– DRAFT –
Spoken Pronunciation Task Force Teleconference

<janina> On my way -- Having Internet issues today

Agenda Review & Announcements

Irfan: we still have some edits to make, once that's done we will have Becky review and then send it to Michael for final review.

janina: I would like us to not keep referring to it as the technical approach but the publication of our specification.

First Public Working Draft (technical approach) CFC

Irfan: do we need to change the name of the document?

https://w3c.github.io/pronunciation/technical-approach

janina: Yes. we should change that to what the eventual specification will be.

janina: "Specification for Spoken Presentation" or something like that

paul_grenier: "Specification for Spoken Presentation in HTML"

janina: we will want a short name for this (that might be it). We want to ask Roy about that when he's back.

<mhakkinen> Spec for Spoken Presentation of HTML Content

Irfan: we don't have a TR url for the doc yet.

janina: we will get a short name and url at the same time.

JF: since EPUB is a superset of HTML, are some people going to think we're excluding EPUB?

janina: if that's a problem, they'll see this first public working draft and I expect they will comment.

NeilS: like you said it's a superset

Irfan: I will add these items in my TODO list. Thanks to Dee for providing the recommendations for spelling, grammar, and editorial mistakes. I'll make these changes today.

Irfan: Paul, you have already made the changes from last week. Except the reference url we talked about this morning.

paul_grenier: I'll take a look at those urls for the external resources and make sure we don't have 'www' in the urls.

janina: 'http' not 's' port 80 not 443.

Irfan: we also need to review compared to the style guide https://www.w3.org/WAI/EO/wiki/Style

janina: as long as we're consistent

Irfan: I'll finish the edits and Becky can review so hopefully we can be done by the end of the week

janina: we'll get a snapshot for the CFC so it doesn't change while we're reviewing and voting on it.

janina: Tom, you mentioned some stuff about the document two weeks ago...

https://github.com/w3c/pronunciation/issues/77

Tom_Babinszki: It still says "sssml"

paul_grenier: we can put these issues next to Dee's list

Irfan: I'm on it

Other Business

NeilS: what's the timeline everyone envisions going forward?

janina: I'm not sure we've talked beyond first public working draft

<Irfan> https://github.com/w3c/pronunciation/wiki/Timeline

janina: it may depend on the comments we get and implimentations

NeilS: is there another implementation in the works beyond text help's implementation?

Irfan: not that I know of.

Irfan: we updated the timeline and I think we're on track with that

janina: how long will we give people to comment?

JF: May 30th will give them 6 weeks

JF: Neil asked about implementations. Is there a possibility of getting a plugin for NVDA?

janina: we've had James Teh on a call before. We may get some Google interest.

JF: do we think a plugin added to NVDA would give us what we need?

NeilS: I'm about to try my hand at a plugin for NVDA, so I might be able to do this.

JF: if we publish in a couple of weeks with a deadline at the end of May, it would be great to have an implementation show up in the Summer months

(I have to go. I don't want to stop discussions, Irfan, please take care of minutes)

<Irfan> SteveNoble_: what are we doing with TextHelp implementation? Is that example where you can go and play from the user perspective.

<Irfan> janina: they support one approach and have agreed to switch the approach if we decide go go for another approach.

<Irfan> SteveNoble_: best I can do to come up with a practice test with SSML tag applied

<Irfan> SteveNoble_: I know some test use it. But you do not have to show underline code that is being used for those example

<Irfan> janina: tools that does the presentation has to read the code

<Irfan> SteveNoble_: I can look for some test content examples

<Irfan> NeilS: do we need to have specific test?

<Irfan> janina: yes

<Irfan> SteveNoble_: its within the test platform itself. it is not like someone taking a test.. it is a secure test

<Irfan> NeilS: its not like someone else has to write an example..

<Irfan> SteveNoble_: they have tool for test content which develop test content and can add ssml. I will see if I can find something publicly available example

<Irfan> JF: I think what we need for examples, we need to show content that is properly marked-up and using the technology and see how it works

<Irfan> JF: if content is behind the firewall, then we need to copy that content and host it publicly and show it which could leverage the technology

<Irfan> SteveNoble_: I will see what I can come up with from pearson.

<Irfan> NeilS: so along-with all these tasks, as per the timeline, there is decision making in November. is there any contingency plan where you can consider the feedback from wider community.

<janina> It's why I called the November date as a gating event

<janina> Neil, I'm sure it won't come out of the blue. There'll be lots of discussion, hopefully

<Irfan> NeilS: its an important date and I want to make sure that I am involved in this process.

<janina> If a direction becomes clear earlier, we're not wedded to November

<Irfan> JF: between now and Nov its an opportunity for engineer to test both the approaches, test it, and make a decision. After nov it will be final.

<SteveNoble_> FYI - here's the login page for the Pearson TTS generator that uses SSML https://ttsgenerator.testnav.com/ (of course, one needs a login...but just mentioning)

<Irfan> JF: based on testing, engineer feedback, the specific will go to approach A or B. we do not need a final definitive public implementation before nov.

<Irfan> JF: I hope everyone understands the timeline

<Irfan> JF: from june 1st to sep1 we may have enough feedback about any specific approach. we do not want to go beyond november at this point of time

<Irfan> NeilS: I might be able to do it

<JF> https://addons.nvda-project.org/index.en.html

<Irfan> NeilS: for NVDA plugin, it is mostly like what they support internally about SSML.

<Irfan> janina: you do not need to implement every feature in order to make it useful as per the spec

<Irfan> JF: as long as it is publicly available it can be hosted on private server. we just need to have a URL where anyone can go and test it.