Re: [admin] Publishing specs before TPAC: CfC start deadline is Oct 15

I've been toying a bit with the current chrome implementation of gamepads,
and been trying to make sense of how it would work for firefox.

There's a few observations I'd like to share:

- Being able to enumerate devices is very convenient. I don't think
Firefoxes implementation went that way which complicates matters. I think
the specification goes for enumeration right?
- The enumeration exposes device IDs, which is again convenient because it
allows to come up with some configuration GUI for the user where devices
aren't just "Device 1" "Device 2" and so on. It would also allow services
to come up with ways to provide presets for games, decoupled from the games
themselves and all that.
- There is the issue that you can miss button presses if they happen fast
enough in chromes implementation, since you get a chance to peek at the
device state only when you call getGamepads()
- Axes in general work fine with the above model, but I think buttons
should fire events.
- There is an additional complication with that, sometimes buttons are
mapped to axes (often a 4-way HOTAS, mapping to two axes). A usage which is
fine if the HOTAS is used for view control, but not so fun if used say for
missile selection.
- This leads me to think that the developer using a device should be able
to either subscribe to events for a control, or to query the current state.

On Mon, Oct 1, 2012 at 1:47 PM, Ted Mielczarek <ted@mielczarek.org> wrote:

> On Wed, Sep 26, 2012 at 10:27 AM, Arthur Barstow <art.barstow@nokia.com>
> wrote:
> > * Gamepad - Scott, Ted - what's the status of the spec and its
> > implementation?
>
> We probably need to discuss a bit more, but I think the spec is pretty
> close to a first version. The one large issue that we haven't tackled
> yet is how button layouts should be described. Scott and I have talked
> about that a bit, but we need to nail it down. I'm reasonably happy
> with the rest of it, I think there's enough there to be useful without
> wandering off into more complex territory.
>
> AFAIK Chrome has shipped an implementation. The Firefox implementation
> is hung up in review still because it's been a side project for me.
> I'd like to get it landed in nightly builds soon.
>
> -Ted
>
>