Meeting minutes
IDREFs discussion
<JenStrickland> I have a question for FazioFazio: where is the latest A11y MM excel spreadsheet? The link I have is outdated.
JenStricklandGeneral and well-deserved Matthew praise from the PWEAG. Clone him.
<JenStrickland> Fazio, where is the latest A11y MM excel spreadsheet? The link I have is outdated.
matatk Recap: we want to go over four thigns. A recap of the possible benefits. One of the a11y concerns. Then discussion of what sort of user research would be helpful (hopefully resulting in an MVP). And how might we follow this up.
matatk Or ust look in the invite.
matatk It could be the case that it improves a11y if duplicate ids are removed or the possibility thereof is removed.
matatk High-level benefits are being discussed.
jyasskin Component system cmae up as a pain point. It can be hard to identify the right id. There are DX problems with having to assign the right ids for the use of this.
<Fazio> JenStrickland AccessibilityMaturity Model Experimental Prototype Assessment Spreadsheet link: https://
jyasskin We should think about how to have a good conversation at TPAC. A11y people tend to be outnumbered and the discussion's goodness will suffer from that.
<Fazio> nevermind links is broken
JenStrickland Having duplicate ids in a document can cause the UX to not be what was designed ordeveloped. It is not just a terrible a11y experience but also breaks what was intended.
janina I just wanted to say that this thing about dupe ids: I don't think anyone is promoting the usefulness of duplicate ids. WE would like to not have the problem at all. You should be able to piont anywhere ina document reliably. I want to talk about documents and not webpages. Documents, fi you run through a validator, you can find ids..l
Accessibility Concerns
matatk Under a11y concenrs, I wanted to put thre to four top-level things.
matatk The last one wants to offer solutions.
matatk The overarching goncern we have is about awareness that there are costs beyond the ones of the proponents of these changes might be aware of.
matatk One area is authoring concerns and one is acutal implementation. Under authoring concerns: satability of the references, questiosn about whetehr most HTML that exists out there in the wild would lend itself to...
matatk Developer undertanding of relative references to. I'm not 100% sure that wwe now how good that would be.
matatk In terms of impelemtnation for UA and AT...
<JenStrickland> For others that are unfamiliar with "foot gun": "Generally a footgun is something that's designed in such a way as to be extremely likely to be used in a way that's going to cause problems for yourself. Yes, it's a play on shooting yourself in the foot, the idea is that you designed a gun designed to shoot yourself in the foot."
matatk Tehre is still quite a burden of implementation for AT and UA. Pssobility of lower performance especially for AT users.
matatk Again, the stability of references
matatk There is providing a mechanism for scopign idrefs. There is also a relative reference approach.
matatk Just looking at providing scoping within the templates for custom elements, that was something proposed but discounted. It would be worth exploring more. It would only require changes to small parts of the stack.
matatk This all could solve a huge chunk of the probelm of duplicate ides.
matatk Let's now think of what research would be helpful.
sarah HOw about splitting out othoring concerns into time of authoring and maintnenance?
<janina> +1 to separating initial authoring from maintanance
matatk One of the things that might be helpful for us to do is to have agood core set of benefits and concerns that we are on the same page about.
<sarah> whatwg/
Fredrik Annoying scribe reminder to link to the calendar nvite.
<sarah> s/> whatwg/
<matatk> Calendar invite for this call - links to lots of intersting and helpful things! https://
<sarah> hah, apparently I can't edit fast enough
matatkWhat would we think would be helpful in terms of research?
jyasskin The template thing is a little worrying. If it only applies to custom elements it leavs out a lot of the component systems in use on the Web. The other one was, I feel like the a11y people who participate on the thread (Sarah) mwerej worried aobut the characteristics of the proposed solution. Elaborate, please?
sarah I'd add user concerns too but they seem already to be covered.
matatk A reference to a coming section.
Research Questions
<jyasskin> The template instantiation suggestion first appeared at whatwg/
matatk So not the task force.
matatk I have a number of RQs grouped into two groups.
matatk Quantifying the problems that we are tryign to solve with these proposals.
matatk And then "Will makign these changes make pages more accessible?".
matatk A few questions in teh first group:
matatk How much are duplicate ids an issue? It would be interesting to see how much they can/ be/are in practice. Tooling, frameworks, provide mechanisms to mitigate against them.
matatk Does the fact that tooling provide those sort of mitigation atually solve the problem
matatk On the relative reerences side about quantifying the probelm: Is most HTML out there acutlaly amenable to relative referencing?
matatk I've go t two int he justifying the dev cost.
matatk Essentially what I've allued to before is that there are a lot of hidden costs that web devs wouldnt be aware of to adopting some of these solutions.
matatk Soem research on what the burden would be there.
matatk And maybe having established what these costs and benefits ouwld be, then looking at what benefits templates would have and other things worth exploring.
matatk The quetion is: do we all have agrreement that those are the questiosn we sould answer?
<Zakim> jyasskin, you wanted to push against research 😇
jyasskin From the proponetns: there is some furstration about being asked to unboundedamounts of research to justify working on a problem when not all proposals are held to that standard.
jyasskin Wehn we ask for research, what sort of design problems should it address/anser.
matatk I would like to have a discussion on how we can support the resarch as a community.
sarah Where is the actual dev pain for research?
sarah Single-file idrefs are easy, trivial. Passing ids betwen files is tricky.
<jyasskin> +1
janina This is very interesting and I'm having al ittle bit of trouble understanding where we think that the problem is. Duplicate ids are a proboem, they show up and cause many bad things. WE find ids very useful, however. I carefully put them...
janina To what degree is mindfulness in the content creation process a part of the probelm and how can tooling help with that? Somehow, the mashups can create thse probelms even from a single web domain. You throw putions of multiple content tracks into a particular web page, you may ineeed get them. Knowing where and how we get these duplicated
delivered to users would be helpful. Where do ehtey come from?
matatk We really should get to the point of what the shape of hte next part ofthis discussion should be.
PaulG I think Sarah hit the nail on the headhere. Duplicate IDs never bothered me as a dev. What Sarah mentioned was that in modern dev stacks, when you're building components in an isolated environemtn, then that si a huge pain point for devs. You hae
PaulG It gets complicated rather quickly.
PaulG Is this a problem in the abstraction frameworks? Could they have started with a11y first?
PaulG Duplicate IDs are possibly an issue. The ral motivator here is what Sarah is talking about. It is when the a11y specialist mentions to a dev: hey, this doesn't work,we need to fix this, and the dev doesn't know how to fix it even remotly. If we can focus research on that pain point...
PaulG Just centering on dupe ids will miss the mark since it doesn't solve their problem.
jyasskin I'm realising that I'm super uncomfortable with the assumption that everyone uses frameworsk to rgenerate HTML. It would be nice if HTML itself supported being able to just write it and say "here is a chunk I am going to repeat a bunch of times and it should act nicely within itself" without assuming everyone has a framework generating ids
for you.
sarah I'm curious to what degree people are using frameworks and to what degree those not using frameworks find this to be an issue.
sarah I think that the possibilities for what the problem is, the question I want answered: Is it the generation of ids? is it the communication across component boundaris of the uid you have generated?
sarah Will return with thte third thing.
matatk I will resist the temptation to say more things.
matatk WE are making progress here!
<Fazio> More discussion at TPAC?
<jyasskin> I think nearly everyone _does_ use frameworks, so my discomfort is a bit weird, but it feels like we should still have an HTML-level way to do these cross-links. Maybe just so that we're guaranteed that no framework forgets it.
matatk Yes, more discussion at TPAC.
matatk WE've already have some offers to thelp with the research.
<Fazio> Happy to see how we can help, Helix Opportunity
How to orchestrate/do the research
<Fazio> We have a globally robust user research operation
matatk I would just very briefly reiaterate that if we could quantify something of what the a11y people think are the challenges here, that would help. While a11y people are best placed to do that, I'm not sure that the burden should be entirely on the a11y people. Maybe we don't have the capacity to do all the answering.
<matatk> sarah
sarah I can go through the two component libraries and document the difficulties and hte jeopardy of using reative references. That I could knock out before TPAC.
Fazio WE coudl create that and make asurvey and see how many people share that problem.
matatk Interesting.
<matatk> +1 to survey about pain points (Sarah)
<jyasskin> +1
<JenStrickland> +1
sarah A survey that si not about "do you want this thing?" but rather "what are yoru pain points?" or "how's your pain?" would be mroe interesing and get more fine-grained results.
Next steps
Fazio We have the resources!
matatk The minutes of this will go out immediately.
matatk I will try to produce a refined outline about this. and be able to share it withe anyone interested.
matatk What do we want to do? A breakout on TPAC? Somethign else before or after?
Fazio Breakout session at TPAC?
jyasskin I think we shoudl do a reakout. The difficulty is doing a balanced structure.
sarah TPAC sounds good. Either a breakout session or an APA session.
matatk Thanks for all the suggestions. We'll look into the timetabling and all that.
matatk I can send out what I said before sometime tomorrow. We can then bake this into a description that can go into a breakout.
<Fazio> After this we need to give a Maturity Model Update please
<Fredrik> +1
<JenStrickland> +1
matatk Awesome!
matatk We're over!
matatk Thanks everyone!
matatk Expect more form me shortly.
Expect more from me chortling.