>It depends on to what extent these MSE APIs are shipping and thus 
>compatibility needs to be preserved.

MSE is in CR and there are shipping implementations.

>I'd love to work with you or the editors on that

I have added [email protected] to this thread.

/paulc

Sent from my Windows Phone
________________________________
From: Domenic Denicola
Sent: 13/10/2014 3:04 PM
To: Paul Cotton; Takeshi Yoshino
Cc: public-webapps; Arthur Barstow; Feras Moussa
Subject: RE: [streams-api] Seeking status of the Streams API spec

From: Paul Cotton [mailto:[email protected]]

> MSE [1] simply used the "Stream" object directly from the previous W3C WD 
> [2].  Since this object no longer is available in [3], what you do recommend 
> that MSE should do?

OK, down to the fun stuff :). Here's my take based on some brief 
investigation...

It depends on to what extent these MSE APIs are shipping and thus compatibility 
needs to be preserved. If they are shipping, something similar to appendStream 
seems OK: change it to take a ReadableStream and you'll be fine.

However, BufferSource is basically a custom WritableStream instance, with many 
parallels I can outline if necessary. So ideally the related MSE APIs would be 
redesigned to either use a WritableStream directly, or to subclass it. That 
would be more idiomatic in how the streams ecosystem works, and in general 
would allow more compositional scenarios through the usual mechanisms (so that 
e.g. people could integrate the BufferSource writable stream into a generic 
stream pipeline, instead of special-casing how they handle it). I'd love to 
work with you or the editors on that, if there are no compat-constraints 
preventing it.

Reply via email to