Re: [RFC] Adding BTN_TOOL_TOUCH to input.h

[Dmitry Torokhov <dmitry.torokhov@gmail.com>]

On Sat, Nov 27, 2010 at 12:46:03PM +0300, Nikolai Kondrashov wrote:
> On 11/27/2010 12:15 PM, Dmitry Torokhov wrote:
> >2000LPI == about 79 l/mm. How is it not useable?
> TBH, I didn't make any real calculations.
> The introduced error isn't very high, indeed. The vertical size of 4x3 inch
> tablet becomes 75 mm instead of 76 mm. So maybe it's alright.
> 
> >In mainline synaptics is the only one that does this. We do not know how
> >many drivers in vendor trees fill and use this data nor we do not know
> >about userspace. The fact that we were not able to locate them does not
> >mean they do not exist. The units/mm are part of input ABI now.
> Well, I can't argue with you here.
> I didn't think about 3rd party drivers, TBH.
> 
> >Right, standarize on existing UOM (which are per/mm) and then implement
> >other UOM if theya re really needed.
> Will this fix qualify for inclusion into 2.6.37?

Yes, I believe it does.

> Also, what do you think about units per radians for rotational axis?

Works for me.

-- 
Dmitry