14:04:24 RRSAgent has joined #adapt 14:04:28 logging to https://www.w3.org/2023/12/05-adapt-irc 14:04:38 zakim, start meeting 14:04:38 RRSAgent, make logs Public 14:04:40 please title this meeting ("meeting: ..."), Sharon 14:05:32 meeting: APA WAI-Adapt Task Force Meeting 5 Dec 2023 14:05:32 rrsagent, make logs public 14:06:03 agenda? 14:06:04 present+ 14:06:04 chair: Sharon 14:06:42 agenda+ WKU for Accessibility Statement: What does an MVP of this use of WKU look like? 
 Link: https://github.com/w3c/adapt/blob/add-explainers/explainers/accessibility- meta.md
Future roadmap for this, Human v Machine readable
 14:07:02 agenda+ WKU for A11Y Reporting: What does an MVP of this use of WKU look like? 
Link: https://github.com/w3c/adapt/blob/add-explainers/explainers/accessibility-reporting.md
 14:07:18 agenda+ WKU vs. Destination: Similarities and Differences, will we need both
 14:08:45 agenda+ Report on Issue 240 Link: https://github.com/w3c/adapt/issues/240 14:08:45 https://github.com/w3c/adapt/issues/240 -> Issue 240 Could we build symbolic annotations with existing Web standards? (by DuncanMacWeb) [i18n-tracker] 14:09:10 agenda+ Continue the concept definition discussion
for 'symbol': adapt-symbol and "Composition or Concatenation" 14:51:58 Roy has joined #adapt 14:59:47 janina has joined #adapt 14:59:53 agenda? 14:59:57 present+ 15:01:45 Abhinav has joined #adapt 15:02:18 Russell has joined #adapt 15:02:23 present+ 15:02:33 matatk has joined #adapt 15:03:45 present+ 15:03:47 present+ 15:04:26 scribe+ 15:04:29 scribe+ 15:04:40 agenda? 15:04:45 https://github.com/w3c/adapt/issues/240 -> Issue 240 Could we build symbolic annotations with existing Web standards? (by DuncanMacWeb) [i18n-tracker] 15:05:08 zakim, next item 15:05:08 agendum 1 -- WKU for Accessibility Statement: What does an MVP of this use of WKU look like? 
 Link: https://github.com/w3c/adapt/blob/add-explainers/explainers/accessibility- 15:05:11 ... meta.md
Future roadmap for this, Human v Machine readable
 -- taken up [from Sharon] 15:05:34 matatk: Did preliminary testing for github discussions. RQTF not very hopeful 15:05:57 matatk: Discussions feature is enabled 15:06:28 matatk: Ack to Abhinav's questions, just not sure what mechanism we can accessibly use 15:06:45 matatk: Think it best we start with Abhinav's questions, and do so asynchronously 15:07:18 matatk: Asks Abhinav if any hot topic now? 15:07:30 AbhinavHappy to take up in github and next week's call 15:08:11 agenda? 15:08:17 https://github.com/w3c/adapt/issues/240 -> Issue 240 Could we build symbolic annotations with existing Web standards? (by DuncanMacWeb) [i18n-tracker] 15:09:19 matatk: Suggesting Info Arch is better monocre 15:09:25 zakim, drop agendum 3 15:09:25 agendum 3, WKU vs. Destination: Similarities and Differences, will we need both
, dropped 15:09:35 zakim, drop item 2 15:09:35 agendum 2, WKU for A11Y Reporting: What does an MVP of this use of WKU look like? 
Link: 15:09:37 ... https://github.com/w3c/adapt/blob/add-explainers/explainers/accessibility-reporting.md
, dropped 15:10:04 subtopic: Schema.org 15:10:49 Sharon: Notes we had issue on Schema, phps 2021? WE did look 15:11:03 Sharon: Notes her email 15:11:06 matatk: On list? 15:11:38 https://schema.org/accessibilitySummary 15:12:38 matatk: Notes used by some groups, esp Epub metadata re a11y as part of W3C specs 15:13:23 matatk: Notes entry for A11y Summary; is it epub? Phps 15:14:10 q? 15:15:02 janina: We would need a wider WAI conversation about what might go into this if it were to become the basis of an official W3C spec. 15:15:32 Lionel_Wolberger has joined #adapt 15:17:47 Abhinav: looking forward to discussing in GitHub 15:18:02 scribe+ 15:18:12 zakim, next item 15:18:12 agendum 4 -- Report on Issue 240 Link: https://github.com/w3c/adapt/issues/240 -- taken up [from Sharon] 15:18:13 https://github.com/w3c/adapt/issues/240 -> Issue 240 Could we build symbolic annotations with existing Web standards? (by DuncanMacWeb) [i18n-tracker] 15:19:06 Russell: No activity on the thread since last time. 15:19:26 q+ to ask if you were able to show Lionel the development you showed us last week 15:19:51 Lionel_Wolberger: Russell and I had a deep dive and I clarified certain critical concepts that we may need to discuss here. We are due to meet again. 15:20:34 ack me 15:20:34 matatk, you wanted to ask if you were able to show Lionel the development you showed us last week 15:21:24 present+ 15:24:06 Ref discussion from last week: https://www.w3.org/2023/11/28-adapt-minutes.html#t03 15:24:48 Russell: *showed us that some Bliss symbols, with their BCI IDs are in Unicode, but 3x as many are not (they're composed, so they're not included in Unicode) 15:26:35 Russell: A single character with an accent may be composed of a letter and an accent codepoint, which are combined. This can work well for display purposes. It makes software that does sorting, for example, more complex, because it has to be aware of the possibilities of multiple code points per character. 15:27:20 Russell: Most of the 1400 in Unicode are symbols in their own right. [ref last week for the other numbers] 15:27:46 Lionel_Wolberger: I suspect the heart of the matter here is our original intention, that the web has all types of symbols, icons, pictrograms - non-character representations of meaning.] 15:27:54 s/meaning.]/meaning./ 15:28:04 Lionel_Wolberger: The space of these is too huge to be enumerated. 15:28:15 ... The concepts we want to represent are also fluid over time. 15:28:55 ... Reminding us that what we want: rather than someone on a web page putting an emoji of someone slapping their face, or a hamburger, they intended to put across the concept of food. 15:29:13 ... This may be better represented by a different symbol in a different cultural context, for example. 15:29:39 ... We need a way to encode the concept, so users around the world can see the concept rendered in the appropriate way. 15:29:46 q? 15:29:48 ... So we need a solid dictionary of these types of concepts. 15:30:01 ... We have a valid proposal that is a reasonably good normative dictionary. 15:30:06 ... We have not yet found a better one. 15:30:14 janina: Including Unicode 15:30:21 Lionel_Wolberger: Agree. It's not a dictionary of that type. 15:30:22 q+ 15:30:36 Lionel_Wolberger: It's an amazing, valuable community of people who curate unambiguous characters and symbols. 15:30:39 q+ 15:31:04 ... It's not a dictionary that's sufficiently robust and descriptive dictionary to serve as an index to any reasonable concept that one may want to expect. 15:31:41 Lionel_Wolberger: Some feel that Unicode provides a sufficient index of concepts. 15:32:03 ack matatk 15:32:31 matatk: Appreciate Lionel's approach, want to play devil's advocate? Why not Unicode? 15:32:57 matatk: unicode may be better? Because you can put existing symbols together, but not in ours 15:33:23 matatk: Though I ack ad hoc is not all that helpful 15:33:33 matatk: was playing devil's advocate there, but good to explore 15:33:46 ack Russell 15:34:38 q+ to capture Russell's geography of symbols as locales 15:34:58 Russell: (1) in terms of locales for people who use symbols: it's not just location in the world, but also the symbol sets that are developed, especially the pictorgraphic ones, there are multiple representations of the same concept in some sets. The reason is that you might use a different picture to represent a concept for an adult vs a child, or a particular child. 15:35:03 rrsagent, make minutes 15:35:04 I have made the request to generate https://www.w3.org/2023/12/05-adapt-minutes.html matatk 15:35:24 Russell: Some of these sets will have maybe tens of drawings to represent the same concept. 15:36:11 Russell: (2) Bliss is about to be part of Unicode but is the only symbol system that's near being part of Unicode. I have come across other systems that are going for a more structured approach, e.g. Widget, the 'schematic version' 15:36:15 rrsagent, make minutes 15:36:16 I have made the request to generate https://www.w3.org/2023/12/05-adapt-minutes.html matatk 15:36:46 Russell: I don't see Unicode being the dictionary a possibility in the near future at least. 15:37:23 ... Using the unicode version of bliss for the representation of our concepts: it's do-able but I advocate against it because of implementation issues that are there. People on Issue 240 have offered some suggestions. 15:37:55 Russell: The BCI IDs are a single number for a single concept [even compound ones] 15:38:10 ack janina 15:38:10 janina, you wanted to capture Russell's geography of symbols as locales 15:38:14 ... But if going with Unicode, there are complications 15:38:48 janina: Re locale: a set of people, a community, could be defined as a locale? Unicode isn't giving us that. 15:39:01 Russell: In some cases you may have a locale of 1. 15:39:29 Russell: A set (maybe Widget) has 50 different representations of dog. Because some people who use symbols want it to look like their doc. 15:39:32 q+ 15:39:37 s/their doc/their dog/ 15:39:43 ack Lionel_Wolberger 15:40:11 Lionel_Wolberger: Appreciative of Russell's deep dive, and the rigor and normativity that I see in Unicode. 15:40:21 ... I'd like to answer the problem statement: can we use Unicode as the dictionary? 15:40:29 ... As I thought about it, I get lost so fast. 15:40:31 q_ 15:40:35 q+ 15:40:51 Lionel_Wolberger: If you look at the emoji description page for hamburger, and it's long! 15:41:08 q+ 15:41:13 ... It reduces to one concept but it isn't one to start with. Which Unicode? Which font? 15:41:20 ack matatk 15:42:03 matatk: thinking cardinalities of different constituents -- Our symbols for adults; our symbols for chn; --but ... 15:42:24 matatk: every time we do that we tweak our mental model but doesn't seem terribly helpful 15:42:40 matatk: see impass with ourselves and with the sw people on thread 15:42:57 matatk: should we try end to end mappings for a few symbols to see whether it works 15:43:08 q+ 15:43:18 q? 15:43:27 matatk: That would show us how it might actually work 15:43:39 ack Russell 15:44:15 Russell: Great idea. I have started a document that I passed to Lionel_Wolberger, maybe will pass to everyone. Sort of doing this already, but could fill it out to show all the cases 15:44:21 ... of using the different systems. 15:44:44 Russell: Also: Lionel_Wolberger asked 'which Unicode?' - you're really saying which language do we want to use to specify these things. 15:45:06 ... You run into all sorts of problems such as synonyms, hominyms, etc. 15:45:25 Russell: I'm talking about using the Bliss version of Unicode vs the IDs 15:45:55 ack janina 15:45:58 Russell: I will walk through the use cases and describe the work, and problems with each. 15:46:21 janina: Concerned that if we go with what matatk is proposing, we use it as an opportunity to show what's missing when going to Unicode. Sounds like a lot missing here. 15:46:42 ... If we really have 3x as many concepts defined in BCI vs Unicode, we need to talk about that. 15:46:45 q+ 15:46:52 q+ 15:47:05 ack matatk 15:47:31 matatk: Could see would still end up with our registry since not all concepts are in unicode 15:47:43 matatk: Could do unicode only; 15:47:54 matatk: Our registry with unicode but not BcI 15:48:01 matatk: Our registry as currently proposed 15:48:39 matatk: Not intending to put this all on Russell! phps no more than 3; so the rest of us could learn and take a few on 15:48:47 matatk: You'll have to correct our work, doubtless! 15:49:03 q? 15:49:15 ack Russell 15:50:13 Russell: When we say there are 3x as many concepts in BCI as in Unicode... we can express every one of the BCI concepts in Unicode. Instead of a single Unicode number, they'd be combined characters. 15:50:24 janina: But the authoring community would have to figure the concepts on their own. 15:51:14 Russell: And if something comes up that's not in our concepts, a new symbol will have to be created. Easier to do in Bliss than Unicode. Still barriers to making a new ID (though you don't have to create the symbol to create an ID). 15:51:36 q? 15:51:55 matatk: rthink we're teasing out something very important -- we need this clearly documented 15:52:23 q+ 15:52:25 Lionel_Wolberger: Schema.org came up - could be that in the linked data space there's a rich semantic vocab that matches the markup we're looking for. 15:52:52 q? 15:53:15 ... I spent time in the world of tuplets. Is the list of references/concepts that we want to provide the community. Is it every word in the dictionary that's not a proper noun? 15:53:21 q- 15:54:20 ... Have we gone to Bliss because, by the nature of their endeavour, they've identified a big subset of vocab that should be used. 15:54:27 matatk: too simple and that makes it hard again, one needs to build up the model in one's head 15:54:34 Russell: The most complete set I know of is ConceptNet. 15:55:15 Lionel_Wolberger: All markup curated in W3C space is a list of concepts that is tremendously useful. 15:55:28 Lionel_Wolberger: We should find a minimum body of concepts. 15:55:38 scribe+ 15:55:59 ... a minimum but sufficiently rich list of concepts to support Adapt's Symbol markup goal 15:56:11 Russell: I don't think there's anything stopping us having both the Unicode and BCI ID representations in the registry 15:56:22 matatk: Ack 15:56:47 matatk: only if it controls energy expended 15:56:57 s/only/even only/ 15:58:31 rrsagent, make minutes 15:58:32 I have made the request to generate https://www.w3.org/2023/12/05-adapt-minutes.html matatk 15:58:38 present- 15:59:57 s/only if it controls energy expended/Think we should have as few possible ID schemes as possible (even if only marginal extra energy requirement, it is there for a long time...)/ 16:00:15 rrsagent, make minutes 16:00:16 I have made the request to generate https://www.w3.org/2023/12/05-adapt-minutes.html matatk 16:02:12 s/even Think/Think/ 16:02:14 rrsagent, make minutes 16:02:15 I have made the request to generate https://www.w3.org/2023/12/05-adapt-minutes.html matatk