Spoken Pronunciation Task Force Teleconference

22 Jul 2020


irfan, Dee, paul_grenier, JF, SteveNoble_, janina, NeilS


<irfan> scribe: Dee

technical approach

Janina: We have many use cases for interoperable pronunciation. We have two approaches with pros and cons. We need to get a sense of consensus of which to follow. We try to get as much feedback from the group as possible. It is not a majority decision.

Mark: Background, solving pronunciation needs in the assessment space. Mechanism to add pronunciation cues that are easy to author.
... Different vendors have different approaches.
... This group began looking at approaches that would work for everyone.
... The group has landed on two approaches.
... SSML inline approach has advantages. Some applications support inlining SSML.
... The second approach is the attribute model.
... There is also a single attribute model.

Both approaches are presented in the explainer docoument.

Paul Grenier (PG): The more natural fit is to have an XML solution. Concerned about authoring.

PG: XML or XML-like tags around HTML
... Reuse is desirable, similar to SVG resuse model.

Mark: Part of the AT challenge: accessing more of the DOM tree.

Neil: Question for Mark, not sure if the attribute will be in the accessibility tree. Is it that the AT will have to do a deeper dive?

Mark: SSML attribute may be easier to query.
... AT vendors will have to comment.

Neil: Can't recommend what was done for Mathtype.
... Aria live regions are areas where pronunciation guidance may be needed/is desired. Ask AT vendors about this.

Peter (Texthelp): Would like to stay away from aria.

PG: Similar issue, use aria label for phone numbers, but this would affect the Braille display.

Peter: Texthelp adopted the json approach and it works well.
... Phone numbers, street addresses, etc. are working.
... Texthelp developer is not available for scheduled meeting.

Janina: perhaps have another meeting in order to have as wide a response as possible.

Data dash discussion. Full HTML5 or reserved word.


Janina: Sending another email to increase replies from stakeholders.

Irfan: Will stakeholders be able to give feedback via email?

Janina: In person conversation is preferred.

Mark: Pearson also implemented an attribute model. IMS standards do support SSML attributes.

Irfan: Next week we will have more vendor/expert feedback.

Alan (Cambium): raw text implementation.

Alan: what is the best way to interact with others outside of the meeting.


issue triage

<paul_grenier> https://glitch.com/edit/#!/ssml-components

<irfan> https://github.com/w3c/pronunciation/issues/64

<paul_grenier> custom element example implementation

Janina: challenge to meet broader needs, not just AT.

Mark: Apple does not support SSML but they do have their own speech markup.
... Software testing tool maps to Mac commands. The data attribute model is not necessarily tied to a synthesizer that has its own markup.
... this is a potential benefit

Other Business

<mhakkinen> I have to drop to chair another meeting. Thanks all!

Janina: agenda will be included in the email to the vendors

trackbot, end meeting

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version (CVS log)
$Date: 2020/07/22 15:03:09 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision of Date 
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00)

Default Present: irfan, Dee, paul_grenier, JF, SteveNoble_, janina, NeilS
Present: irfan Dee paul_grenier JF SteveNoble_ janina NeilS
Found Scribe: Dee
Inferring ScribeNick: Dee
Found Date: 22 Jul 2020
People with action items: 

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.

WARNING: IRC log location not specified!  (You can ignore this 
warning if you do not want the generated minutes to contain 
a link to the original IRC log.)

[End of scribe.perl diagnostic output]