Re: Thoughts on mrow default intent

On 12/16/20 5:49 PM, David Farmer wrote:
> 
>> 3. I would even suggest stepping away from the desire to have "content trees" in the 
>> default
>> rules for anything that isn't trivial or extremely constrained. Neil's examples 1,2,3,6,7
>> above can be narrated without any additional annotations, and without defaults, just 
>> reading
>> through the presentation tree as usual. So why do anything at all for them?
> 
> Some of us (me!)  [ "me with emphasis" not "me factorial" ]
> primarily care about enabling correct pronunciation by a screen
> reader.  But others (perhaps the majority in this group but perhaps
> not in the general population) want to enable computation.  We have

I'm finding myself agreeing with both Moritz and David, here.
While I'm more in tune with computation than accessibility,
I'm sympathetic to the need for accessibility.

For computation, we already have Content MathML and OpenMath.
Indeed, they both have major problems with adoption. Exactly
why is still unclear to me, and thus I can propose no solution yet.
But it is becoming clear to me that while re-encoding either one in an
intent/semantic attribute is possible, it's not really much of
a solution, either, and just as likely to be adopted.

For accessibility, it seems intuitive that just a few strategic
clues ought to suffice, but which clues? I find myself always
sliding back to a content-oriented approach, since at least I
know what it "means".
I still have no clear idea of exactly what information needs to
be added specifically for true accessibility.   Just as the
trees and information needed for presentation & content are quite
different, it seems that there are different points of view
on accessibility that also imply quite different information sets.
"For this example you should say This"; "No, That"; and for this
other indistinguishable example, you'd want something different entirely.
To the extent that there's even a "tree" involved, it appears
sometimes to track more closely to the presentation tree,
other times to the content tree, depending on exactly what
you want to be vocalized.

[Sorry, you a11y experts likely know exactly what you mean, want
and need, but it's failing to come across to me]

> already seen that those goals are not completely aligned.  An example
> from a recent call is:
> 
> \int dx/x
> 
> and
> 
> \int 1/x dx
> 
> were said to have the same "intent", while I doubt authors
> intend those to be pronounced the same.
> 
> I recognize the value of enabling statements to be read into a
> computer for computation, even though that is not my use case.
> But I don't want the computational use of "intent" to lead to
> incorrect pronunciation.
> 


-- 
bruce.miller@nist.gov
http://math.nist.gov/~BMiller/

Received on Thursday, 17 December 2020 00:39:33 UTC