[ossrp-control] Re: What Is A Screen Reader?

Wouldn't this be right up genetic programming's alley?

Take care,
Sina

-----Original Message-----
From: ossrp-control-bounce@xxxxxxxxxxxxx
[mailto:ossrp-control-bounce@xxxxxxxxxxxxx] On Behalf Of Will Pearson
Sent: Sunday, May 29, 2005 3:44 PM
To: ossrp-control@xxxxxxxxxxxxx
Subject: [ossrp-control] Re: What Is A Screen Reader?

Hi Pete,

Apologies for the delay in getting back to you.  I've been being a good grad
student to, and getting some work done on a haptech project I'm currently
involved with *smile*.

I agree that an interface designed around the presentation modality being
used is the best approach, but I would tend to view this as requiring human
intervention to produce.  The need for human intervention comes down to
semantics and the ability to decide on the most appropriate encoding scheme
for those semantics in a given context.  Additionally, if it's a conversion
between two different types of physical presentation, then the human ability
to learn and decode the encoding scheme used in the physical presentation of
the interface being converted.  Automatic and semi-autonomous conversion is
possible in situations where the encoding scheme is constant in a given
context, but it would still be semi-autonomous in nature as it would require
a human to initially figure out the encoding scheme being used and to write
some rule based AI detailing that scheme.  The consistent nature may pose
some difficulties in applying this to applications, although it can be used
for things such as diagrams, where I've proposed a system for conversion of
SVG diagrams.  The semantic encoding scheme used in application interfaces
isn't always constant throughout the scope of an application, so there would
have to be multiple sets of AI rules for accurate extraction of the
semantics contained within an interface, which would likely only be narrowly
less work than creating an interface by hand.  It will be interesting to see
what happens with a couple of AI projects that are currently being
researched.  The systems being investigated are capable of building
hypotheses and then evaluating them to see whether they're true in a given
context.  This is interesting, as it's a step towards automating the process
that humans use to determine if an semantic encoding scheme is valid in a
given context.  So, you could teach the system a set of encoding schemes and
allow it to determine which to use for a given context.

Will
----- Original Message -----
From: "Peter Parente" <parente@xxxxxxxxxx>
To: <ossrp-control@xxxxxxxxxxxxx>
Sent: Saturday, May 07, 2005 12:47 AM
Subject: [ossrp-control] Re: What Is A Screen Reader?


> Hi Will,
>
> Thanks for the feedback and the references. I've been a good little grad 
> student and read Mynatt/Edward's papers and Miller's paper before. I'll 
> have to look for your paper in the HCI book when it comes out.
>
> I think I see what you're getting at with a tiered approach too. One of 
> the benefits of modeling programs in terms of domain and medium 
> independent interaction patterns in Clique is that you don't necessarily 
> have to have a GUI program underneath. If you wanted to, you could use 
> Clique as a platform for authoring audio-based applications from scratch. 
> Thus, if a developer wanted to migrate from a mostly automatic adaptation 
> of his graphical interface to a more audio aware application, he could do 
> so within the same Clique framework.
>
> Pete
> To post to the list, send a message to:
> ossrp-control@xxxxxxxxxxxxx
> To unsubscribe, send a message to:
> ossrp-control-request@xxxxxxxxxxxxx
> and set the subject field of the message to "unsubscribe" (without the 
> quotes
>
> 


To post to the list, send a message to:
ossrp-control@xxxxxxxxxxxxx
To unsubscribe, send a message to:
ossrp-control-request@xxxxxxxxxxxxx
and set the subject field of the message to "unsubscribe" (without the
quotes



To post to the list, send a message to:
ossrp-control@xxxxxxxxxxxxx
To unsubscribe, send a message to:
ossrp-control-request@xxxxxxxxxxxxx
and set the subject field of the message to "unsubscribe" (without the quotes

Other related posts: