Re: Split TTS and Speech Recognition?

Hi, Glen–

I'm not trying to be pesky about this, and I'm not going to get pushy. 
But I'd like you to reconsider this, and I'd like to hear from others 
what they think (especially implementers).


On 10/8/13 8:40 PM, Glen Shires wrote:
> A unified spec hasn't slowed implementations, as there are currently
> browsers that implement the ASR portion and not the TTS portion, and
> browsers that implement the TTS portion and not the ASR portion.

This would seem to be an argument for splitting them up, not keeping 
them together. They are moving at different rates.


> (And speech aside, there are many examples where implementors
> implement a spec in parts.)

Yes, but this is not good for web developers. It's to be avoided, if 
possible. With my web developer hat on, this is really frustrating. This 
is why CSS took a more modular approach, which is working pretty well in 
terms of consistency and interoperability.


> Also, keeping TTS and ASR together avoids the problem of having to sync
> things up in the future.

Speaking from a position of ignorance and curiosity, what things need to 
be synced up between TTS and ASR? They seem pretty orthogonal from my 
reading of the spec.


> As the unified spec matures, it may have a
> better chance of finding a unified home in one of the major W3C groups,
> such as HTML.

I'm not sure I follow your reasoning there. Why would a single spec have 
a better chance of being adopted by a WG than 2 smaller specs?


Is there some concern that one would get implemented, and not the other, 
so keeping them together might incent implementers to do both?


Finally, I just want to be clear that this request is not me speaking 
with my W3C hat on; I'm speaking solely as an interested web developer 
who wants his apps to work in as many browsers as possible, and who's 
mostly using the TTS stuff.

Regards-
-Doug


> Glen
>
>
> On Tue, Oct 8, 2013 at 9:28 AM, Doug Schepers <schepers@w3.org
> <mailto:schepers@w3.org>> wrote:
>
>     Hi, folks–
>
>     I'd like to propose that the text-to-speech feature be split out
>     from the Web Speech API spec; it's more or less orthogonal with the
>     speech recognition aspect of the spec, and while there are still
>     open issues that are being discussed, I think it's more stable in
>     terms of implementations, and could move forward more quickly on its
>     own.
>
>     I have been using both TTS and speech recognition in some of my
>     recent apps, and I think both are very cool and useful; I think both
>     will be great for accessibility, as well. TTS is much simpler,
>     though, and I think we could get more implementations right away if
>     we split it out. I really want to see both succeed, at their own pace.
>
>     (As an aside, I made a "talking calculator" back in 2004 using SVG
>     and the Microsoft IE TTS API; it no longer works, but it hints to me
>     that it wouldn't be too hard for Microsoft to implement the more
>     modern TTS functionality in IE, if the path ahead were clear for them.)
>
>     In light of the recent news that the W3C Web Speech WG is not going
>     to be formed [1], I think the work should still be done in the Web
>     Speech Community Group, though maybe when it's mature enough, it
>     could move to an existing W3C WG to become a Recommendation.
>
>     (I don't have a strong feeling about which group this might fit in,
>     but a few spring to mind: the WebApps WG, the Audio WG, or the HTML
>     WG to take advantage of the new CC-BY licensing being experimented
>     on there. It could even be its own WG, though that seems like
>     overkill to me.)
>
>     If any of this resonates with this group, I'm happy to help with it
>     unofficially, with my W3C staff experience. (If it were ultimately
>     moved into the Audio WG, then I could give my official help, since
>     that's one of my working groups. :P)
>
>     [1]
>     http://lists.w3.org/Archives/__Public/public-new-work/__2013Oct/0004.html
>     <http://lists.w3.org/Archives/Public/public-new-work/2013Oct/0004.html>
>
>     Regards-
>     -Doug
>
>

Received on Wednesday, 9 October 2013 07:46:43 UTC