Re: [mediacapture-main] What is the purpose of MediaTrackCapabilities displaySurface? (#915)

> If, on the current track the range/set consists of a single value, we run into the question of "value" vs. ["value"]. That's an orthogonal question. If you wish to discuss it, let's do so separately.

I believe that's inaccurate. e.g. [channelCount](https://w3c.github.io/mediacapture-main/#def-constraint-channelCount) is well-defined to take a range. The [editor's mistake](https://github.com/w3c/mediacapture-main/commit/059a9211d1622094dd8857db9beb477976f971dd) seemed limited to "properties that cannot ever be changed", to use your definition (facingMode was later corrected).

I don't think it's orthogonal either, as it hopefully explains part of the OP's _"spec currently says getCapabilities() should return the current [displaySurface](https://github.com/w3c/mediacapture-main/issues/1) value"_ confusion. I.e. it might have been less confusing if it had said to return an array with the current settings value in it (since that's the only value that can be set with applyConstraints).

 I'm still not clear on whether this issue chases utility or what "makes sense", because those aren't always the same thing:

E.g. it arguably "makes sense" for getSettings() and getCapabilities() to have the same number members, and it "makes sense" that inherent and non-inherent properties follow the same rules. But there's little practical utility in it, except maybe listing them in a generic app. 

There's no small irony having to explain the ConstrainablePattern and say this, but it's not clear to me this meets the bar for overriding earlier decisions.

> * Properties that cannot ever be changed, not even theoretically, should not be a capability. One example is displaySurface.

So [deviceId](https://w3c.github.io/mediacapture-main/#dom-mediatrackcapabilities-deviceid) wouldn't be a capability, but [facingMode](https://w3c.github.io/mediacapture-main/#dom-mediatrackcapabilities-facingmode) would, since it says: _"A camera can report multiple facing modes. For example, in a high-end telepresence solution with several cameras facing the user, a camera to the [left](https://w3c.github.io/mediacapture-main/#dom-videofacingmodeenum-left) of the [user](https://w3c.github.io/mediacapture-main/#dom-videofacingmodeenum-user) can report both "left" and "user"."_?

I.e. it can *in theory* be set to [e.g. "user" or "left"](https://w3c.github.io/mediacapture-main/#example-11) (without effect), even though it's an inherent property of the camera.

-- 
GitHub Notification of comment by jan-ivar
Please view or discuss this issue at https://github.com/w3c/mediacapture-main/issues/915#issuecomment-1302693059 using your GitHub account


-- 
Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config

Received on Thursday, 3 November 2022 21:36:34 UTC