Meeting minutes
matatk: Summarises work from yesterday
… Plan for today; recap document status and outline what is required
… Planning communications and reviewing matrix
… Session later with symbols expert
… More matrix work time later which might not be required. Time could be used for Explainer or Examples
… For the Explainer we explored what was needed yesterday - it is more than tweaks so may be better done out of meeting
… We could also split up into groups to work on separate topics
… Also, what else is required, for example updating the Registry document
Document statuses and needs
Lisa: Would be good to look at worked mock ups and examples
… This will be required later when going through CR
matatk: Lisa, there is a page on the wiki where Russell discussed issues he faced when making a font
https://
… This would be helpful as a NOTE track document and this would be the basis for such a thing
… Mock ups is important. There are other examples form the demo authoring tool.
… Annalu also suggested having some real world examples
janina: Could use the WAI homepage
Lisa: It would be good to outline what materials are going to be delivered so people have advanced notice
<Lisa> need to step away (physiotherapy)
kevin: Could use the WAI website for some of the narrative description
matatk: Could have something in WAI website that directs people to the right information
… Official document in TR space and WAI website can provide a more filtered view
kevin: That is worth exploring further
matatk: Lionel could you look at the 'shop window' material and see what needs changed
Lionel_Wolberger: Immediate goals is to prepare for our two meetings today
… Explainer may help with those?
matatk: I think largely we are prepared for the meeting today. The work on the matrix has been shared with folks
… Others involved are aware of the history so may not need much more upfront information to be presented
janina2: It might be worth exploring what objections might be raised
matatk: We should formalise how we are using the time later today
Lionel_Wolberger: I think it is worth spending some time working out our responses to those objections
… Would also be worth exploring language
kevin: Worth doing first draft of glossary
matatk: Between 1 and 3 - spend hour digesting information from EA, 30mins Planning further communications and 30mins working on glossary.
Lionel_Wolberger: We also determined that you can have multiple <rt>'s presented horizantally
Lionel_Wolberger: If an author chooses to add a symbol to a page. There is a subtlety in the order of events: the author will chose a symbol and then think of the concept?
matatk: The authoring tool could/should be suggesting symbols for concepts in page
janina2: What we need to avoid is people using aesthetics to chose for a symbol
Lionel_Wolberger: So the subtlety is that people are choosing symbols with string identifiers of the concepts. Is the string a Unicode code point?
matatk: As in the thing that goes into the markup?
Lionel_Wolberger: Yes
Lionel_Wolberger: Yes. So that Unicode code point I can look up in the Adapt Symbol Registry maintained by W3C as a normative source of concept code points in the Authorized Vocabulary
… I am going to consume and render that page. I use a browser that has the symbol and a code point on the page
matatk: No. A Unicode code point is just a number. It isn't rendered.
Lionel_Wolberger: But when I authored I chose a symbol and a code point?
matatk: The authoring tool at minimum might just give a text concept which might not even show the symbol itself.
… Users with Bliss font would then see their associated symbol
Annalu: If you have a font and you type a word and you change the font - you may get jibberish unless you change it to a Latin font in which you just get a different script.
… What we are looking to do is have the same symbols when we change the font
janina2: I think what we need to ensure is communicated is that the author is not choosing the symbol to use, just the concept
Lionel_Wolberger: Currently, I wish to decorate a sentence with a symbol. I choose a symbol based on what I think will visually work.
… What we are saying is that we are now saying that chosen pictograms are not to be defined by the author
janina2: This is not personal expression like emojis. This is expressing a concept in the users perspective.
Lionel_Wolberger: I think that the author is also selecting the symbol
matatk: They might be in practice but people shouldn't be doing this
Lionel_Wolberger: That is a big change from how people are currently doing this?
janina2: The reason we will have to explain to authors is that we will be more about identifying symbols that are part of a 'language'
matatk: We can replace <ruby> with discussion on this later
… Note that I don't think this is an issue for any of our experts
Lionel_Wolberger: To give an example, I have written something about it being sunny today and I write the Hebrew symbols for sunny. When I type these on my keyboard it types the right code point for this charatcer
… This code point is then rendered appropriately across all devices. I then also add an image of the sun to aid communication.
janina2: I think this is fine except where choosing an ARASAAC symbol which has additional meaning
Lionel_Wolberger: But in the future when I paste in the concept it pastes in the concept code
matatk: We should do a mock-up of an authoring tool that lets the user look up a concept, and shows what it *might* look like - by showing it in a number of symbol sets.
kevin: If selecting some text, like 'sunny day' the authoring tool would then pick the right concept and symbols would appear, as the tool put in the code point(s) for the author.
janina2: I think Lionel_Wolberger wants to leverage the interest on emoji
kevin: This is separate to the contextual info that Adapt layers on top.
Lionel_Wolberger: What it author wants to include an actual image of a symbol in the doc - can it be ignored?
kevin: If they put it in the doc, it's document content. It stays there.
matatk: Examples are important to show this. Selecting the right examples will allow for good demonstrations. We should probably include an example of this scenario
… Another thing that we could do is we could add some symbols to some text that includes emojis
… Is this now a good time for Annalu to share her example?
Annalu: [Demos how to do conjunctions]
… Bliss AV 12378 maps to 'animal, beast'. Then Bliss AV 9011 maps to 'indicator_(plural)'
… Indicators generally sit half a square above the presentation square and is presented in a quarter sized square itself
… So technically when I writing Bliss I put the indicator in the middle of the symbol being modified
… This needs to be handled in the font presentation
… So 'animals' does not appear in the AV.
… What I do is put the indicator first followed by the thing being modified
Ruby code: <ruby>plural_ind<rt>12345</rt></ruby>"< ruby>animal<rt>6789</rt></ruby>
kevin: Could also do - <ruby>animals<rt>12345<rt>6789</rt></ruby>
Lionel_Wolberger: Now we need to think about how this is presented
matatk: Don't need to worry about whether this is the approach that implementors will support or not. If they do, great, if they don't then we can explore more
matatk: Annalu this is extremely helpful as it makes the concepts that are being explore much clearer
… When we do best practice we will need to think about what examples
Annalu: This is quite complicated to look at and may be better suited for a version 2
… Creating an authoring tool that would manage this would be an excellent project
matatk: Another challenge is that when new web formats then we try to be as permissive as possible. We need to be careful that we are not handing people ways to break things for themselves.
… The authoring tool could help you do this but the individual is unlikely to know enough about Bliss in order to do conjunctions
… Some sort of expert system could help with this
Annalu: Yes, so it could present how to do plurals for authors
Lionel_Wolberger: So the scratch pad I use adds decorations to rows to aid comprehension. For example a green checkmark is added to pros.
Expert session
matatk: Q:How 'localisation' of symbols - e.g. to specific groups or regions (within a wider country/symbol set) - currently works.
<Lisa> Lisa: pulling out EA's update for coga https://
EA: AI presents a great opportunity.
matatk: There may be a symbol set for a country and there might be regional variations, and within that there may be personalised changes to the symbol set.
… In high level terms, how does this work?
EA: There are basic open-licensed symbol sets
… These sets are very often not a full vocabulary. They have a good core vocabulary of 500-2000 symbols.
… There are variations based on age, targeted for children or adults
… There are then collections associated with country themes of 2-300 symbols
… They may be added to over time. As a collection they would mirror another symbol set so that they can be used with a wider symbol set
… Authors have control of their symbol set. There is the facility to do definitions. There is ConceptNet which allows variations on a symbol to be managed, for example adding 'seek' to 'look'
… Can also manage the images themselves but our preference is to have SVGs
… We are building a system to have bulk upload SVGs and then allow for the download of PNGs
matatk: That localizaiton can be and usual is separate from the parent symbol set?
EA: Yes, and is totally controlled by the people who made it locally
… The designs are a stylised versions of the wider symbol set. So a stylised version of ARASAAC
… All we do is facilitate the process of making symbols in their own language and cultural context
… We are very careful with ARASAAC as they are run by a government organization. They are largely happy with how collections are managed though
matatk: Useful to add 'collection' to our glossary
EA: A symbol set means that there is a core vocabulary (top 500-1000 words) and fringe words pertinent to the group being worked with
… For example, localization might have a suitable currency symbol
janina: The reason to tease this out is to see whether any of this 'personalization'
EA: There was a question on social media as to why there was no symbol for cerebral palsy. We made some images for this but received no response.
… Part of the issue is that most images come from the relevant group
… If you want to add an image to a collection you can also keep it private; for example a family dog or parent images
… So there is personalization going on but it is harder to represent
janina: Where people are uploading or sharing, I presume it is much more formulated how it is integrated?
EA: Not at the moment but it is an area that we are exploring
Annalu: When someone has personal images on their communication board or device, I don't think that necessarily impacts on the global representation of what we are trying to do here.
… Should we be in the future be giving the option of local rendering for people.
EA: Interesting question.
… What we are thinking of doing because of localization and poor internet access is allowing symbols to be retrieved or added from a local drive.
… If you have text that is summarized does the user have the skills and cognition as well as the ability to find the right image
Annalu: Yes, and there may be some people who do not have the ability or skills to navigate to the material in this way
EA: Yes, this might be a 'gatekeeper' role who is selecting these symbols
Annalu: yes, this is one of our use cases to consider
Lisa: The old implementation did support overriding from a local or personal symbol set
… So for example a siren as a critical service, load this as a local picture with additional informaiton
… I think this does have value but not sure if it is a blocker
… Not sure about AI in critical serivce delivery
<Lisa> can it be handled at the user end via a user agent
EA: What we have is a way of making a wide range of fire hydrants or ambulances for many different countries
… That is being explored as an experiment at the moment. We are looking for global user feedback on this at the moment
… The aim is that these are critiqued as appropriate symbols
… Hopefully this will be completed by the end of the year
… So a question I am thinking about is if these are incorporated as part of collection or in core
janina: Just testing my understanding, it strikes me that from this TFs goals we really don't care what the image is just that people can get to the image that they understand
… Our concern is that the concept is being correctly mapped
… Trying to ensure that associated a numeric identifier to a concept and thus to an image
… What I am hearing is that the identifier doesn't change but the image is appropriate presented by the users context
EA: Symbols in our DB have their own IDs that are mapped to a concept.
… The translation of a particular word (or words) to a concept is difficult to achieve in particular languages
matatk: We are explicitly excluding translation. A translated page could be marked up separately
EA: Ok, in that case the word/label would have to have a separate ID. I will need to consider this.
<Zakim> matatk, you wanted to ask about local religious festivals
matatk: For local religious festivals; we are using the Bliss AV as our list of concepts. One reason is that it is comprehensive.
… We talked about the possibility that in some symbol sets there may be a symbol that doesn't correspond to a Bliss concept that could be requested to be added
… Local religious festivals may be one of those things that may not be broadly represented in Bliss
EA: The good thing about Bliss is that you can write the symbols online. There is no doubt that there are many festivals that we are not covering in pictographics or Bliss
… The difficulty then is that if someone comes across that then they won't get that symbol at the time of reading.
… This would need to have something that would be something similar as a workaround or no symbol is presented
… This is never going to be 100%
matatk: We want to make sure that we are not introducing things that would make this situation worse
… The authoring flow that we foresee is that there is a page and the non-expert author is reliant on the authoring tool to help them add concepts to their text
… The authoring tool provides Bliss code points to cover the concepts. The author doesn't need to know what the Bliss symbol is.
… When the user renders the concept then their symbol set replaces the code point appropriately.
… However if there is no festival symbol then what is presented?
EA: Take Thaipusam (a Singaporean festival). This isn't a well known festival. AI comes up with something or the author could create a Bliss symbol that represents this
… This could be designed in Bliss as it is generative language.
… The user then would need to map the Bliss symbols to a local symbol.
matatk: Just one more thing, this reinforces that Bliss is generative and authors could come up with new things.
… Many people might want to refer to that festival. For the system to work universally then there needs to be a away for authors to feed new concept mappings back into core.
… But what if the user doesn't have a local symbol for this
… The concept look up could be in authoring tools but is it a big deal with mappings not being included quickly?
EA: No, this wouldn't be a huge deal.
… The problem isn't generating the symbol. It would be possible to take an image and create a symbol in any symbol set.
… Would need to also ensure that a symbol is added with a unique code.
Lionel_Wolberger: How did you chose Concept.net?
EA: It is open-source, it is added to and it has a wide ranging vocabulary. It also has automatic translations
… It is a huge database
Lionel_Wolberger: For later - Does concept.net meet W3C normative
<Lisa> i think it is by pricton orignaly and used in the original wwaac project
Lionel_Wolberger: Searching globalsymbols.com for money I end up with three different symbols.
<Lisa> trying to incorporate every concept
Lionel_Wolberger: Inspecting the HTML shows that the images are included as CSS background images.
<Lisa> AND availible in RDT or OWL
Lionel_Wolberger: The focus of our work is that UAs could get unique IDs from sites like this to be included/interpreted when content is rendered.
… [Demoing <ruby> with Bliss code point in <rt> and alternative using `adapt-symbol` and code point in that]
… Any opinion on these?
EA: The only concern is that Bliss concept can require 2 or 3 symbols
Lionel_Wolberger: Have you seen <ruby> used in this way?
EA: I am sure this would work until concepts that require more than one symbol or a missing a symbol
Lionel_Wolberger: When there is no symbol the current specification is that a suggestion is provided. One of the nice things about using concept.net then you could traverse the tree upward to get an alternative representation.
matatk: Given that the first version of our proposed approach seeks minimal symbolisation for maximal results, what would be the most important content to highlight with symbols for our user population?
… E.g. we are thinking of a single symbol per row, for chapters in videos, or step-by-step procedures; the symbol functions as a summary to aid comprehension.
EA: Yes
Annalu: To reiterate we are not translating, we are symbolising
EA: Yes, similarly to translating BSL for example
matatk: Interesting we have been looking at symbols above but need to think about ways of presenting symbols in different ways.
matatk: One other thing to touch upon, we have been suggested this particular approach using Bliss and <ruby>.
<Lisa> the problem with concept net is it is too big
matatk: If we can overcome the identified concerns later today then we need to be clear that this approach only uses Bliss
<Lisa> too many similar concepts
matatk: If we need to connect this to concept.net then this is something this approach doesn't address
<Lisa> of course most do not have symbols assosiated
<Lisa> and how to you select between similar concepts
<Lisa> practicaly, when we tried to build it, it just became overwelming
Annalu: Can we put this on the side just now. This introduces a challenge in how authors select symbols, and how one translates properly
Lisa: We have tried some implementation with concept.net. The system is huge though.
… So, lets say you have 10,000 symbols and millions of concepts this creates problems for authors to select from.
… This seemed to bring a huge amount of problems. Theoretically it could work but in practice selecting symbols would be problematic
… We haven't precluded concept.net in our approach yet
<Zakim> matatk, you wanted to check on where the concept net / looking for an alternative would fit in to the authoring / viewing flow
matatk: Aim is to avoiding precluding things in advertantly.
… The author isn't trying to find a symbol but just find the concept.
EA: Concept.net if you have a word like 'spring' you could find a symbol that has a coil, or time of year, or to jump.
… The user doesn't know which one applies to the text. That is where the gatekeepers come in.
… The dilemma is that whoever is using the system, you don't know if it is the AAC user who doesn't understand the text or the gatekeeper who does
<Lisa> it is the auther who specifices the concept
<Lisa> note they will get it wrong
janina: We are assuming that it is the author is finding and identifying the appropriate symbol
<Lisa> you can cascade sybol sets (thatb is what we did)
<Lisa> at the user agent.
Annalu: If the user doesn't have a symbol then they could go to the symbol provider to request a symbol
<Lisa> user needs to select the possition
matatk: So keying off Bliss, not doing translation, need to be able to position <ruby>, talked about process of customisation.
… The customisation is done by different entities so expecting multiple sources in this is a bit of an ask.
Lionel_Wolberger: You mentioned that things are moving quickly and AI is quite the change agent.
… People are commonly decorating with emojis but I am told that this is not the same as symbols.
… We are interested in symbols as an aid to comprehension.
… I am interested on how emojis are now being used in this way.
EA: If you go to our site there are references to this and emojis are returned as possible options to searched for concepts
https://
EA: If this is coming from an AAC user perspective then this is great as this is what people are using
janina: And them linking this to a Bliss ID is fine?
EA: No problem at all
… AAC images can be strange but they are very personal
Lionel_Wolberger: So decorating with an emoji is similar to an AAC image?
EA: No, they are different ways of communicating
Lionel_Wolberger: The author tends to pick an image that they have in mind for the concept that they are selecting
EA: That is fine because it is a concept
matatk: Have you ever seen an emoji with an AAC symbol?
<Lisa> btw, coga comunity gave us the following feedback: Emoji alt text is often inadequate. Clearer guidance is needed on using emojis effectively in content, ensuring they don't replace essential text information. The current pattern (4.4.12) does not feel sufficient for members.
<Lisa> Problem: Emoji alt text may be inaccurate or insufficient. Their use in newsletters raises concerns about clarity and context. Simply pointing with an arrow emoji might not be clear enough.
<Lisa> Member: “In Content Usable, emoji are mentioned in the 4.4.12 Explain Implied Content (Pattern) but it just states “Note that standard emojis often come with an explanation or alternative text” but I’m not sure that’s accurate or very helpful.”
<Lisa> Proposed Solutions:
<Lisa> More detailed descriptive text for emojis.
<Lisa> Avoid using emojis to replace text.
<Lisa> Clarify the use of emojis in 4.4.12.
<Lisa> Clarify what "standard" emojis are.
<Lisa> Note: emojis don't have alt text by default - it must be added, which is covered in this WCAG 2.1 technique [Providing text alternatives for emojis, emoticons, ASCII art, and leetspeak]
EA: Technically possible to do
Annalu: Emoji convey some emotion related to the associated message.
… This gives people a wider array of tools
EA: If you can't do a thumbs up then you could add a symbol to convey this meaning
Annalu: This is a way to add to non-verbal communication
janina: Keen that we give good vocabulary as we write supporting material
<Lisa> simbolaztion
janina: Agreed this is not translation, what is it we are doing?
Annalu: Symbolization
<Lisa> +1 for symbolization
EA: It is definitely not 'decorating'
janina: So the process is 'symbollizing'
EA: Have to respect that this is someones language and we need to respect that
Lionel_Wolberger: Symbolization (symbolisation) is well defined. E.g., https://
EA: Similar to BSL. AAC could/should be an accepted language
Annalu: We need to remember that this field has been around for 50 years
FYI: Code sample of inline Ruby --> <!DOCTYPE html><html lang="en"><head> <meta charset="UTF-8"> <title>Inline Ruby Demo</title> <style> ruby { font-size: 1.2em; } .inline-ruby { display: inline-flex; align-items: center; gap: 0.25em; font-family: sans-serif; } .annotation { font-size: 1.2em;
background-color: #eef; padding: 0 4px; border-radius: 4px; } </style></head><body> <p> <strong>Standard Ruby:</strong><br> <ruby> cat <rt>🐱</rt> </ruby> </p> <p> <strong>Simulated Inline Ruby:</strong><br> <span class="inline-ruby"> cat <span class="annotation">🐱</span> </span>
</p></body></html>
Annalu: There is a slight tension between BSL and AAC communities but these is also a shared recognition of the importance of the words that we are using
<Lisa> i think it is a concept that we are recoding
Reviewing the expert session
Lisa: The idea of them changing the representation but not the concept seems alien - has this ever happened?
Annalu: I think the best way to think about this is spelling words. The concept remains the same, but the way we concatenate a seuqnece of bliss characters to reflect that concept may change. It doesn't happen often, but it may change. And we have a way to reflect that in the AV via either the gloss or the English word associated with the concept,
followed by the word 'OLD' which means its an old spelling.
Annalu: Some bliss codepoints point to concepts; some concepts are identified by a sequence of charcters (codepoints).
matatk: How often does it happen?
Annalu: Very seldom - I can look that up.
Lionel_Wolberger: When it happens, what's the workaround?
matatk: Symbol sets would have to change because Bliss changed the appearance of something. This isn't appropriate. But is it worse - existing content will break if the symbol sets update?
matatk: So the steps:
… I search for 'tea' in my authoring tool
… It returns 'drink' 'leaf'
… This is added to the code
… Later Bliss decides to change the 'spelling' for tea
… So we are keying on the spelling of a word where the spelling of the word can change in the future
Annalu: So if website was looking up on that ID then it would go to 'tea (old)'
matatk: But old webpages would still point to the old spelling
Lionel_Wolberger: So I am trying to look up the concept of 'tea' and the authoring tool returns two unicode code points for 'drink' + 'leaf'.
… Later a user views the page and the look up shows those two symbols
… Later Bliss changes 'tea' to 'glass' + 'leaf'
… Later user views the page again and the symbol still has the same two symbols as originally.
matatk: Correct but everytime Bliss does this then symbol set fonts need to update to recognize the change
Annalu: If we look at the way the webpage is formatted, then we have 'tea' followed by two code points.
… <ruby>tea <rt> code point for drink <rt> code point for leaf</rt></ruby>
janina: So not wrong but historic
matatk: My font which displays my symbols, when it sees 'drink' + 'leaf' then it displays my ARASAAC symbol for 'tea'. After an update the AV is updated and this changes the reference
Annalu: I have done the calculation... in the 50-year history of Bliss words, 11% are marked as old.
Annalu: Unicode is the alphabet, and the AV is the dictionary. What we're asking symbol set creators to do is to create their font based on the dictionary rather than the alphabet.
kevin: Sounds like this can work.
matatk: Yep, seems so!
Lionel_Wolberger: Quote
In the past 50 years of language development and planning, a number of glyphs
have been changed. The Blissymbol Reference Guide Supplement (BR1 2004:1) specifies that old glyphs
must be supported for users who prefer them: “The original forms will remain in the BCI Authorized
Vocabulary and can be retained by Bliss users if they so choose”.
From: https://
ISO/IEC JTC1/SC2/WG2 N5228 2023-06-23 Parag. 10
Missing equivalent symbol
Annalu: The symbol font author shall define the "no symbol found" symbol
… the rendering system could display the default bliss symbol
matatk: Can a font define a null character?
Annalu: This does not sound like the platform should behave this way
… In my experience, the missing symbol should be rendered as BLANK. If the page is white, the square is white; if the page is black, the square is black
… the end-user does not experience any symbol at all
<Lisa> I have a coga related meeting at 3-4 your time. I will be back for the remote meeting
Just updated the agenda for later per our discussion: https://
Meeting practice
(agenda above - this bit won't need much scribing)
Accessible Name and PUA
Lionel_Wolberger: Accessible name is likely better supported in the PUA implementation
Revisiting Issue 240
Annalu: Shows a "psychology" example
<Lisa> need to juump to another call. will be back for the meeting
Lionel_Wolberger: Reposting an example of <rt> .. <rt> .. </rt> https://
Meeting with colleagues representing other W3C groups and WHATWG
[Introductions]
matatk: Welcome to all. We have been working through identified issues and have some questions to explore a bit further with everyone
matatk: We have gone through user cases and situations to explore where challenges or questions might exist
… We have added the list of questions into the agenda
… We had a proposal for an attribute that allowed to include an integer reference to a Bliss id in an authorised vocabulary
… It was suggested that there was already a mechanism to map from an ID to a font character
… Bliss 'words' can be made from individual characters all of which will be in the Unicode font
… We identified an issue where Bliss may change concept mapping
<jcraig> Discussion related to w3c/
<gb> Issue 240 Could we build symbolic annotations with existing Web standards? (by DuncanMacWeb) [i18n-tracker] [Symbols]
matatk: We are indexing from a symbol ID rather than a concept ID. However, on exploration Bliss does not delete old 'words' so older content will still render just with older 'words'.
… Over the last 50 years of Bliss only 11% of words have been depricated and changed
jcraig: First thing was Bliss integer look up; what about ARASAAC, is that referencing Bliss?
Annalu: We are modelling how different symbol sets could be compatible to this system. EA joined this morning and she was positive that other symbol sets could be compatible with the approach being propossed
Lionel_Wolberger: We are coordinating with ARASAAC through globalsymbols.
… The goal is to symbolize content. The approach is to allow authors to key-in symbol identifiers that could rendered as ARASAAC, Mulberry or personalized symbols.
matatk: We chose to base this work on Bliss because it has a wide ranging set of concepts and a mature approach to expanding the set
… Someone could make a font that keys from Bliss code points to their own images
jcraig: Bliss being a superset is helpful
<Lisa> another point bliss isnt bound to a particular language - trys for universal use
jcraig: The Bliss integer seems a reasonable path for a polyfil
… Longer term the lookup will be Unicode. This may align directly with Bliss integer code.
Annalu: When Matthew says integer look up he means the Unicode mapping of each concept not from the Bliss ID
jcraig: Just to check, the goal is to use Unicode codes, have you already been granted a block?
Annalu: No, but we are conflating the route to approval to use Unicode with exploring the system to understand how this could work
annevk: So the look up will go from ID to series of code points but the Unicode block isn't defined yet
jcraig: Any polyfil is fine for me but just checking that any concept or look up could be used effectively with anty symbolic char set
annevk: Different fonts produce different symbol sets which are keyed off the same concepts
All: yes
matatk: So the Bliss 'alphabet' rather than the whole concept set will be added to Unicode ideally
… So 'tea' is two 'letter' - 'drink' and 'leaf'. ARASAAC could take these two code points and use these two to map to a single glyph that presents 'tea'
jcraig: Is the syntax robust to have 'drink' + 'leaf' means something different from 'leaf' + 'drink' and sentence structure?
Annalu: Correct and RTL matters and would make sense
Annalu: Yes, your two examples would be different concepts so 'leaf' + 'drink' might be some strange drinkable plant thing.
jcraig: So similar to emoji 'male' + 'doctor'. And the different 'drink' + 'tea' would be a different symbol to 'leaf' + 'drink'.
Annalu: Yes, different concept
Russell: It is classifier + specifier
matatk: Need to be clear that Bliss is a language and can be used to write complex text
… We are not trying to use it in this way for now though. This is not about translation but about symbolization
… For example helping users to understand parts of content, steps in a process, headings etc
… On the RTL issue we have some things to check internally
… We don't think this is a big issue though
jcraig: So using symbols for direction arrows how does that chage
Annalu: That is already considered in the Unicode proposal
Russell: In the Unicode proposal there is contention around the repetition of the latin alphabet but this is seen as an important inclusion
matatk: So, that was where we were in terms of how things might be addressed
… Another option explore was instead of having one code point or multiple code points, we could use the Private Use Area to map out core points
… Don't know if this would present issues for example around performance
jcraig: As a polyfil or the end game solution?
Russell: To expand, the downside is that a registry of the PUA would have to be maintained. The benefit is a one-to-one concept lookup.
… The real issue could be, although there are approx 6500 concepts defined in Bliss there are times that new words need to be defined.
… By using the Bliss Unicode spell a new word would require everyone to be dependent on BCI to come up with the word before it could be used.
… Alternatively, everyone could come up with a 'word' based on their chosen approach and this could change when an official BCI symbol was defined
… This does make this registry absolutely required and it is not maintained by Unicode
… The whole implementation in terms of font and <ruby> would be the same but would use PUA registry assignments instead of Bliss chars
jcraig: Being able to respond to new word needs quickly is understandable. That part could be done as a polyfil even when full set makes it into Unicode
… I don't think using PUA is a good use long term
… Creating this as a permanent 'fix' isn't ideal
… however using a polyfil as an interim fix for quick added words is good
matatk: If we are using Unicode alphabet then no more symbols will need to be added into Unicode
jcraig: What you are calling the alphabet, does this mean that joined concepts is part of Unicode?
… Also, you said nothing else will need to be added, it is worth adding a buffer just in case
matatk: Agreed but this is effectively putting a-z for Bliss into Unicode so we shouldn't need to add much more if anything
Annalu: So we are looking at the Bliss char set in combination with the Authorized Vocabulary (the dictionary).
… The AV is a living document that is actively managed and not added into Unicode
… We are asking symbol set developers to use the AV to key from the code point spellings in there.
… They can then create their own font with their images keyed from this.
… So we can create new concepts and new words. However, the Unicode entries for Bliss are the building blocks and don't change.
<jcraig> re: dictionary: 6² vs 6^2, or 6/2 vs 6÷2
Russel: Still a good idea to have some space... just in case
Annalu: Agreed although the desire is to not do this. Any modification to this isn't something that will happen anything like as regular as words are added
jcraig: Just on the dictionary analogy, you don't need to update a rending engine in order to have a dictionary
… These are going to be rendered sequentially and any system can rendered these things