Re: [mediacapture-image] Add pan and tilt constraints

@jan-ivar, I don't think there's anything about being stateful in this spec. Your suggestion would require the implementation of the spec to keep a separate state, where as I believe current implementations just pass the values to the OS-level APIs and read them back accordingly. The problem with keeping state is that the state could be changed by something outside of the implementation, or simply another tab, which could lead to inconsistency.

The note you linked to, while interesting, is more about 1-to-1 mappings it would seem. The end also seems to suggest that `getSettings()` would still return the true value (although it doesn't say it has to).

I think that solution would slightly side-step the issue, but isn't an absolute fix, so I'm not sure it would be worth doing a half-measure if it adds extra complexity to implementations. The main issue is that order of operation when using `step` can cause rounding problems with floats, so while the mapping might save the programmer some of the time, a different path through the code may perform the math in a different order which would result in a different value, and we're back to square one.

-- 
GitHub Notification of comment by dsanders11
Please view or discuss this issue at https://github.com/w3c/mediacapture-image/pull/182#issuecomment-310811551 using your GitHub account

Received on Saturday, 24 June 2017 03:48:19 UTC