07:53:09 RRSAgent has joined #apa 07:53:13 logging to https://www.w3.org/2025/11/12-apa-irc 07:53:13 RRSAgent, do not leave 07:53:14 RRSAgent, this meeting spans midnight 07:53:14 RRSAgent, make logs public 07:53:16 Meeting: Fixing HTML element references (IDREFs) on the Web 07:53:16 Chair: Matthew Atkinson, Lea Verou, Jeffrey Yasskin 07:53:16 Agenda: https://github.com/w3c/tpac2025-breakouts/issues/61 07:53:16 Zakim has joined #apa 07:53:17 Zakim, clear agenda 07:53:17 agenda cleared 07:53:17 Zakim, agenda+ Pick a scribe 07:53:19 agendum 1 added 07:53:20 Zakim, agenda+ Reminders: code of conduct, health policies, recorded session policy 07:53:20 agendum 2 added 07:53:20 Zakim, agenda+ Goal of this session 07:53:21 agendum 3 added 07:53:21 Zakim, agenda+ Discussion 07:53:21 agendum 4 added 07:53:21 Zakim, agenda+ Next steps / where discussion continues 07:53:22 agendum 5 added 07:53:22 Zakim, agenda+ Adjourn / Use IRC command: Zakim, end meeting 07:53:22 agendum 6 added 07:53:23 breakout-bot has left #apa 08:26:34 ZoeBijl has joined #APA 08:27:59 tantek-projector has joined #apa 08:28:47 Jamie has joined #apa 08:29:25 matatk has joined #apa 08:29:37 zakim, start meeting 08:29:37 RRSAgent, make logs Public 08:29:38 please title this meeting ("meeting: ..."), matatk 08:29:42 meeting: Fixing IDREFs 08:29:47 present+ 08:29:50 present+ 08:29:50 chair: matatk 08:29:56 = 08:30:00 s/=// 08:30:01 kizu has joined #apa 08:31:21 present+ 08:31:24 present+ 08:31:27 scribe: ZoeBijl 08:32:05 q+ 08:32:07 q- 08:32:12 present+ 08:32:18 alisonmaher has joined #apa 08:32:24 present+ 08:32:28 sarah has joined #apa 08:32:29 present+ 08:32:32 front-endian-jane has joined #apa 08:32:36 daniel-mac has joined #apa 08:32:39 present+ 08:32:41 present+ 08:32:47 zcorpan has joined #apa 08:32:48 MA: let’s get started 08:32:50 present+ 08:33:05 present+ 08:33:06 present+ 08:33:07 LeoLee has joined #apa 08:33:08 rrsagent, make minutes 08:33:10 I have made the request to generate https://www.w3.org/2025/11/12-apa-minutes.html ZoeBijl 08:33:18 lea has joined #apa 08:33:18 present+ 08:33:18 present+ 08:33:18 present+ 08:33:18 sidvishnoi has joined #apa 08:33:23 nigel has joined #apa 08:33:27 jamesn has joined #apa 08:33:31 we have a presentation that we all worked on 08:33:50 we got input van alice, sarah, ??, and jefery 08:33:55 s/??/noam/ 08:33:55 s/??/Lea, Noam/ 08:33:55 some important information, borrowing from florian 08:34:05 you’ve all agreed to the CoC 08:34:17 s/jefery/Jeffrey/ 08:34:19 some ground rules 08:34:22 no qq+ 08:34:25 we want problems 08:34:27 not solutions 08:34:34 we are going to listen to the problems 08:34:38 we are going to collect them 08:34:39 fantasai has joined #apa 08:34:57 not judge them 08:35:05 all insights are valid 08:35:11 oliverdunk has joined #apa 08:35:15 goals: 08:35:34 1. find concensus regarding the painpoints of IDRefs 08:35:34 2. ?? 08:35:34 3. ?? 08:35:49 (see slides) 08:36:00 is there a link to the slides? 08:36:08 non-goals: (see slides) 08:36:12 I have made the request to generate https://www.w3.org/2025/11/12-apa-minutes.html fantasai 08:36:20 agenda+ what (if anything) needs fixing? 08:36:27 agenda+ prior art 08:36:54 slides: https://raw.githack.com/w3c/apa/main/presentations/2025/TPAC/fixing-idrefs/index.html 08:37:00 thank you! 08:37:09 take up next item 08:37:17 zakim, show agenda 08:37:17 I see 8 items remaining on the agenda: 08:37:18 1. Pick a scribe [from breakout-bot] 08:37:18 2. Reminders: code of conduct, health policies, recorded session policy [from breakout-bot] 08:37:18 3. Goal of this session [from breakout-bot] 08:37:19 4. Discussion [from breakout-bot] 08:37:19 5. Next steps / where discussion continues [from breakout-bot] 08:37:19 6. Adjourn / Use IRC command: Zakim, end meeting [from breakout-bot] 08:37:20 7. what (if anything) needs fixing? [from ZoeBijl] 08:37:20 8. prior art [from ZoeBijl] 08:37:27 take up item 7 08:37:41 repeated content requires rewriting IDs 08:37:51 we have a lot of use cases around SVG 08:38:04 For anyone who wants to dig in, here is the actual author quotes: https://github.com/WICG/idrefs/blob/main/research.md 08:38:10 also combining content from multiple sources 08:38:17 s/here is/here are/ 08:38:22 more general themes from developers 08:38:27 saku has joined #apa 08:38:42 no platform support for generating/managing uinique ids 08:38:43 ??2 08:39:04 it’s unergonomic usig IDs 08:39:18 in general what we’re concerned about 08:39:21 on the DX side 08:39:24 s/??2/not ergonomic to associate labels with inputs using ids because authors often just don't bother creating accessibility issues/ 08:39:27 finding the pain points 08:39:33 to do that we need to find out what they are 08:40:29 MA: ?? 08:40:42 there’s also a runtime permance hit potentially 08:40:45 those are some general things 08:40:53 take up next item 08:41:01 take up item 8 08:41:22 useID() is a hook function provided by component frameworks Vue and React 08:41:29 * shows example of how to use* 08:42:16 you can also do something similar with prefixing 08:42:36 basically have an id for a form, and then prefix all descendants with the form id 08:42:43 s/React, generates a unique id on demand/ 08:42:53 present+ 08:43:10 s/React/React, generates a unique id on demand/ 08:43:25 we also have scoped ID rewriting 08:43:39 similar to useID 08:44:27 zakim, clear agenda 08:44:27 agenda cleared 08:44:37 we also have selectors 08:44:38 tantek has joined #apa 08:44:49 you can use selectors to express relationships 08:45:03 it can be useful for things relatively nearer in the DOM 08:45:07 like labels and inputs 08:45:27 *shows htmx example with hx-target* 08:46:05 q+ to ask which ATs don't label association through nesting 08:46:28 some questions… 08:46:36 these are just suggestions 08:46:48 - what are the pain points? 08:46:57 - is the problem generation or. communication? 08:47:07 - what appreoaches have been used in frameworks or tooling 08:47:09 - ??1 08:47:12 - ??2 08:47:13 - ??3 08:47:16 s/communication/communication across boundaries/ 08:47:18 - what are performance costs 08:47:29 - how would any particular change benefit users 08:47:34 - ??4 08:47:42 - some research notes 08:47:55 s/??1/What are the (UA, AT) development costs? 08:47:59 (these are in the slides too) 08:48:09 s/??2/could devtools or offline tooling be improved to help the problem? 08:48:17 s/??3/What are the performance costs? 08:48:22 broad themes from developers were summarised before 08:48:43 but there’s more in the appendix of this slide deck 08:48:55 reminder: our goal is to come to concensus on the problem space 08:48:56 s/??4/What other constraitns would you want on the potential solution space? 08:49:02 that might require research 08:49:22 we’re here to listen 08:49:33 there’s a WICG repo for this 08:49:35 q? 08:49:39 q? 08:49:49 ack zcorpan 08:49:50 zcorpan, you wanted to ask which ATs don't label association through nesting 08:49:51 present+ 08:49:55 zcorpan: you showed an input and label in the examples 08:50:03 i know there are other elements that use id and association 08:50:12 theparkagen has joined #apa 08:50:21 the main problem with nested input and label that it’s not supported in all AT 08:50:28 my question is in, which AT is this broken? 08:50:40 if this is fixed, will that take away the problem? 08:50:48 MA: input + label is a big problem for sure 08:50:54 q? 08:50:55 but there are many others that need work 08:50:57 q+ 08:50:58 q+ to ask about if we are assuming the developer is cooperating 08:51:01 even if the wrapping worked 08:51:07 that’s not how ARIA stuff works 08:51:26 https://www.tpgi.com/should-form-labels-be-wrapped-or-separate/ 08:51:40 AB: it’s a long standing problem with Voice Control 08:51:54 also for Dragon Naturally Speaking 08:52:07 James, do you know why Voice Control doesn’t work with label nesting? 08:52:12 JC: it’s a bug 08:52:18 present+ 08:52:19 AB: it’s a long standing issue 08:52:24 first heard this in 2011 or so 08:52:29 they should absolutely be fixed 08:52:33 q? 08:52:36 whetehr we should hold our breath i don’t know 08:52:46 a lot of these tools are not frequently updated 08:52:53 q+ 08:52:53 zcorpan: ?? 08:53:03 doesn’t mean we shouldn’t do anything else 08:53:08 q+ to flag that any solution to this problem needs to take mutations into account and the performance impact of managing those 08:53:10 ack fantasai 08:53:10 fantasai, you wanted to add prior art from CSS 08:53:24 fantasai: css has started to have features that need to link two elements together 08:53:31 like ?? 08:53:36 another is anchor positioning 08:53:42 so css has properties that name a box 08:53:50 for the purpose of anchor position or animations 08:53:57 it’s similar to the problems with IDs 08:54:03 we knew authors would repeat a pattern 08:54:08 like comments in a form 08:54:20 we have for each of the places where we have naming 08:54:32 we have a scoping property 08:54:32 s/zcorpan: ??/zcorpan: To anne's previous point, if the problem is that there's a bug in implementations, introducing a new thing that needs to be implemented may not be the best way to solve it, instead push for the bug to be fixed./ 08:54:36 it traps those ids 08:54:40 you can’t go higher than these 08:54:50 q? 08:55:02 that’s an existing prior art that’s probabbly relavant here 08:55:05 ack lea 08:55:10 lea: wanted to reply to simon 08:55:12 qv? 08:55:18 with problems relevant to ?? 08:55:26 nigel has joined #apa 08:55:29 we have a lot of upcoming features that use id ?? 08:55:36 nigel has joined #apa 08:55:40 s/id ??/id references/ 08:55:48 the problems would not go away 08:55:51 q? 08:55:57 q+ 08:56:07 ack oliverdunk 08:56:07 oliverdunk, you wanted to ask about if we are assuming the developer is cooperating 08:56:18 s/like ??/like scroll animations, tying scrolling of one element to animation of another element/ 08:56:25 OD: it seems we’re mostly talking about making it easier for devs to make these associations? 08:56:28 if a dev has a heading 08:56:34 and wants to make a link to the heading 08:56:41 q+ text fragments 08:56:44 it would be useful if there was a browser way to do that 08:56:52 s/a scoping property/a scoping property, which takes values none | all | list of IDs 08:57:13 MA: can you clarify stable reference? 08:57:20 OD: you can’t have ?? 08:57:21 ack text 08:57:24 ack fragments 08:57:29 q+ zcorpan 08:57:34 ack noamr 08:57:46 noamr: i asked developers about this 08:57:56 and they said i would never use this because it uses idrefs 08:58:09 they’re afraid of importing things 08:58:22 there’s a group of people that use IDrefs without knowning 08:58:39 ?? 08:59:00 there’s a sound in the room that if it breaks AT we can’t touch it 08:59:14 q+ 08:59:14 i don’t understand how the relationships is between browser and AT 08:59:22 q+ 08:59:30 is that constraint here to make it easier to appraoch a solution? 08:59:48 MA: i think it depends what the problems are 08:59:56 but it migth be a reasonable questions 09:00:07 you’ve probably experienced pain for using IDREFs 09:00:23 it’s reasonable to assume that bugs in AT will be fixed 09:00:29 ack Jamie 09:00:29 Jamie, you wanted to flag that any solution to this problem needs to take mutations into account and the performance impact of managing those 09:00:33 JT: in terms of contraints 09:00:48 any solution to the problem will not be can i get the element 09:00:56 but more also if someone mutates the tree after that 09:01:01 what’s the performance hit 09:01:06 q+ to ask noam for clarification wrt resolving between the browser and AT (and bypassing the normal platform API as the in-betweener?) 09:01:09 wanted to raise that as a constraint 09:01:18 ack frehner 09:01:20 ack fantasai 09:01:24 AF: it does feel weird to come with problems 09:01:28 ack frehner 09:01:32 one problem that you… 09:01:47 was also solving was the issue of people generating IDs with math.random 09:02:01 but than the server side generated ID wouldn’t be the same as the client side id 09:02:09 second problem is in my case right now 09:02:17 s/you.../useID/ 09:02:27 parts of our webpage allows 3rd party developers to embed code 09:02:48 but we’re scared of them adding a button/popup outside of the sandbox we’ve given them 09:02:55 so we need to?? 09:03:10 AK: ?? 09:03:22 AF: ?? 09:03:30 there’s a lot we tried to do there 09:03:50 there are concerns on how to make sure their IDs don’t conflict with ours 09:04:08 *further discussion* 09:04:13 qv? 09:04:13 s/AK: ??/AK: You already have to sanitize the IDs to prevent various security issues, no?/ 09:04:18 one more problem… 09:04:30 to noamr’s point earlier 09:04:39 we have a lot of issues in our react codebase 09:04:53 devs think of their components as scoped 09:05:02 we frequently found, that because rendered code isn’t scoped, IDs conflict 09:05:11 s/AF: ??/AF: Yes, we run it in a safe environment, sanitize the output, etc, but there are still problems around IDREFs that could be improved/ 09:05:19 MA: thank you! 09:05:32 what i’m hearing is a communication use case we hadn’t heard yet 09:05:35 q? 09:05:37 and something else around scoping 09:05:41 but harder than we were thinking of 09:05:41 q? 09:05:48 ack zcorpan 09:05:51 I also hear that we're not going to notice these collisions with HTML Archive data, since they're generated client-side 09:05:57 SP: wanted to respond to the linking in a heading 09:06:06 and the way browsers do this is with ?? 09:06:14 MA: that’s some prior art we didn’t mention 09:06:15 s/??/text fragments/ 09:06:21 SP: we could use it for element association 09:06:24 q+ to note the use-case of table td element "headers" attribute IDREF *list* to IDs of th elements 09:06:31 MA: it’s good to know about the prior art 09:06:32 q- 09:06:36 I have made the request to generate https://www.w3.org/2025/11/12-apa-minutes.html fantasai 09:06:45 ack sarah 09:06:52 q+ to also note https://developer.mozilla.org/en-US/docs/Web/HTML/Reference/Elements/td#headers 09:06:53 SH: another pain, or annoyance 09:07:01 we make component libraries 09:07:06 so reusable things 09:07:18 lots of places where an author will need to reference a specific component 09:07:37 say you have an input and a button, but you want to add an error state 09:07:41 +1 to Sarah's problem 09:07:47 you need a consumer of the sandbox component to reference into it 09:08:03 when we generate a generated ID you need to pass it around 09:08:08 if we could ?? as strings 09:08:27 it seems very solveable 09:08:28 q? 09:08:33 ack me 09:08:33 jcraig, you wanted to ask noam for clarification wrt resolving between the browser and AT (and bypassing the normal platform API as the in-betweener?) 09:08:35 and in that case ?? 09:08:36 ack jcraig 09:08:54 JC: clarifying question to noamr 09:09:02 s/?? as strings/have static strings/ 09:09:05 negotiating a reference between teh browser and AT 09:09:09 JRJurman has joined #apa 09:09:10 i misunderstood what you meant 09:09:26 noamr: was asking the questions that ?? 09:09:32 where are the issues in the architecture? 09:09:38 is it between the AT and the OS? 09:09:42 between the AT and browser? 09:09:49 felt like i didn’t understand 09:09:56 JC: i think i can try and clarify 09:10:11 on apple paltforms we try to have the AT talk to the platform API directly 09:10:18 and be as browser unaware as possible 09:10:33 s/paltforms/platforms/g 09:10:33 then the browser engine can wend to the accessiblity API directly 09:10:53 q+ 09:10:54 i guess to your questions, where are the bugs 09:10:54 s/wend/vend/ 09:10:58 all of the above 09:11:08 q+ to flag that while most AT uses browser accessibility engine, some do not 09:11:36 MA: with respect to that one, i don’t know if this is current, but the scope of some of the AT vendors is to ?? 09:11:50 sometimes it’s just a bug 09:11:52 q? 09:12:15 AT = assistive technology 09:12:33 q+ for possible link to extensions 09:12:41 ack tantek 09:12:41 tantek, you wanted to note the use-case of table td element "headers" attribute IDREF *list* to IDs of th elements and to also note 09:12:43 ... https://developer.mozilla.org/en-US/docs/Web/HTML/Reference/Elements/td#headers 09:12:52 TC: to add to the ?? problems 09:13:00 i’ve had some challenges with the IDs list 09:13:07 int he ?? 09:13:22 that’s a one to many IDREFs to ID relationship 09:13:29 i don’t know how ATs handle that if at all 09:13:32 MA: quite well! 09:13:42 JC: that’s been around for about 30 years… 09:13:45 ack jyasskin 09:13:58 s/the ?? problems/the IDREF problems 09:14:01 s/int he??/in the headers attribute of the td element/ 09:14:06 jyasskin: i think we ahev this impresion that a lot of bugs take a long time to fi 09:14:22 s/fi/fix/ 09:14:24 we have ?? where it’s all in the browser 09:14:33 if we try to add something to the aria references 09:14:38 is that something we can just do 09:14:43 s/int he ??/in the 'headers' attribute of the td element where it's a list of IDREFs that reference IDs of th elements/ 09:14:43 or does it need AT vendors? 09:14:48 q? 09:14:49 JT: it depends 09:14:51 ack Jamie 09:14:51 Jamie, you wanted to flag that while most AT uses browser accessibility engine, some do not 09:14:54 *everyones favourite answer* 09:15:15 dare i say most AT use the browser accessibility engine 09:15:24 there are some AT that do not do this 09:15:35 present+ 09:15:53 i guess it doesn’t matter what we do in the browser 09:16:02 those AT will not support it 09:16:13 ack oliverdunk 09:16:13 oliverdunk, you wanted to discuss possible link to extensions 09:16:20 OD: there’s an interestin case in extensions 09:16:26 if you right-click on an element 09:16:35 and there’s an extension context 09:16:43 you can pass the id of the clicked element 09:16:57 that API is supported in Gecko 09:17:11 not supported in Chrome 09:17:20 MA: clarification question?? 09:17:28 OD: we can generate our own ID for that 09:17:35 but if there was some standard to that we would sue it 09:17:36 node_id 09:17:37 q? 09:17:38 s/sue/use/ 09:17:43 zakim, who is here? 09:17:43 Present: matatk, ZoeBijl, jyasskin, kizu, noamr, alisonmaher, alice, frehner, bkardell, masonf, zcorpan, kzms, front-endian-jane, LeoLee, lea, plinss, hober, tantek 09:17:44 s/sue/use/ 09:17:47