Re: Auditory interface ideas, what would help?

From: Andreas Stefik <stefika@xxxxxxxxx>
To: programmingblind@xxxxxxxxxxxxx
Date: Tue, 1 Sep 2009 21:00:30 -0500
Susan,

My team and I are working hard to get our tools integrated into NetBeans.
We've made enormous progress this summer but still have several months to go
before we can even consider a release (and realistically, probably longer).
Our work is currently being done at Southern Illinois University
Edwardsville (Where my grad student and I are at), and at Central Washington
University, where one of my old students is doing some work on the virtual
machine we've built. We have also been very fortunate to have Louie Most, a
retired blind programmer that lives in washington help to keep us honest
about all things blind and who has been contributing some text-to-speech
code on Macintosh (although we could use some help with other platforms!).

To answer your question though, what we've been trying to figure out is the
best possible performance one can achieve without visual stimuli for
programming. When we first took measurements of a blind programming
environment (Visual Studio 2005), under our experimental conditions, we
found about a 80% drop in performance compared to visual programming.
However, with our environment at the time, we were able to drop that down to
about 30% in the programming language C. I'm leaving out a lot of details
from our experiments, but in short, we considered this to be a pretty
stellar step in the right direction.

That was a year and a half ago and at the time, we were using custom, home
grown, environments. Now, we've been working incredibly hard to get a new,
and substantially more powerful and robust, version of our tool into the
NetBeans programming environment. Right now this includes our own custom
programming language, virtual machine, and a plethora of other tools that
are being experimented with to see if they help (e.g., code navigators,
special codes in the virtual machine, changing around programming language
constructs, editor "hints" as you type).

In my ideal world, what will happen is that we'll be able to make progress
using custom languages, then eventually help some larger corporations get
the results from our experiments integrated into other languages. That's a
tall order, though, right now, there's only three of us, and designing IDE's
and custom compilers/debuggers is not, even remotely, easy.

Cheers,

Stefik



On Tue, Sep 1, 2009 at 9:02 AM, Stanzel, Susan - Kansas City, MO <
susan.stanzel@xxxxxxxxxxxx> wrote:

>  Hi Listers,
>
> What IDE does this work with? I am learning java and feel at a
> disadvantage.
>
> Susie Stanzel
> Programmer U.S.D.A.
>
>  ------------------------------
> *From:* programmingblind-bounce@xxxxxxxxxxxxx [mailto:
> programmingblind-bounce@xxxxxxxxxxxxx] *On Behalf Of *Andreas Stefik
> *Sent:* Monday, August 31, 2009 10:28 PM
> *To:* programmingblind@xxxxxxxxxxxxx
> *Subject:* Re: Auditory interface ideas, what would help?
>
> Vladyslav,
>
> Thanks for pointing this out. Ironically, my dissertation was done, in
> part, on this very topic. We've got a whole army of techniques for handling
> it and have put in several hundred hours of user testing on it.
>
> So, yaa, in short, I couldn't agree more, virtually anything related to
> scoping is a real pain in audio. In our current version of the tool, we get
> pretty good results with helping folks identify things related to scope, but
> I think there's some improvements that can be made to our design as well.
>
> Anyway, good point,
>
> Stefik
>
> On Mon, Aug 31, 2009 at 4:35 PM, Vladyslav Kutsenko <kutsenko@xxxxxxxxx>wrote:
>
>> Hello Andreas,
>>
>> Andreas Stefik schrieb:
>>
>>> I've got an overwhelming number of ideas about what could help, but I
>>> can't implement them all and would like to get some feedback from the
>>> community. Specifically, we're looking for some ideas in two main areas,
>>> code completion and auditory navigation.
>>>
>>> 2. Auditory navigation --- I have a graduate student currently working on
>>> a blind code navigation system. Right now, the system allows you to jump
>>> around the source window and find variable declarations, method
>>> declarations, and other similar things, but we want to expand the navigation
>>> to make things easier.
>>>
>>
>> I think it is very important to get information on the context the code is
>> in. So for example, it is sometimes quite time consuming to read code which
>> contains  numerous into each other embedded variable scopes. In my
>> experience, maintaining of such code, often results in algorithm errors. The
>> information about the level of scope recursion which could also contain
>> additional information about the scope properties, e.g. "if block", "for
>> loop ...) would be in my opinion of huge importance.
>>
>>>
>>> Thank you for your work!
>>
>> Vladyslav
>>
>>
>> __________
>> View the list's information and change your settings at
>> //www.freelists.org/list/programmingblind
>>
>>
>
References:
- RE: Auditory interface ideas, what would help?
  - From: Stanzel, Susan - Kansas City, MO
Re: Auditory interface ideas, what would help?

Other related posts:
