- From: Young, Milan <Milan.Young@nuance.com>
- Date: Fri, 25 May 2012 18:37:45 +0000
- To: Hans Wennborg <hwennborg@google.com>, "public-speech-api@w3.org" <public-speech-api@w3.org>
Sorry for not getting back to you earlier on these changes. In any case, yes, I agree with the updates. Thank you -----Original Message----- From: Hans Wennborg [mailto:hwennborg@google.com] Sent: Friday, May 25, 2012 10:13 AM To: public-speech-api@w3.org Subject: Re: Calling start() on an already started SpeechRecognition object On Thu, May 24, 2012 at 1:55 PM, Hans Wennborg <hwennborg@google.com> wrote: > On Wed, May 16, 2012 at 2:37 PM, Hans Wennborg <hwennborg@google.com> wrote: >> Hi all, >> >> The spec should say what happens when calling start() on an already >> started SpeechRecognition object, and calling stop() or abort() on a >> non-started object. >> >> Do we want to throw an exception? Fire an error event? Do nothing? > > If there are no objections, I'll update the draft tomorrow to say that > calling start() on an already started SpeechRecognition object causes > an InvalidStateError [1] exception to be thrown, and the same for > stop() and abort() on a non-started object. I've updated the spec: http://dvcs.w3.org/hg/speech-api/rev/386b0c5dd555 I changed my mind about throwing exceptions for calling stop() or abort() on already stopped (or non-started) objects. I think we should just ignore such calls instead. This way, web pages can call these functions as part of clean-up code without worrying about catching an exception or trying to keep track of whether the object was started or not. Thanks, Hans
Received on Friday, 25 May 2012 18:38:16 UTC