Re: Proposal: Fire a toneDiscarded event while discarding the invalid DTMF values.

On 13 Jun 2014, at 15:18, Harald Alvestrand <harald@alvestrand.no> wrote:

> On 06/13/2014 03:54 PM, tim panton wrote:
>> On 13 Jun 2014, at 14:35, Makaraju, Maridi Raju (Raju) <Raju.Makaraju@alcatel-lucent.com> wrote:
>>> So, I am sorry, but the argument of "Javascript can do it, so browser should not" does not sound like a good argument to me.
>> That is exactly my argument. (better expressed than I’ve managed so far ;-) )
>> 
>> If it can be implemented by the average javascript developer in 10 lines or can be neatly put into a simple library we shouldn’t
>> put it in the browser.
>> 
>> Every thing we add at this stage (and I’ve been saying this for 9 months) delays the standard, and complicates testing. 
>> You have to be _very_ sure your favourite feature is worth the delay.
>> 
>> 
> 
> Tim, DTMF was added in order to have a surface to command the browser to
> send RFC 4733 telephone-events, because people said that was required to
> implement use case #3.4.2 ("Fedex Call").
> 
> Telephone-events can't be generated using WebAudio.

Yes, I know I lost that argument. (Someone pointed me to a
commonly used device that supports 4733 but not tone-detect, unless you pay extra).

> 
> The interface we have in the spec now was finished a year or so ago. It
> was thought at the time to be an interface that solved the problem, and
> had as few bells and whistles above that as we could get away with.

Yup.

> 
> To my mind, every message we spend on revisiting this inteface is a
> waste of time. We satisfied our use case, we should go on.

+1 

> 
> Some people are saying we should change the text.
> They need to show that the changed text fits better with satisfying use
> case #3.4.2, OR that it makes life significantly easier for either
> browser developers or Javascript developers.
> 
> If not, I see no point in changing it.
> 
> Given that -use-cases is not changing, I don't see ANY justification for
> revisiting the issue of whether we have this API or not.

Agreed, my original point was that the additional features of filtering the invalid symbols
could be done in 1 or 2 lines of javascript and so did not belong in the browser.

Tim.

> 
> 
> 
> -- 
> Surveillance is pervasive. Go Dark.