RE: agreement: user disposes; disagreement: author proposes [was: Re: When actions speak louder than words]

Charles McCathieNevile wrote: 
> I am happy for it to be optional, but believe that is should be there
> - it is useful to us, and anyone else who has an implementation that
> doesn't conflict with basic system functionality.
> 

And here lies a major issue - what good does an author suggested
(hinted) @key provide if it conflicts with:

	1) User determined keys (Conflict Resolution scheme awards to
end user)
	2) Software determined keys (Conflict Resolution scheme awards
to end user)

...while at the very same time you have well meaning but un-informed
content authors simultaneously providing the same "hinting" to the end
user via help files, "user manuals", underlining the letter, or, as
folks like Jesper Tverskov suggested, using the "First Letter"
(http://www.smackthemouse.com/20021031 - dated, I agree).

I question the usefulness of a hint that more often than not MAY be
wrong for Opera, will CERTAINLY be wrong for some other
configuration(s), and, according to the specs will likely be over-ridden
anyway.  

I keep going back to my little chart
(http://wats.ca/resources/accesskeysandkeystrokes/38):  If you were a
content author that needed to "hint" a custom access @key, which one
would you choose, and how would you convey that information to the end
user?  Remember, it won't be from one of the "pre-determined" common
collection @roles already defined by the W3C, but something custom.  As
author, you provide the "machine hint" via @key in your source code, but
what about the human user?  How do you tell them that the accelerator is
available?  You can't tell them that it is key "foo", because it may or
may not be.  So what, exactly do you tell them? (And how?)

I understand that software developers would like a hinting from the
content author - it removes the need for the software to propose one.
But *if* the hint is inappropriate, then what?  Now the software must
reject the suggestion (more overhead?), and then you are back to square
one...  The software must provide a means for the end user to supply an
appropriate user-supplied key mapping, or simply reject the
functionality all-together.  So, would it not follow that, since the
software needs to provide the custom mapping ability to the end user
anyway, that you stop there?  Why do you need/want to add more code to
your (user agent) application?

JF
--
John Foliot  foliot@wats.ca
Web Accessibility Specialist / Co-founder of WATS.ca
Web Accessibility Testing and Services
http://www.wats.ca   
Phone: 1-613-482-7053

Received on Monday, 2 January 2006 15:15:25 UTC