interacting with the cursor:

[Janina Sajka]

Hi, David:

OK. I think you're getting to the nub ov the issue. So, I'll try one
more time. My answers are below, in line with your last post ...

david poehlman writes:
> Janina,
> 
> If I hear you correctly, if you came upon a platform which bbehaved as I 
> describe with no aility to adjust it, you'd say that that would be fine? 

I have come across such a platform, and it was not fine with me,
therefore I abandoned it in favor of one which provided me the freedoms
I prefer.

This, in my view, is the difference between what you're saying and what
I'm saying. Fix the underlying problem, not its symptoms.

In particular, and most importantly, I am unwilling to see any work
around requirements from a closed environment creep into my free
environment as some kind of best behavior guidance or accessibility
standard. This is why I say "I'm sorry if you're having problems with
Windows, but you could ditch those problems by ditching Windows. If you
choose to stick with Windows, than don't push off its problems on me."

> I'm not actually trying to force a standard but would entertain notions of 
> best practice in this area.  I remember during the early dos days, artic and 
> vert if I remember rightly were shunned partly because they were confusing 
> in how they implemented this.  Of course, there were other complaints as 
> well and they were the screen readers that were available at the time.  The 
> cry among users I remember at the time was that editing was difficult for 
> this reason though.

I don't recall much about those particular ATs, but still fail to see
how this could be taken seriously. It would be a problem if the AT were
inconsistent. I don't know why it should be a problem if one AT did
things one way, while a different one did them differently. Most
importantly, I'm not convinced that a "best practice" could be so
succinctly defined, because I expect much depends on the user's cultural
background. So, best ask i18n first.

And, regardless, I fail to see how this is accessibility as opposed to
application usability, which is out of scope for accessibility. It is
certainly easy for us in accessibility to stray, and I'm cautioning that
we should not do so here.

> 
> Would you say that a best practice would be to be able to interact with the 
> character you hear when you move from character to character?  I know I'm 
> getting at this the rong way round, but if you had to compare the two 
> methods of interaction, which would be less error prone, which would be more 
> desirable?  Which would be preferable?...

I think consistancy is the primary requirement. The second requirement,
imho, is to hear the element on which action is taken. In the case of
cursor movement left and right, that we be a char traversed in a text
doc. Up and down would not be a char traversed. Del and backspace would
be some kind of deletion action, so the char heard would be the one that
has been taken away.

But, again, it does not trouble me if del and backspace behave
differently in different environments, it only troubles me if the
behavior is inconsistent and/or unconfigurable.

				Janina