Re: aria-kbdshortcuts feedback

> On Feb 24, 2016, at 9:12 AM, Dominic Mazzoni <dmazzoni@google.com> wrote:
> 
> Thanks.
> 
> I'm open to whatever name is decided on. Might be worth considering aria-shortcuts and aria-keys for something shorter, but I'm fine with aria-keyboardshortcuts.
> 
> On Wed, Feb 24, 2016 at 7:31 AM Richard Schwerdtfeger <richschwer@gmail.com <mailto:richschwer@gmail.com>> wrote:
> I would imagine that anyone implementing keyboard shortcuts would need to know their are limitations beyond knowing what the languages is for the page, regarding the keyboard being used. Given that you know the language you are targeting do you feel it is enough to be able to determine your keyboard shortcuts when authoring a page? 
> 
> No. It's true that there isn't currently a way for the web author to know what keyboard layout the user has. That should be fixed.

How do you propose to resolve this? 

> However, that doesn't sound like a reason not to have this ARIA attribute.

We should avoid specifying new Web API that is locale specific.

The context of the spec restriction I objected to was this vague non-requirement. It's a very hand-wavy "authors will figure it out" expectation,  which does not reflect the precision we expect from a W3C spec. Clearly there is some aspect of this feature that is incomplete or underspecified. I don't think it should be published until the problem is solved or addressed

>> It is incorrect to specify "%" or "Shift+%". However, note that on some international keyboards the percent sign may be an unmodified key, in which case "%" and "Shift+%" would be correct on those keyboards.


The section has a bunch of editorial mistakes too. For example:

>> Modifier keys must be specified exactly 

>> When modifier keys are part of a shortcut, they must come first. 

>> The non-modifier key is required and must come last. 


MUST by whom? RFC-2129 statements require an actor: either author, user agent, or assistive technology. All of the RFC-2119 requirements in this section need to be rewritten. 

Much of the normative prose is worded like informative notes. For example:

>> It is incorrect to specify "%" or "Shift+%". 
>> If the key that needs to be specified is illegal in the host language or would cause a string to be terminated, use the string escaping sequence of the host language to specify it. 

Are these requirements? If so, there should be an unambiguous RFC-2119 statement preceding these examples. If not, the informative context needs to be more clear, probably contained in a Note.

James


> Authors don't have a way to know keyboard layouts now, but that doesn't stop them from adding keyboard shortcuts. Overall I'd say they try to stick to keys that are more universal. The worst case is that a few shortcuts can't be pressed in a few locales, but that's a mild inconvenience.
> 
> Adding this ARIA attribute now allows authors to express the shortcuts they have, and if future APIs allow them to adapt these shortcuts to different keyboard layouts better, they could do so without affecting this spec at all.
> 
> I'd be in favor of including some language reminding authors that punctuation differs a lot from one keyboard to the next and any shortcut involving a non-alphanumeric may present compatibility problems.
> 
> Have you vetted this with the Google i18N experts?
> 
> No, but that's a good idea. I'll reach out right now and let you know if I hear anything.
>  
> - Dominic
> 
> 
> I personally don’t have issues with James name change to aria-keyboardshortcuts other than it is very long or the use of the word control (this is spelled out on Mac keyboards). 
> 
> We are working toward locking ARIA 1.1 down so that we can move on to ARIA 2.0, web component support, etc. 
> 
> Rich
> 
>> Begin forwarded message:
>> 
>> From: James Craig <jcraig@apple.com <mailto:jcraig@apple.com>>
>> Subject: aria-kbdshortcuts feedback
>> Date: February 24, 2016 at 3:33:06 AM CST
>> To: ARIA Working Group <public-aria@w3.org <mailto:public-aria@w3.org>>
>> Resent-From: public-aria@w3.org <mailto:public-aria@w3.org>
>> 
>> Issue #1: Name: kbdshortcuts. With the notable exception/oversight of the "img" role, ARIA doesn't use abbreviations like "kbd" in role or attribute names. This one should be changed to aria-keyboardshortcuts or something shorter like aria-hotkeys.
>> 
>> Issue #2: Spec examples make it seem as if both "Control" and "Ctrl" are valid values. My interpretation of the KeyboardEvent spec is that only "Control" is valid.
>> 
>> Issue #3: This prose:
>> 
>>> When specifying a key on the keyboard that changes when you hold down a modifier key other than an alphabetic key, you must specify the unmodified key name. For example, on most U.S. English keyboards, the percent sign "%" can be printed by pressing Shift+5. The correct way to specify this shortcut is "Shift+5". It is incorrect to specify "%" or "Shift+%". However, note that on some international keyboards the percent sign may be an unmodified key, in which case "%" and "Shift+%" would be correct on those keyboards.
>> 
>> 
>> If I recall correctly, I raised this specific example on the list last year as a reason for not including the property in ARIA 1.1. It is not possible for the web application to know which keyboard is being used and therefore not possibly to implement this feature in a i18n-friendly manner. Simply stating in prose that there is a problem does not resolve the problem. 
>> 
>> Please don't publish this feature without consulting some of the W3C i18n experts.
>> 
>> Issue #4: Editorial: Spec does not list [ARIA 1.1] on this property.
>> 
>> 
>