W3C home > Mailing lists > Public > public-media-capture@w3.org > July 2014

Re: Bare constraint values - KISS

From: Jan-Ivar Bruaroey <jib@mozilla.com>
Date: Fri, 11 Jul 2014 17:09:53 -0400
Message-ID: <53C052A1.60905@mozilla.com>
To: Stefan Håkansson LK <stefan.lk.hakansson@ericsson.com>, "public-media-capture@w3.org" <public-media-capture@w3.org>
On 7/11/14 4:07 AM, Stefan Håkansson LK wrote:
>> I too have had concerns about mandatory in the
>> past, and I share your concern a little bit, but I think these
>> particular training-wheels get in the way of driving.
> As I said in my first response: this part is really only a minor concern
> for me and perhaps I should not have brought it up at all. (But I can't
> escape the feeling that we've gone full circle on this: did you not
> introduce the mandatory/footgun problem originally?)

The original mandatory footgun was false negatives, which 
getSupportedDevices solves, and I'm happy with that solution. 
Ironically, it matches something I suggested back in November. The 
syntax changes since then have largely chased webIDL compliance in 
harder ways on the assumption that the simpler getSupportedDevices 
solution would not be acceptable.

Now that that is fixed, is having mandatory as the default itself a 
footgun? Perhaps a smaller one, but I'm less concerned by it, because 
that pilot error should be immediately obvious, and it trains people 
quickly to widen their net using min/max and sequences of choices (put 
in wide hard edges first, then fine-tune what passes using advanced). 
gUM fails because the user actually doesn't have what you wanted, which 
seems simple and expected.

> I think however that we should treat screen/application sharing as other
> media types than camera video in the interest of getting done with the
> latter without having it be on hold until we understand the former fully.

Agreed, though unfortunately both Chrome and Firefox use a MediaSource 
constraint right now to control screensharing, and that doesn't really 
work unless MediaSource is treated internally as plain-means-required, 
which actually violates the constraints model we have at the moment.

> Stefan

.: Jan-Ivar :.
Received on Friday, 11 July 2014 21:10:19 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:26:28 UTC