Re: Bug 20109 - Permitted values for pressure

On Monday, February 4, 2013, Jacob Rossi wrote:

>  Hey folks,****
>
> ** **
>
> We took a second look at this issue and have a suggested tweak to the
> change we made [1]. In most pen enabled applications, mouse generally
> emulates *half* the maximum pressure of pen.  As an example, Paint
> provides half the stroke width when using a mouse as compared with maximum
> pressure of a pen. This is generally in line with guidance Microsoft has
> given for pen drawing for quite some time: minimum pressure should be 50%
> stroke width, half-way pressure is 100% (e.g. what mouse should produce),
> and maximum pressure is 150% [2].****
>
> ** **
>
> Given that, I reactivated this bug for us to consider making the value 0.5
> rather than 1 when the device is in the active buttons state.
>

Having the default be intermediate between minimum and maximum makes sense
to me, and will probably result in better behaving apps in general.

Similar to the pressure issue, the spec currently allows implementations to
> provide a width/height that’s “typical of the device type” when the
> hardware doesn’t support contact geometry [3]. But it doesn’t require it.
> IE10  does not simulate width/height for devices that do not provide
> geometry. But I think we’d probably make IE do so in our next release,
> similar to emulating pressure. I’d like to consider making the MAY
> (optional) a SHOULD (recommended) in the spec for emulating width/height. I
> think the emulation we’d use in IE is:****
>
> ** **
>
> Mouse, Pen – 1x1 width****
>
> Touch without hw support for geometry – about 40x40 CSS pixels ***
>

Yes, I like this change (the less UA-defined-behavior the better).  Are you
aware of any sites today that treat no-radius differently from some
reasonable default radius (eg. I could imagine some UI features changing
slightly when it's known that radius information is available)?  If so then
we may need to consider some mechanism to determine pointer capabilities,
but without a compelling use case it's probably not worth the complexity.

> **
>
> This “smart default” approach should also apply to tiltX/tiltY, but the
> existing default of 0 seems appropriate for those (e.g. the stylus is
> normal to the screen).****
>
> ** **
>
> -Jacob****
>
> ** **
>
> *** Adjusted for zoom. 40 CSS pixels is approximately 11mm, the average
> width of a touch contact [4].****
>
> ** **
>
> [1] http://dvcs.w3.org/hg/pointerevents/rev/47212a0ea91c****
>
> [2] http://msdn.microsoft.com/en-us/subscriptions/cc300776.aspx#S11 ****
>
> [3]
> https://dvcs.w3.org/hg/pointerevents/raw-file/tip/pointerEvents.html#widl-PointerEvent-width
> ****
>
> [4] http://bit.ly/win8touchguidance ****
>
> ** **
>