Re: minimum target size for mouse

Here's a forst crack at a rewrite of 2.5.4

The size of the target in relation to the visible display at the default
viewport size is at least:

- 44px by 44px for touch events
- 20px by 20px for mouse/pointer events (Level AA)

Note: If an activation target is to be used for touch and pointer/mouse
without automatic adjustment, then the larger size is required.

Editor note: The 20px value for mouse/pointer is a placeholder. We are
seeking research on this and outside input. We also have to define the
difference between a touch event and a mouse event, particularly in html
and responsive environments. Maybe some language in the Understanding
something like this:

"On platform where there is no distinction between touch and pointer, the
pointer size prevails unless it can't be used for touch..." or something
similar...
edit suggestions welcome... I started a wiki here:
https://www.w3.org/WAI/GL/mobile-a11y-tf/wiki/Proposed_revision_of_2.5.4#Proposed_2.5.4

On Thu, Apr 21, 2016 at 4:01 PM, Patrick H. Lauke <redux@splintered.co.uk>
wrote:

> On 21/04/2016 17:17, David MacDonald wrote:
>
>>
>>         The feeling of the the mobile task force is that our touch size
>>         SC should include a minimum size for a pointer/mouse target in
>>         addition to touch. We're not sure what that minimum should be,
>>         whether it should be the same 44px as the touch or smaller. It
>>         was the end of the call so were were not able to flesh this out
>>         sufficiently... Os I'm creating a discussion here. If there is
>>         momentum I'll set up a WIKI page.
>>
>
> As theoretically a mouse pointer has an "active" surface of 1px x 1px, it
> comes down to defining a minimum size that is "reasonable" enough to ensure
> that somebody who doesn't have a very steady hand can still hit the
> activation targets comfortably without the need to be absolutely
> pixel-perfect in their mouse positioning, and (relating to the other SC
> relating to accidental activation/ability to bail out of an activation) a
> size that is reasonably big to also accommodate slight movement of the
> mouse during the mouse button down / mouse button up movement (i.e. big
> enough that a user even with slightly shaky/unsteady hands can go through
> the down/up click movement without accidentally moving the pointer back out
> of the click target area).
>
> How much a mouse moves, in terms of CSS pixels, will depend on the type of
> mouse (for instance, there are "high-dpi" mice that potentially move very
> far in terms of CSS pixels for very little actual physical movement of the
> mouse) and the OS/UA settings (e.g. mouse sensitivity, mouse
> acceleration)...so once again, we have a moving target (no pun intended).
> But, if pressed to pull a figure out of a hat, I'd start around at least
> 15px x 15px, maybe even 20px x 20px.
>
> There is some discussion about
>>         how would the author know the what is activating the control on
>>         responsive sites that share icons for touch AND pointer. The
>>         current wording is.
>>
>
> There is currently no way to know "a priori" what input type a user will
> be using. And there are multi-input devices (such as Surface) that have
> both touchscreen and mouse/trackpad input.
>
> There are CSS Media Queries 4 Interaction Media Features, but these can be
> quite flawed / lead to incorrect assumptions (see
> https://dev.opera.com/articles/media-features/).
>
> One possible technique is to use libraries such as what-input
> https://github.com/ten1seven/what-input which at least let you
> programmatically determine what input a user is using right now, but these
> obviously only work once the user started an interaction (they started to
> put their finger on the touchscreen, they moved the mouse) at which point
> you probably don't want to change the size of controls on the fly / under
> their feet/finger/mouse pointer.
>
> Long story short: in the responsive site scenario, you'd design for the
> least accurate/coarsest input type, that being touch.
>
> (and I'm not the only one who's been saying that for a while - for
> instance, this from Josh Clark's "Designing for touch": "ASSUME EVERY
> SCREEN IS A
> TOUCHSCREEN (AND THEN S OME)"
> https://abookapart.com/products/designing-for-touch)
>
> P
>
>
>         2.5.4 Target Size: A touch target
>>         <https://w3c.github.io/Mobile-A11y-Extension/#def-touch-target>
>> is
>>         at least 44px by 44px of the visible display at the default
>>         viewport size. (Level AA)
>>
>> What about this amendment, If it's an ok construct, then what value
>> should be for the pointer minimum size.
>>
>>
>>         2.5.4 Target Size: A touch target
>>         <https://w3c.github.io/Mobile-A11y-Extension/#def-touch-target>
>> is
>>         at least 44px by 44px and the pointer/mouse target is at least
>>         <x pixels> by <y pixels> of the visible display at the default
>>         viewport size. (Level AA)
>>
>>
>
> --
> Patrick H. Lauke
>
> www.splintered.co.uk | https://github.com/patrickhlauke
> http://flickr.com/photos/redux/ | http://redux.deviantart.com
> twitter: @patrick_h_lauke | skype: patrick_h_lauke
>
>
>