W3C

– DRAFT –
Spoken Pronunciation Task Force Teleconference

30 June 2021

Attendees

Present
Alan, Dee, Irfan, janina, JF, mhakkinen, NeilS, SteveNoble_
Regrets
Dee
Chair
Irfan Ali
Scribe
Irfan

Meeting minutes

Agenda Review & Announcements

Conferences

https://accessinghigherground.org/irfan-ali-spoken-presentation-in-html/

outreach to AT vendor for feedback

mhakkinen: have reached out to Joanie but havent heard back

mhakkinen: reached out to MS also and trying to get the feedback.

mhakkinen: Janina was going to reach out to NVDA

janina: may be reaching out to publishing people in MS will help

james-gable-nwea: I can reach out to Doug in MS

mhakkinen: Ask Doug that we will appreciate the feedback on our technical document. this is time sensitive

mhakkinen: He can add his comments or create a ticket and mention his idea about the two approaches that we have in our document

JF: ePub people are working on some proposal. They are focusing on reading system which reads the ePub document but doesn't provide the navigational stuff. It may be part of larger picture.

mhakkinen: are they proposing anything beyond what they already have?

mhakkinen: we did meet them but I didn't get any impression that they have anything specific planned

JF: feedback I got from that they wont get any traction from reading system vendors but I do not have any specific information. May be I can dig that info in my meeting with ePub

mhakkinen: my recollection is building pre-epud reading system. but non of them implemented any kind of SSML support.

<JF> https://w3c.github.io/epub-specs/epub33/tts/

mhakkinen: other organization is Amazon. They have read aloud capability in kindle. not sure if they are doing anything behind the scene to enable correct pronunciation

mhakkinen: John- if you find any new info that ePub is doing related to pronunciation please share.

JF: Not sure if we want to put the agenda item for APA call.

janina: this is not the week to do that

JF: I am walkign into ePub group with ideas and look forward.

Review Feedback

https://github.com/w3c/pronunciation/issues/88

janina: this topic is hard of an agenda of meeting which is scheduled on Thursday setup by Shadi. I proposed to frontier who may be interested in the topic that existing API are insufficient. I agree that current API dont help. we need something. I also understand that AOM does not work. If the university of lisbon people are interested and we can get some funding that we can hope to take it forward.

mhakkinen: we need to engage owner of UI automation such as MS. folks on Apple, NVDA and others should be involved in this conversation

janina: how faster we can make it happen

janina: Aaron from google have the same ideas about current API

mhakkinen: who at MS is owner of UI automation?

Irfan will identify this in ARIA working group call.

<JF> HUGE +1 to polite response now

https://github.com/w3c/pronunciation/issues/89

janina: we thought about it and have details in our gap-analysis document.

mhakkinen: I do not know how would that work for multi-attr but for single-attr is easy.

JF: that might be something that we might be lean fairly heavy

mhakkinen: paul brought it early in the discussion which could be a case to work upon

https://github.com/w3c/pronunciation/issues/93

mhakkinen: at this point we can thank for the feedback. we will wait to hear the feedback from other AT vendors

https://github.com/w3c/pronunciation/issues/94

Alan: didnt understand how data-ssml-voice-language works

Alan: it is confusing to me, data-ssml-say-as-* - This, in the index, indicates to me that all data-ssml-say-as- prefixes will have some * filled in indicated the actual one being used, but I think it is intended that data-ssml-say-as is also a legitimate attribute containing one of the six format types.

janina: +1 to adding examples

mhakkinen: I wonder if this is the case to map all of the SSML function to single or multi attr.

mhakkinen: voice language attr. it may be including in SSML, we don't include anything which has no existing mechanism.

Irfan: we can add more examples after our feedback review

janina: once we lock in direction that we want people to implement our approach.

JF: if there is an overlap, how does conflict resolution play out. We may need to think of the use-case.

janina: we may not support in our specification. we differ the HTML lang. I am wonder if we have reason. we want to be specific to the point about lang attribute.

janina: do we want to deferred html lang attribute.

NeilS: we should go to ssml path

Agenda Review & Announcements

Conferences

https://github.com/w3c/pronunciation/issues/94 adding this item in next week agenda. starting point

Minutes manually created (not a transcript), formatted by scribe.perl version 136 (Thu May 27 13:50:24 2021 UTC).

Diagnostics

No scribenick or scribe found. Guessed: Irfan

Maybe present: james-gable-nwea