Re: Adding Web Audio API Spec to W3C Repository

On Fri, Jun 10, 2011 at 9:41 PM, Ian Hickson <ian@hixie.ch> wrote:

> On Fri, 10 Jun 2011, Chris Rogers wrote:
> > On Fri, Jun 10, 2011 at 6:41 PM, Ian Hickson <ian@hixie.ch> wrote:
> > >
> > > (Sorry for breaking threading, I wasn't on the list until now.)
> > >
> > > One last thing: I'm not convinced that audio APIs are really the next
> > > most important thing to work on on the Web platform. There are tons of
> > > new APIs already being proposed; so much so that implementation
> > > quality is suffering (the biggest indicator of that being the
> > > continued lack of comprehensive test suites for large parts of the
> > > platform, including such fundamentals as HTML).
> >
> > Hi Ian.  For what it's worth, a comprehensive test suite is very much a
> > priority for us.  We've adapted both WebKit's mac and chrome port of
> > DumpRenderTree and run-webkit-layout-tests to handle audio within the
> > last week.  I've written seven "layout" tests so far, and have many more
> > planned for the near future.  So, whatever reason there is for *not*
> > having an audio api, this is not a good reason.
>
> I wasn't referring to the lack of an Audio API test suite, but to the
> general state of multivendor test suites for all Web technologies.


Ok, but still I think it's a bit unfair to use this lack of multivendor test
suites for "all Web technologies" as a reason not to have an audio API,
especially since we're actively working on audio test suites.

As for implementation quality, what more could be done?  The WebKit and
Chromium review process is *extremely* rigorous and the code has been
re-designed and re-factored to an excruciating level.  And while I'm sure
there will still be bugs which extensive layout testing and security review
will catch, they will be fixed.

Implementations exist for Safari mac port, Chrome Mac OS X, Windows, Linux,
soon Chrome OS, WebKit GTK port, all with non-trivial differing audio
back-ends.  Serious progress has been put into layout tests even this week.
 Mozilla and Opera could also have implementations and I am certainly
willing to share code and provide assistance.

Robert brings up points which should honestly and respectfully be worked
through and you're quite right that what matters in the end is what is
implemented in the various browsers.  Over time, I hope that some of these
differences will be reconciled.  But guess what?  We're working on audio
APIs here!

Chris