigel: Want to talk about the wider issues around user choice and playback, for additional context
… The way subtitles and captions are distributed is in individual tracks that the user can switch between
… There isn't a clear model defined for user preferences about information users need. CTA working on something. An algorithm would be needed to map those preferences to player behaviour
… In practice, I think very few people use the forced content approach
… Usually, two tracks are made available, rather than multiplexing them together and having the player switch
Pierre: I've seen both dedicated tracks and mixed tracks, it depends
Nigel: From a player perspective, if we want to give richer information to enable richer presentation decisions, that's a bigger industry change
Pierre: Discussed this many times. There's no industry consensus on a model for player behaviour for timed text tracks
… Each part of the ecosystem has their own way to handle it
… Player behaviour is which track to enable, and how, and based on user preference. There's no standard for this
… Until there's such a standard, it's hard to come up with anything but the most basic signalling
Nigel: I think it's possible to signal this. Whether people signalling subtitles and captions will take it up, might be difficult. Similar getting players to implement
… Work we're doing in DAPT, in producing the transcript we're producing a metadata scheme to allow granular labelling of what the text represents in the related media
… a heirarchical scheme
… it allows every bit of the transcript to be labelled
… And tracks translations
… You could add the same info to IMSC documents, without needing a spec change, treat as foreign metadata (defined in DAPT)
… We could point to that in the spec
… I don't think that would make anyone adopt it, but it's a start
Janina: How much is potentially getting multiplexed? Subtitles, captions, other timed metadata - information about sounds happening, for example
James: Modern captions are usually implemented as SDH, because they have all the benefits of localisation, dialog and dialog-plus
Nigel: Terminology issue around subtitles and captions
Nigel: We don't have additional annotations, no
<pal> Track selection algorithm: https://
Matthew: The idea of a Note to start things off is good. Schedule more discussion to see if we can do more than that
Pierre: To note, IMSC as opposed to TTML is intended to capture practice, so we'd be looking here for actual practice
Horizontal review for new media features
cpn: In the Media WG we are producing specifications that may or may not have been reviewed,
… and I'd like to organise those reviews.
… What kind of self-review should we do in the WG before we bring it to you?
… I ask because you have the MAUR and a long list of questions that you ask,
… so I'm seeking advice. What would you like to see from us as a WG before we raise an issue with you
matatk: Is this before FPWD?
<jcraig> MAUR: https://
cpn: Some are, some are not
matatk: After an early design review from TAG (for pre-FPWD),
… we will get to know about it through that mechanism.
… Appreciate the question.
… Almost certain that if you did a TAG accessibility questionnaire it would need a review,
… because it's a media feature.
… As you point out there are lots of MAUR questions and additional APA check list.
… Is this already established or will need a TAG review?
cpn: Why is this coupled to TAG review?
matatk: TAG gets a lot of reviews of things at the stage as being proposed as possible features.
cpn: Browser features: chromium has a process of requiring TAG review for new features.
… Not all browsers require that.
matatk: May be misunderstanding what a new media feature is
… Thought you meant things that will end up as changes to specs
… Challenge is HR comes after a lot of work
… Super cool you want earlier input, the question is what shape does that take?
… If it's the sort of thing that goes through a TAG design review they would flag it up to us.
… If not, then we need a process that keeps us up to date about the thinking.
cpn: That makes sense, I can take that to the WG.
… To say we're doing something interesting from an a11y point of view.
… I can suggest to the group that it does it through the TAG.
matatk: Could do that, might make sense to let APA or whoever is best placed to review it,
… start with that, and if it's too much and we can't cope, then dial it down.
… Would be a shame to miss things.
… We could look at, if you have a section in your calls for new work, we could just look at those
… sections of your minutes.
… Then we can ask you if we have questions.
… Not sure how helpful that is
cpn: Genuinely very helpful
janina1: I like the idea of hearing about new features early.
… We can give more useful feedback if we see them early
matatk: If we get a few of these through then we may be able to narrow down the core concerns to ask about
cpn: There was also a question about things that have shipped.
… When we go through the HR process and get to CR stage, is there a self-review that you'd like
… us to complete?
matatk: will answer, james first on the q
jcraig: 2 issues for APA for review. One generally happens anyway, which is this "something interesting
… from a11y point of view, let's get a review"
… More common issue is the group doesn't consider that there could be an a11y impact at all.
… Some examples recently include display: contents in CSS, caused a lot of implementation challenges.
… Those are harder to anticipate because the proposers aren't necessarily thinking about it from
… a perspective of a11y, so a11y people with domain knowledge need to be in those groups as
… much as possible. "Toss over the fence" is reasonable in certain scenarios, but there are a number of
… people in CSS WG and HTML WG who are doing that earlier, which can be more effective.
<Zakim> matatk, you wanted to talk about things that shipped
jcraig: Asking for a formal review process is a useful feature of the process as well.
matatk: Totally +1
… Seems like if you're asking you already know the review is needed so you've already
… solved the problem, but more communication between the groups especially for early features
… is a good idea.
cpn: Do you expect us to go through the MAUR line by line prior to asking for a review?
matatk: No, not with that level of granularity
… Stuff that's already shipped: important question more widely.
… We need to review where we can have the most impact, which means doing early,
… if already shipped, we might be stuck.
… Thank you for raising this.
Regulations
Janina: We discussed EU regulations some months back
… A bit of a breakdown of communication
… We want to make sure what we've documented for accessibility is well understood and things aren't left out
Matthew: We might need to come back to this issue
… Have a future meeting
James: Which groups would need to be involved?
Janina: I think it related to television
James: RTT point-to-point has been deployed in the US for a few years now, doesn't support multi-party RTT
… There was some EAA requirement where that part of the protocol needed updating
… RFC 9071
… That's been added, but no implementations yet
… I don't think there's a gap in multi-party RTT anymore, not widely deployed yet
… We recently added user requirements related to multi-party RTT, a document ETSI TR 103 708
Matthew: We'll check on this
Symbols in media chapter titles
https://
Matthew: There is work and discussion ongoing on how best to implement this, with WHATWG
… ADAPT task force, to bring symbols to web content
… The challenge is that it's not as simple as adding images to webpages
… There are many symbol sets
Janina: There are specific symbol sets, in academic or rehabilitation situations. People have learned one set, and rely on it heavily
… Not emojis, but there are similarities
Matthew: They're used as ways to select a symbol and produce speech
… Also added to content to help people understand the content
… Examples of BLISS and ARASAAC
… How to provide these in normal web content? Could be an extra element or attribute associated with the text we want to augment with the symbol
… Could mark up semantically, to link to the text displayed
… This is a common thing for people with cognitive issues, so prefer audio or video content, so the killer app is to make the media chapters more understandable
… Hence the question to the media groups. Ways being developed to surface chapter titles
cpn: this relates to a feature we discussed in Media WG this morning
… which may have already shipped,
… which is Chapter information in Media Session API.
… That relates to UI surfaced by the browser chrome, not in-page.
… In order to provide these symbols in that context there would need to be browser chrome UI
… designed and then a mechanism for the page to provide the symbols.
… It would need the implementors to be onboard to honour these symbols in the UI
… Then allow the page authors to provide the info.
jcraig: I've probably had a copy of this card from the last 4 TPACs.
… Every time, I and others, including the ones this week, suggested these symbols need to be in Unicode.
… After the last time, I looked up, and it was proposed in Unicode in the late '90s.
… If someone had followed through they would already be in Unicode.
… I know there are some supersets that probably will never land.
… That needs to be how its done, including how to combine these symbols.
… Also using Ruby stacking is a good way to associate the characters with the underlying text.
… You can use Ruby stacking in the captions as well.
… Backing up a bit, I'm not sure, other than what we discussed before, Unicode and Ruby,
… when that's widely deployed we could consider doing it at the OS level for native apps.
… Is there more you're asking?
matatk: I understand that media in this proposed metadata standard for media session API, that
… the chapter titles are strings so we can't put HTML markup in there, so we would need a
… separate property for the symbol, as a different string.
… Great to put them in captions in the future maybe.
cpn: I think you're right you can't do HTML markup in chapter titles.
janina1: That's why we're bringing it up here.
… Still hammering out details with WHATWG.
… Once they are hammered out we will continue to have that gap until
… markup is permitted in those chapters.
matatk: +1 to jcraig we're more constrained in this area.
… What's the likelihood of being able to add this extra field to the API?
cpn: I can't answer, needs to go back to the WG.
… Depends on the UX designers and the browser chrome implementers.
… Can figure out the mechanism, if we need HTML or a plain text description with a symbol.
… This is all UI that's provided at that level, needs their support, and a WG discussion to figure it out.
Gary: A lot of what players like Video.js which have a chapters UI do is used WebVTT chapters
… Would be easier if they were Unicode. Could add to WebVTT potentially, if Unicode isn't good enough?
<Zakim> matatk, you wanted to ask about extensions
Gary: For Media Session, it's possible we could use the chapters kind in WebVTT, have one with text or text with symbols
<Zakim> jcraig, you wanted to discuss recent vtt metadata changes
Matthew: I assume a browser extension could access this
James: We have a WebVTT pull request, to allow disambiguation of WebVTT metadata, e.g., lists of thumbnails for scrubbing a timeline. Chapters is another ad-hoc usage
… Could add alt text or localised alt text
… In this case, we could do something with the chapters usage, include the symbols as as a symbolic image alternative to the chapter text
<jcraig> Original light-sensitivity flash mitigation use case: w3c/
<jcraig> Metadata disambiguation w3c/
<jcraig> Pull Request w3c/
Update on Immersive Captions
<kaz> Recommendations for accessible captions in 360 degree video
Matthew: Immersive Captions CG produced a report
… APA has adopted that report, and people from the group have joined APA to continue work on it
Janina: It will be on the Note track
<Roy> w3c/
Matthew: Any collaboration between Media groups and Immersive Web groups
Nigel: We've had one or two meetings, but there isn't any formal work. Immersive Captions group hasn't come to TTWG, neither has anything happened in the other direction
… There has been some consideration about immersive captions and immersive space, 360 video
… One of our members had an interest in that, but their priorities changed
… If formal spec work is needed on caption formats, we should start the conversation
Matthew: We'll try to communicate more with you
… Any communication between Media groups and Immersive Web groups
cpn: Not really! Nigel partly answered the question, because the person who was actively looking at this
… was in the MEIG and we had a couple of joint meetings with the Immersive Web WG,
… teleconferences and TPAC a number of years ago.
… Since then, no, not really anything.
… On the Media WG side, not so much, I think the answer is no.
… There was a conversation this week about the HTML <model> element, and if it is a media element.
… Much more of an architectural discussion.
… I'm happy to facilitate the conversation.
matatk: Piqued our curiosity, we thought if this is happening there must be other collaboration going on.
… You've answered that, so we will listen out.
… I reiterate my apology on behalf of us because we're responsible for it but did not communicate
… this to you.
Atsushi: As Immersive Web team contact, we had some discussion between accessibility group and Immersive Captions group and media group, around 2019
… Conclusion was no specific feature needed. It was displaying caption with a video, possible to implement using the DOM layer spec
… If there's any new feature needed, Immersive Web CG or WG is happy to discuss
Matthew: Interested to follow the work of APA group
ack
<kaz> WebVMT Note by the Spatial Data on the Web WG
Kaz: There was other work by Spatial Data on the Web group, WebVMT, we might want to look at that as well
MAUR
Matthew: MAUR hasn't been updated since 2015. Looking for funding to enable that
https://
Nigel: How much funding?
Janina: 20-30% FTE to do the editing work, for a year maybe
Conclusion
Matthew: Thank you all for contributing. Look forward to continuing to work with you