RE: [WebAIM] More data on accesskeys

Charles McCathieNevile wrote:
> Guess John (at least) was wondering where I was. Couple of thoughts...

Na, I figured you were busy, but would get back at it <grin>.


> 
> The basic idea is that accesskey,
> as it exists in HTML 4 (but not in the b0rken FF or even worse MS
> implementations) is workable. There is nothing intrinsically wrong
> with an author giving a hint for a key - although the user agent
> should decide to accept or replace it... It is kind of rough to
> expect someone using a device with a joystick in hindi to find the
> ÙR key is a useful shortcut. 

But Chaals, there *is* a cognitive issue if I, as a content author state:

"To do _foobar_ use accesskey <span style="text-decoration: underline;>¨¦
</span>" (&eacute;)

How many English speaking users know how to access that key?  Is that really
an appropriate accesskey in the first place?  It troubles me greatly that a
content author presumes to know which key should be the best for the end
user - "hinted" or not.  Your own example reveals the folly in categorically
suggesting that a specific key is the "proper" key.  I get that you envision
it as a hint primarily to the user-agent, but what system is in place for
those who cannot use the hint?  There may be a technical/mechanical
solution, but if I indicate that a specific key does something (either by
stating it, or "visually hinting" it by underlining a letter), and it is
over-ridden, how do we deal with the "confusion" factor?  I have not heard
the "graceful fallback" solution to this conundrum.


> 
> This means
> the guys writing the implied japanese site in my accesskey example
> above don't have to worry in advance about whether WHATWG, W3C,
> microformats.org or whoever can read their explanation of what their
> particular value means, because you have a way of applying it
> afterwards that is lightweight enough to implement in a browser if
> you want to provide real-time discovery. 

OK... But if there is "real-time discovery" capability in the browsers, why
again do we even need to hint a key?  onLoad: "Prompt: Attention user, would
you like to map a keystroke shortcut to this function?" - user selects key -
"Alert: that key is already taken by _foo_, are you sure you wish to change
it?" Yes/No; else "Alert: Thank you, that key has now been mapped to _bar_"
(or whatever... You get the idea) - and onward we go...  Does the condition
that triggers the initial prompt have to be an @key?  One would think not.

>>> I personally like the idea of role as it more closely defines what
>>> we are talking about, but if one of the goals of HTML 5 is to
>>> re-use as much of what we have now (rather than starting over
>>> again) then REL works too...
>> 
>> That's sums it up nicely!
> 
> Yep, in general. I think the decentralised and distributed
> extensibility story is really important, but there are plenty of ways
> to skin that cat. 

All without directly telling the end user what to do, only that the capacity
to do so exists.

Cheers!

JF

Received on Saturday, 2 December 2006 01:26:09 UTC