Google Developers Working On Gaming Protocol For Wayland

[sarmad]

Wait, what is Google using Wayland for? Is ChromeOS using Wayland?

[Michael]

Wait, what is Google using Wayland for? Is ChromeOS using Wayland?

As linked to from the article: https://www.phoronix.com/scan.php?pa...16-ARC-Wayland

[eydee]

Nice. We'd finally have a proper way for the screensaver to be suppressed in response to gamepad/joystick input rather than requiring some imperfect surrogate such as "suppress screensaver when something fullscreens" or "games/launchers are responsible for suppressing the screensaver".

Simply not using screensavers should be enough. Current hardware doesn't need it anymore.

[Hi-Angel]

Nice. We'd finally have a proper way for the screensaver to be suppressed in response to gamepad/joystick input rather than requiring some imperfect surrogate such as "suppress screensaver when something fullscreens" or "games/launchers are responsible for suppressing the screensaver".

Fullscreen screensaver suppress is not an inferior solution. Because you don't want screensaver to popup for watching a video. I am not saying that we don't need a protocol to react to gamepad presses — I'm just pointing that ditching out the previous behavior would be silly.

That said, boo on the W3C gamepad API's "Standard Gamepad" mapping. XInput is the new de facto standard on Windows and SDL's Gamepad API is making it the de facto standard on other platforms, but the W3C's solution is an inferior subset. (4 axes and 17 buttons can only represent XInput's 8 axes and 11 buttons if you represent the D-Pad as 4 buttons and downgrade the triggers from axes to buttons... though at least it can support everything a Steam Controller can say if you represent it as 16 buttons, four axes, and two sources of pointer events.)

If you're sure in that, this sounds like something worth mentioning on ML

[ssokolow]

Simply not using screensavers should be enough. Current hardware doesn't need it anymore.

It saves power to turn off the monitors when idle and, when I decide to fall asleep while watching a movie, I prefer the light sources in my room to follow me into sleep. I just happen to lump DPMI in under "screensavers" since all input idle detection is lumped under the XSCREENSAVER X11 extension.

[ssokolow]

Fullscreen screensaver suppress is not an inferior solution. Because you don't want screensaver to popup for watching a video.

I currently disable MPV's explicit screensaver suppression because it interferes with detecting system idle. I don't really miss it.

I think it's because I've been boycotting the MPAA for so long that I've grown a habit of putting YouTube on one monitor while doing work on the other two.

That said, I suppose I could just maximize the window without fullscreening it.

If you're sure in that, this sounds like something worth mentioning on ML

I never join mailing lists. I have enough trouble dealing with this mess of an inbox without them. Do you know if they're available through GMane's NNTP bridge?

[ssokolow]

Actually it looks like the W3C API has an analogue (double) value for every button - it can express far more information than the SDL or XInput ones where the buttons are a single bit.

Better, but it still can't get the semantic representation correct if there aren't enough axes for both analog sticks AND the D-Pad.

(Yes, I'm aware that a DDR pad needs the D-Pad to be presented as discrete buttons, but you'd think that a "standard" mapping would be able to represent XInput's semantic meanings faithfully and, even with button pressure, that's still two analog sticks and a D-Pad vying for axis status.)

Also, here's my vote (as a Canadian) for "analog" to be the contrast to "digital" so "analogue" can be reserved for its original meaning. (likewise with "dialog" and "dialogue")

[liam]

Better, but it still can't get the semantic representation correct if there aren't enough axes for both analog sticks AND the D-Pad.

(Yes, I'm aware that a DDR pad needs the D-Pad to be presented as discrete buttons, but you'd think that a "standard" mapping would be able to represent XInput's semantic meanings faithfully and, even with button pressure, that's still two analog sticks and a D-Pad vying for axis status.)

Also, here's my vote (as a Canadian) for "analog" to be the contrast to "digital" so "analogue" can be reserved for its original meaning. (likewise with "dialog" and "dialogue")

https://w3c.github.io/gamepad/#gamepad-interface

So, I didn't see where the spec limited either buttons or axes. The canonical standard gamepad (thus far, the only mapping defined and corresponding to the Xbox/ps layouts) has 4 axes and 17 buttons. The spec just shows two types of input types (gamepadbutton and axes) with the former comprising a an array of two booleans with a double value (range be normalized), and the later being an array of just the double value.

[Hi-Angel]

I never join mailing lists. I have enough trouble dealing with this mess of an inbox without them. Do you know if they're available through GMane's NNTP bridge?

Sorry, I don't know. You could mention it on IRC channel #wayland of Freenode, though. I know, you'd probably think that devs definitely knows the specifics better, I was thinking the same. But after they managed to forget about primary clipboard, and made windows titlebars into client-side decorations — so now there have to be some protocol between Qt, GTK, EFL, and whatever to synchronize them — I think everyone's opinion is important.

[Hi-Angel]

ssokolow I linked your comment on IRC.