W3C home > Mailing lists > Public > www-html-editor@w3.org > January to March 2009

RE: [XHTMLAccess] i18n comment 2: Keycode or character

From: Phillips, Addison <addison@amazon.com>
Date: Wed, 7 Jan 2009 13:34:00 -0800
To: "Phillips, Addison" <addison@amazon.com>, Steven Pemberton <steven.pemberton@cwi.nl>, "ishida@w3.org" <ishida@w3.org>, "www-html-editor@w3.org" <www-html-editor@w3.org>, "public-i18n-core@w3.org" <public-i18n-core@w3.org>
Message-ID: <4D25F22093241741BC1D0EEBC2DBB1DA017CF1AB02@EX-SEA5-D.ant.amazon.com>
Hi Steven and HTML WG,

This note is on behalf of the Internationalization Core WG. 

We recently received your responses to our comments on the XHTML Access Module and we reviewed them at a recent teleconference [1]. While some progress has been made, we're still not entirely satisfied with the results. Our focus is on Section 3.1.2 [2].

We recognize that this is a difficult problem in part because it hasn't been solved in a consistently recognized "best practices" manner: different platforms and operating environments have taken different approaches whose details vary when dealing with keyboard events and such. Notably, we've been engaged with the folks working on DOM Events as they struggle with similar issues. (Which is why one sees the text one does in [3]!!)

One of the main problems here is that there is often a difference between the "key codes" produced by key events (key up, key down, etc.) and the "char codes" that result from various key presses (i.e. "key typed" events). Try out [4] with different keyboard layouts, for example.

Comments on the current text follow:

<q>
This attribute assigns a key mapping to an access shortcut. An access key is a single character from the document character set.
</q>

This might not be the way to express this. Some visual characters are composed of more than one code point. Some physical keys on keyboards produce multiple characters (or no visual characters at all). And so forth. Linking the characters to the document's character set is probably not a good idea either (unless by "document character set" you mean X(HT)ML's character set, which is Unicode). It might be better to say something like:

<q>
This attribute assigns a key mapping to an access shortcut. The key mapping consists of a single Unicode code point (character). Typically the key mapping is expected to be accessible to the user via a single keystroke, although activating it might involve pressing or holding down multiple keys. The invocation of access keys depends on the implementation. For instance, on some systems one may have to press an "alt" or "cmd" key in addition to the access key.

Authors are cautioned that not all characters are appropriate as access key values, since they cannot be accessed directly from the keyboard. Other characters only appear when combined with base characters. Examples of these might include combining vowels or tone marks, such as used in Arabic, Southeast Asian, or Indic scripts. These are more difficult to communicate to users because, while they can often be typed independently, they are not typically displayed independently and the user might not know which character is intended as the key mapping. Finally, any key available on one keyboard might not be available on a different keyboard layout.
</q>

Later the text says:

<q>The character assigned to a key, and its relationship to a role or id attribute SHOULD be treated as an author suggestion. </q>

This should probably say: "The key mapping and its..." or possibly "The key attribute and its..."

In the remainder of this section, the phrases "key assignment", "key", "assignment", "key binding", etc. are used to mean the key attribute value, which, in turn, means a character (because the attribute value is defined to be a Unicode code point).

Ultimately, we think you're on the right track here. The Internationalization working group would be happy to review text or work with your WG in some other way to help resolve these issues.

Kind regards,

Addison (for I18N Core)

[1] http://www.w3.org/2008/11/26-core-minutes.html#item06 
[2] http://www.w3.org/MarkUp/2008/WD-xhtml-access-20080526/#sec_3.1.2.

    [a] http://www.w3.org/MarkUp/2008/ED-xhtml-access-20081023/#A_key 
[3] http://www.w3.org/TR/DOM-Level-2-Events/events.html#Events-eventgroupings-keyevents 
[4] http://rishida.net/utils/keyevents/index.html


Addison Phillips
Globalization Architect -- Lab126

Internationalization is not a feature.
It is an architecture.


> -----Original Message-----
> From: public-i18n-core-request@w3.org [mailto:public-i18n-core-
> request@w3.org] On Behalf Of Phillips, Addison
> Sent: Wednesday, November 26, 2008 7:28 AM
> To: Steven Pemberton; ishida@w3.org; www-html-editor@w3.org;
> public-i18n-core@w3.org
> Subject: RE: [XHTMLAccess] i18n comment 2: Keycode or character
> 
> (personal response)
> 
> I think this text moves in the right direction, but think that
> there may still be problems with it. Mainly, I think it is now
> unclear how the 'key' attribute is supposed to work, given that the
> word key is both disclaimed and also used to mean (or at least
> imply) actual keypresses.
> 
> It should be noted that there is not a well-defined solution to
> this problem. WebAPI has been struggling with this also. In
> practice, how physical key events and character input are related
> is normally handled at a fairly low level in the system. Higher
> level software that attempts to listen and respond to key press
> events often ends up damaging or disabling more complex input
> systems, such as the IMEs (input method editors) used to compose
> e.g. East Asian text.
> 
> (chair hat ON)
> 
> Thanks for the response. The I18N WG will review your response and
> text in detail. Our next teleconference is today.
> 
> Addison
> 
> Addison Phillips
> Globalization Architect -- Lab126
> Chair -- W3C Internationalization Core WG
> 
> Internationalization is not a feature.
> It is an architecture.
> 
> 
> > -----Original Message-----
> > From: public-i18n-core-request@w3.org [mailto:public-i18n-core-
> > request@w3.org] On Behalf Of Steven Pemberton
> > Sent: Wednesday, November 26, 2008 6:46 AM
> > To: ishida@w3.org; www-html-editor@w3.org; public-i18n-
> core@w3.org
> > Subject: Re: [XHTMLAccess] i18n comment 2: Keycode or character
> >
> >
> > Thanks.
> >
> > We have tried to address this by making certain that people
> > understand
> > that "key" is
> >    an abstraction and does not correlate to a "key code".
> >
> > Please see the latest editor's draft for full details.
> > http://www.w3.org/MarkUp/2008/ED-xhtml-access-20081023/

> >
> > Best wishes,
> >
> > Steven Pemberton
> >
> >
> > On Wed, 06 Aug 2008 09:12:28 +0200, <ishida@w3.org> wrote:
> >
> > >
> > > Comment from the i18n review of:
> > > http://www.w3.org/MarkUp/2008/WD-xhtml-access-20080526/

> > >
> > > Comment 2
> > > At
> > > http://www.w3.org/International/reviews/0806-xhtml-

> > access/Overview.html
> > > Editorial/substantive: S
> > > Tracked by: RI
> > >
> > > Location in reviewed document:
> > > 3.1.2
> > > [http://www.w3.org/MarkUp/2008/WD-xhtml-access-

> > 20080526/#sec_3.1.2.]
> > >
> > > Comment:
> > > It isn't clear that this section has taken into account the
> > potential
> > > difference between key codes and the characters that may result
> > from a
> > > key press on a given keyboard. It seems to assume that the
> > character on
> > > a key cap == the key code identifier == the character produced
> by
> > > pressing that key == the character that is the value of the key
> > > attribute.
> > >
> > > This is not always the case when you take into account a
> variety
> > of
> > > keyboards serving various different locales.
> > >
> > > Please provide some precision as to how a key attribute value
> is
> > > associated with keyboard events. (Note that this has proved to
> be
> > a
> > > difficult topic for the specification of DOM3 keyboard events.)
> > >
> > >
> > >
> >
> >

Received on Wednesday, 7 January 2009 21:34:40 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 27 March 2012 18:17:59 GMT