Re[2]: Quick (hopefully) question about WCAG 2 glossary definition of "keyboard interface"

Doh! Never say 'Quick' anything - the thread is jinxed *grin....

------ Original Message ------
From: "Detlev Fischer" <detlev.fischer@testkreis.de>
To: gregg@raisingthefloor.org; redux@splintered.co.uk
Cc: public-mobile-a11y-tf@w3.org; team-wcag-editors@w3.org
Sent: 30/06/2016 11:31:26
Subject: Re: Quick (hopefully) question about WCAG 2 glossary definition 
of "keyboard interface"

>One difference between sequential navigation via swiping and keyboard 
>sequential navigation (tabbing, arrowing) is predictability. In several 
>user tests we have experienced issues of focus loss or accidental focus 
>jumps even with users who know the swipe gestures and can in other 
>circumstances apply them correctly. Swiping (especially with Android / 
>Talkback) often resets the focus to what happens to be at the position 
>where the finger came down. This is partly due to problems of touch 
>implementation (especially compared with the better implemetation under 
>iOS/Voiceover), partly due to processing speed (test devices were Moto 
>E, Galaxy Note 3, Nexus 4, etc, so not the latest and fastest 
>hardware).
>
>This may still not be a general reason to keep keyboard and touch 
>sequential navigation apart - just saying...
>
>Detlev
>
>Compared to that, tabbing is a fairly predictable action.
>
>>  And I'm not saying we can ignore keyboard. What I AM saying is that 
>>it's
>>  irrelevant that "swipe" is a "gesture", the same way that for 
>>instance
>>  you don't explicitly call out "voice control" or similar...because
>>  regardless of the actual mechanism, the end result is what counts: 
>>the
>>  user controls (using a non-pointer device, so NOT a mouse or touch or
>>  pen) the focus, and can activate controls etc. In essence the OS/UA
>>  themselves handle the peculiarities of recognising gesture, voice, 
>>etc,
>>  and translate it at a high level into "move the focus, trigger the 
>>click
>>  action, etc".NNN
>>
>>>  With keyboard access I can use an alternate keyboard, a sip and puff
>>>  alternate keyboard, a speech control alternate keyboard,  a 
>>>miniature
>>>  keyboard, a large keyboard and many other input devices to control 
>>>my
>>>  iPhone (or any mobile device).   All of these and much more are cut 
>>>off
>>>  if you say  “swipes can replace keyboard access”
>>
>>  And I never said "swipes can replace keyboard access". I'm saying 
>>that
>>  for users who can use the touchscreen (but may be blind/visually
>>  impaired), using VoiceOver and swipe gestures is equivalent to using 
>>an
>>  external keyboard, or switch control, or any other peripheral, in 
>>that
>>  the OS translates that gesture for them into a "move the focus" 
>>action.
>>  Thus, at a functional level, it is equivalent to "keyboard" access 
>>(just
>>  that it does not fire an actual keystroke event)
>>
>>>  Again - because iPhone provides a keyboard equiv for each navigation
>>>  gesture — the iPhone is completely keyboard operable and meets WCAG 
>>>2.0
>>>  keyboard SC.    If it did not - and swipes were the only way then it
>>>  would fail.
>>
>>  I'm not talking about "does the iPhone pass/fail WCAG 2.0's SCs". I'm
>>  saying that unless you want to see all keyboard SCs being exactly
>>  duplicated unnecessarily to also cover sequential navigation 
>>mechanisms
>>  which are exactly the same, from a functional standpoint, as 
>>sequential
>>  navigation using an actual input device that fires keystrokes, the
>>  definition of "keyboard" can be extended to allow for input 
>>mechanisms
>>  like VoiceOver+swipe gestures. This does not remove or weaken the
>>  requirement that content needs to also work with an actual bluetooth
>>  keyboard. In fact, it tightens the requirement, as without a 
>>bluetooth
>>  keyboard, VO users can't simply press arbitrary keys, like a 
>>particular
>>  letter, or an arrow key, meaning that content needs to be made to 
>>work
>>  at a minimum with just reacting to focus/click/blur events (because
>>  paradoxically, things that currently quite happily pass the letter of
>>  WCAG 2.0 because they are "keyboard accessible" don't work in 
>>situations
>>  where a user is navigating using VoiceOver/TalkBack on a touchscreen,
>>  because they can't press specific keys - see
>>  https://w3c.github.io/aria-in-html/#aria-touch)
>>
>>>  Here we are talking about web content though — and it needs to be
>>>  accessible on both Pcs and mobile.    So you need keyboard access 
>>>for
>>>  all the places and mobile devices taht don’t have swipe nav.
>>
>>  You seem to be misunderstanding how content is made to work with
>>  VoiceOver + swipe (and note, once more, I'm explicitly talking about
>>  swipe WHEN AT IS ENABLED, as the AT is the one which then handles the
>>  gesture and translates it into high-level "move focus, activate 
>>element,
>>  etc" instructions that are passed on to the browser): you don't code
>>  special swipe nav detection into your site. You take care to use 
>>simple
>>  focus/blur/click handlers...the same way you would to make it 
>>"keyboard
>>  accessible" on PC.
>>
>>>  IF YOU ARE TALKING ABOUT ADDING A NEW requirement  (that all content 
>>>be
>>>  navigable BY BOTH the keyboard AND gestures )— then you have  
>>>different
>>>  problem.  The gesture control is not from the web content but from 
>>>the
>>>  iPhone and the author has no control of that….
>>
>>  Exactly, the author has no control over that, as the AT will handle 
>>the
>>  gestures in the case of AT+touch. Which is my point: functionally, 
>>the
>>  other has no control, but the author also doesn't need to do anything
>>  special that they're not already doing to handle keyboard.
>>
>>>  Help me here.   We are going back and forth here but I’m not sure I
>>>  really know what you are proposing to add or subtract.    Can you 
>>>tell
>>>  me specifically the wording you are proposing (or the final effect 
>>>you
>>>  want the wording to have?)
>>
>>  At this point, I wanted to understand why "keystrokes" was explicitly
>>  added to the definition. Now that I know, I'll be proposing a slight
>>  rewording in the coming week or so. That should give a hopefully 
>>clearer
>>  idea of where I'm heading with this.
>>
>>>  It is always good to include more options.  But dropping the basic 
>>>one
>>>  is not.
>>
>>  Again, I'm not dropping anything.
>>
>>>  Again, including other built in access approaches is always good.
>>
>>  That's the one. I'm proposing to carefully expand the definition of
>>  "keyboard" to include the AT+touch scenario, where the AT is handling
>>  the gesture recognition and then translates that into high-level 
>>"move
>>  focus, activate the element, ..." signals to the browser.
>>
>>>  REQUIRING them — means everyone has to do all of them on everything.
>>>   Be sure that is possible.
>>
>>  As the work necessary to make content work in the AT+touch scenario 
>>is
>>  functionally the same as that of making things work with a regular
>>  keyboard (with some clarification needed about which specific events
>>  need to be targetted, which is something that needs to happen in the
>>  non-normative/understanding/how to meet docs), it is possible.
>>
>>>  ELIMINATING the basic one for one that is usable by less people - 
>>>would
>>>  be a significant reduction in access span.
>>
>>  Not eliminating anything here.
>>
>>>  Again - can you say what IT is?
>>>
>>>  Adding a new requirement in addition to keyboard interface access?
>>>
>>>  IT above (saying that navigation doesnt have to all be possible from
>>>  just the keyboard)  (that you said doesnt weaken it )   would 
>>>eliminate
>>>  access for all the people using the approaches I enumerated above.
>>>
>>>  Or are you saying something different ?
>>
>>  Yes, see above.
>>
>>>>
>>>>>  I DO think it is GREAT that gestures also be possible.   Just like 
>>>>>it is
>>>>>  great that there are mouse ways to do things done by keyboard.   
>>>>>Both
>>>>>  should be available to users.  But keyboard access should always 
>>>>>be one
>>>>>  option.
>>>>
>>>>  And keyboard access would still be one of the options, as it's the
>>>>  most common non-pointer/sequential navigation interface in
>>>>  circulation. Just that on devices which are primarily touchscreen
>>>>  driven the most common non-pointer/navigation paradigm is 
>>>>functionally
>>>>  the same, but does not send "keystrokes" - the effect is the same
>>>>  though, in that it moves the focus.
>>>
>>>  when you say OPTIONS — do you mean
>>>
>>>  1) OPTION for the user (and therefore a requirement of the author?)
>>>
>>>      Then we are on the same page.  You are requiring BOTH the 
>>>keyboard
>>>      AND gesture?
>>>      My questions then are
>>>
>>>        * how does the author provide gesture access to his content?
>>>        * how does the author provide gesture access to his content on 
>>>PCs?
>>
>>  Yes, this is the option. And as outlined above, functionally this is
>>  completely transparent to the author as it's the AT (in the AT+touch
>>  scenario I'm talking about here) that interprets the gestures and 
>>sends
>>  high-level commands to the UA, which have the same effect that 
>>"actual
>>  physical keyboard" navigation has.
>>
>>>  2) If you mean OPTION of the author — then the author would be able 
>>>to
>>>  have some navigation accessible via swipe and not keyboard and that
>>>  would cut out all the people above again.
>>>
>>>      then it is not an option for the user — the author decides how 
>>>the
>>>      user much be able to access the content
>>
>>  No, that's not what I'm proposing.
>>
>>  In short: now that you've clarified the presence of "keystroke" 
>>(which
>>  is the one fundamental stumbling block I saw for even attempting to
>>  propose a change), I'll work on an actual proposal for your/Mobile
>>  TF/WCAG GL consideration.
>>
>>  Thanks,
>>
>>  P
>>  --
>>  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
>>
>
>

Received on Thursday, 30 June 2016 10:35:08 UTC