Personalization Task Force Teleconference

11 April 2022


becky, CharlesL1, janina_, Lionel_Wolberger, Lisa, Matthew_Atkinson, mike_beganyi

Meeting minutes

<Lisa> (but I need to leave a bit early)

Logistics: Easter/Passover/Ramadan holiday season

<Lisa> I wont be around. taking it off

Janina: Our focus should be the additions to the document that came up in email

<Matthew_Atkinson> I _might_ be able to make it, not 100% sure. Officially it's a holiday here, though if I am actually at home I would tend to come.

Decided: We will meet.

RESOLUTION: We will meet next week as usual, for those members not attending, have a good vacation.

Naming progress https://github.com/w3c/wai-personalization-standards/issues/9

sharon: I understand we are moving forward with WAI-Adapt

janina_: There was some additional discussion with Judy on the WAI-CC call

janina_: It was agreed that we do not need an expansion, but we do want a pithy phrase to put on the slide where we introduce the standard at TPAC
… we suggested phrases as well

janina_: We've got all the pieces, someone needs to capture that 'pithy phrase'

sharon: I will go through the lists and get the phrase.

Content Module Implementations Status

sharon: Discussions with Phill at IBM are going well, need to clarify further.

janina_: With access-board we got commitment to markup the ADA-complaint form.
… we need content and two user-agent implementations.

Lionel_Wolberger: UserWay is going to be a renderer of this new markup.

janina_: A reminder, renderers should render any markup.

Lisa: Ayeleth created a git long ago, she created a script. It featured a button to show symbols

<Matthew_Atkinson> I think the link is: https://github.com/ayelet-seeman/coga.personalisation (there are several repos). Also we linked the HTML example in our TPAC talk.

<Matthew_Atkinson> https://rawgit.com/ayelet-seeman/coga.personalisation/demo/conactUs.html

Lisa: so it is both a markup and a rendering.
… a shopping cart mockup

Matthew_Atkinson: What exactly does IBM expect to render, if they do render?

janina_: Given that the specification has x number of 'must' statements, imagine we number those statements
… then for each x(1) must statement, we must show two independent user-agent rendering implementations
… there can be one content source that both render
… the ADA complaint form being marked up would even be worthy of a press release
… as well as an IBM employment form

Lisa: +1. It's very exciting.

janina_: Press reports are not as interested in the renderer as the content.

janina_: We need in writing the exact terms from Bliss. We may need to go back and review aspects of the W3C process
… keep in mind this is enabling many various registries of symbols

Lisa: We have an excel spreadsheet, linked in the Wiki

<Lisa> https://github.com/w3c/personalization-semantics/wiki/Implementations-of-Semantics#implementations-in-development

Lisa: this is the link to the spreadsheet, on the wiki
… you can see that Vodafone and Atos made implementation pages

Lisa: What about wikipedia? I may still have contacts to developers over there.

janina_: Bliss symbols should at minimum give us a dump periodically in PDF format
… we would want a commitment from them to continue to maintain this registry
… a best case scenario is that Bliss will commit to retaining a Bliss registry on W3.org

Roy_: I did speak with them about that topic, I will find the email.

Implications of issue 203 for the data-symbol attribute (multi-value vs single value) https://github.com/w3c/personalization-semantics/issues/203 sharon]

sharon: Citing the note:
… Which of these is correct? The English "ham sandwich" or the Spanish, which literally translates to "sandwich of ham." In the case sentence and object structure of a symbolic representation, does it matter?

<span lang="en" data-symbol="16223 16718">ham sandwich</span>

<span lang="es" data-symbol="16718 12324 16223">sándwich de jamón</span>

janina_: Janina notes that answering this question is answering a question from Apple, worth replying to

Matthew_Atkinson: We should have replied long ago

Matthew_Atkinson: The data symbol attribute can potentially take multiple values.

Lisa: We have a page in German and in Swedish (the cup of tea page)
… see there that 'cup of tea' can be combined to become a single symbol
… and sometimes you need multiple symbols

Matthew_Atkinson: If multiple symbols are listed, are they meant to be construed as combining to denote a single symbol?

Lisa: Depends on the symbol set

<Zakim> janina_, you wanted to react to Lisa

janina_: Apple does also make authoring tools, so they might get involved with symbol authoring
… my inclination is to (a) apologize to James for the delay (b) point him to the ongoing discussion (c) mention the rendering opportunity and (d) the registry need
… but then point out (e) it is not translation, just different uses of symbols

<Matthew_Atkinson> +1 to janina_ ACKing and clarifying with James.

CharlesL1: Rendering should be 1:1. Symbol '12345' is always a dog, etc.
… I do not think the HTML developer will do the markup. It will be a Bliss symbol expert hired to make the page Bliss-symbol-friendly.

Matthew_Atkinson: Does that imply that our system is 'broken', since it requires a Bliss SME?

janina_: Respectfully disagree. It should be easy to look-up symbol values and put them in.
… no need to know in advance the end-user's preferred set of symbols

Matthew_Atkinson: The mapping challenge falls on the AAC Symbol Set provider. Then, if they are lacking a Bliss ID, they might define that this Bliss idea '23456' maps to these three symbols.
… this is not well explained.

<Lisa> i need to drop off. sorry

janina_: This lookup table registry, -- assuming success -- will encourage symbol set authors to reach a wider audience and perhaps generate more symbols

<CharlesL1> +1 to more clarification on how symbols will work in the spec.

<CharlesL1> +1 to reaching out to JC

sharon: Do we need an issue to be sure we action the need for explanatory sections that Matt cited?

Lionel_Wolberger: I have a bit of confusion between (a) multiple symbol values in a single markup, in addition to (b) multiple symbols constituting a single aggregated symbol

CharlesL1: I have similar concerns to the ones Lionel aired.

Summary of resolutions

  1. We will meet next week as usual, for those members not attending, have a good vacation.
Minutes manually created (not a transcript), formatted by scribe.perl version 185 (Thu Dec 2 18:51:55 2021 UTC).


Maybe present: Decided, Janina, Roy_, sharon