Agenda Review & Announcements
<JF> +1 to the transcription service - it *IS* good
janina: we have a new capability.. thanks to zoom. MIT is paying for it. Roy is the host. You need to be a host in oder to invoke the transcription. There is no plan to archive the transcript though. we have it available for us and can use it anytime. Our facilitator should be able to enable it.
NeilS: I think we need to record the call in-order to see more features.
janina: remind me at APA call.
janina: this is something different and a lot better and have so many additional features
JF: transcription is very good
janina: lets make sure that co-facilitator have the access to use these features
Action Updates https://
www.w3.org/ WAI/ APA/ task-forces/ pronunciation/ track/ actions/ open
paul_grenier: Mark made some updated and I have made some edits in the verbage.
paul_grenier: it fixed the formatting issues with our content. It is much cleaner now
paul_grenier: there is rom for improvement and if you see something please take a look.
<paul_grenier> IrfanA: I have created a separate document for the json schema but now need to update the main document and link to the schema doc.
<paul_grenier> IrfanA: we need someone from w3 to corrected the error in the document
<paul_grenier> janina: a staff contact can help us put the doc in the right place.
<paul_grenier> IrfanA: I will contact MichaelC since Roy is out. There are also two warnings related to ReSpec from deprecated files and maybe there's a new version. So, I'll reach out to resolve these issues.
<paul_grenier> IrfanA: I can reach out today.
<paul_grenier> janina: MichaelC can create a stub document that we then add content to
<paul_grenier> IrfanA: I read many people have similar issues from github comments
<paul_grenier> Dee: "NOTE" appears twice in the green note section
<mhakkinen> Yes, there are a few typos like "Note" and a spurious /H5 visible in the text
<paul_grenier> Dee: I looked at the SSML from Mark's demos and "once upon" isn't playing in the audio. https://
<mhakkinen> I need to recut a couple of the audios and fix the example text, such as the Say-As. Will check the audio links.
<paul_grenier> Dee: it's the phoneme sample I was looking at
<paul_grenier> NeilS: it would be helpful to see the code in the content
<mhakkinen> If anyone has additional sample they would like to add, please add as an issue.
<paul_grenier> Dee: I used dev tools. Does this get added to the technical approach doc?
paul_grenier: for now we are going to have these examples in stand-alone page. it is also helpful if we can include the code in example.
<mhakkinen> My concern with adding visible code in the samples, as it is a sample intended for reading by tools. Code is messy to listen to. Perhaps I can put the code in a details? Expose it if you want to see it?
tom_babinszki: it sounds to me that very often these examples are going to be used, so providing step by step guide to use will help users
paul_grenier: perhaps providing instruction for developers to learn how ssml works or link to this information will help users
tom_babinszki: i used it using python script
paul_grenier: to use developer console, there are several ways
paul_grenier: may be we can create a tool just like we did for the hackathon which can be tried by users to understand how SSML works
tom_babinszki: would it help to have some kind of texanomy or something
paul_grenier: Janina may be something like maple street doctor or something. May be we can create a github and we all can add our inputs
<janina> Found the following re Python cli for AWS:
paul_grenier: we have some examples where some city names spell the same way but pronunciation is totally different
<janina> Still looking ...
paul_grenier: dont know is voice packs is any problem. as long as content switches in the markup and TTS tech is setup, it will read it okay. however, if there is no switch or technology is not available .. then may need to me massaging the pronunciation then we have problems.
JF: for the ppl who are not native english speakers, can we articulate a request statement
janina: I can get some examples from my inbox. there is no trouble with french and english. however, error handling is something we need to talk about. Hungarian is some example.
<JF> NVDA does NOT ship with a JP language pack, but you can download one
janina: we want to support multiple languages but never discussed the error handling.
Dee: I have background in German. I may ne bale to call some examples.
janina: if you install language packs, you can get an answer for pronunciation
mhakkinen: there are two separate issues here. we see adoption for support of screen readers. we are focusing on the capabilities that are beyond language packs.
mhakkinen: thinking of two documents, technical and guide for implementors
Call for Consensus (CFC)
<paul_grenier> IrfanA: based on our timeline, by the end of Feb. we need to send this doc for CFC. We're in pretty good shape. Thanks, Mark. If we're going to create another document, like a guide for implementors and any updated examples, we need to do it by next week.
<paul_grenier> mhakkinen: i'd love to see more examples that we an include but the samples page is meant to be a way to compare someone's implementation.
<paul_grenier> mhakkinen: i can put some stuff in an issue and share what I'm thinking for another document for implementors
<paul_grenier> IrfanA: can we agree to send for CFC by Feb 24?
<paul_grenier> janina: yes. we also have to put the messaging together with Sean Henry
<paul_grenier> IrfanA: I have the template
<paul_grenier> janina: you should start that now.
<paul_grenier> IrfanA: i'll share the template with paul and we'll get that started