- From: cowwoc <cowwoc@bbs.darktech.org>
- Date: Sun, 27 Jan 2013 16:32:52 -0500
- To: public-webrtc@w3.org
+1 for breaking API changes until it's final. I'd rather put up
with API breakage now than a poorly-designed API when it is finalized.
Gili
On 27/01/2013 2:53 PM, Matthew Kaufman (SKYPE) wrote:
> From: Harald Alvestrand [harald@alvestrand.no]
>
> We're getting more and more code that depends on the API. We'd like to
> make sure that we can allow that code to run, but preferably we'd want
> to do that by adding polyfills instead of messing up the implementation.
>
> Sorry. I'm pretty sure the specification isn't anywhere near final, and the fact that you've been selling it as nearly so and developers are creating "more and more code that depends on the API" really is your (Google's) problem and shouldn't be limiting the WGs ability to create the appropriate API here.
>
> If you think adding an error callback for addIceCandidate is a complicated API change, just wait until the IETF actually takes up some of the work necessary for trickle ICE to even coexist with both existing ICE implementations and the SDP that describes existing ICE. I note that the first full IETF meeting in which MMUSIC has considered these proposals is still some months off.
>
> Matthew Kaufman
>
Received on Sunday, 27 January 2013 21:33:38 UTC