Re: Draft WebAudio API review text

On Thu, Jul 18, 2013 at 8:45 AM, Domenic Denicola <
domenic@domenicdenicola.com> wrote:

> This is lovely. Go TAG!!
>
> Here are some small editorial critiques I think would make the feedback
> more effective:
>
> - It might be helpful to show code written "after the fixes." E.g., show
> code that uses the proposed `PannerNode` constructor directly, in addition
> to showing the desugaring of `createPannerNode` in terms of it.


Done.


> Or code that uses the promise version of `decodeAudioData` directly,
> instead of the callback version desugared in terms of an internal promise
> version. For that last one, a compelling example might be doing multiple
> decodings at once with `Promise.every`,


Done.


> similar to the HTML5 spec's recently-added example for `createImageBitmap`.
>
> - The layering section starts strong by talking about the connection
> between `<audio>` and web audio. But the bullet-pointed questions start
> talking about a bunch of stuff related to multiple contexts and hardware,
> and the connection there is hard to follow. Worse, it's not clear that
> these questions are answered by the following proposed redesign. Basically
> this section lost me---which might just be me not knowing the source
> material well enough, but might indicate it needs a bit more editorial and
> cohesiveness work.
>

Have moved some of the questions out.


> - The "Other Considerations" section could also benefit from some "after"
> code; right now it contains the awkward "before" code, but doesn't quite
> show what the world would look like with the API fixed.
>
> Finally, one potential additional spec improvement:
>
> - It seems like there are many other opportunities for using promises to
> replace one-off events. For example, OfflineAudioContext and, I believe,
> the onended events. I admit I might not completely understanding how these
> work though, so promises might not apply there.
>

Done.

Received on Wednesday, 24 July 2013 22:16:03 UTC