Requirements on mandatory constraints (ACTION-27)

Hi,

at the TPAC meeting, we discussed that it would be helpful do put up the 
requirements we try to meet with mandatory constraints. This has been 
done before (when the concept was originally proposed), but that is a 
long time ago and things may have changed since.

What we identified during the discussion was (as far as I remember):

a. Whenever a mandatory constraint cannot be fulfilled there needs to be 
sufficient information back the developer so they can adapt their code.

b. The app developer should be able detect whether a given mandatory
constraint is understood by the User Agent.

c. Not all constraints should be useable in mandatory mode.

d. The risk of by mistake using Mandatory constraints in a way that
inadvertently locks out users should be minimized

(The reason for all of them except d. are obvious. For d. I think we 
discussed that it is impossible for the UA to know the direction of USB 
cams so for example facingMode should not be allowed to be a mandatory 
constraint.)

We also discussed that the info leakage when a mandatory constraint 
can't be met should be as low as possible (but still sufficient for the 
app to be able to make something intelligent).

Three questions:

1. Are the four requirements I list above in line with what people 
think? Any comments are welcome.

2. Can someone come up with a requirement on information leakage?

3. Are there additional requirements related to mandatory constraints
(and they should be motivated by use cases)?


Stefan

[1] http://www.w3.org/2011/04/webrtc/mediacap/track/actions/27

Received on Monday, 25 November 2013 08:53:21 UTC