Re: A History of Variables

  • From: Marlon Brandão de Sousa <splyt.lists@xxxxxxxxx>
  • To: programmingblind@xxxxxxxxxxxxx
  • Date: Thu, 15 Oct 2009 08:18:00 -0300

Hello Andreas,
First of all let me to explain a little bite about what is context keys.
When one is blind, in general they lack information. This happens
mainly because while a sighted individual can quickly see groups of
information and select the part they need, because we have sequential
access to information, getting what we need usually become a painful
process, cinse we need to go through a lot of garbage to hear the info
we need.
The problem is that, while a peace of info can be in a given moment a
garbage, it can become a necessary peace in another moment. This
forses all access tecnology to make a hole bunch of keystrokes
available, each one to give access to a peace of info.
Now supose the following:
1- I need to try a program that I have never seem befor running. I
would set break points and probably use the arrows in a tree structure
to browse pretty much everything that is happening inside the
software.

2- I have worked on the software trhee mmonths and know even the
assembly code generated by memory. For some stupid reazon something
isn't working as it should, so I actvate the debugger and run directly
to the part where I know the probleem is. In this case the quick keys
will allow me almost imediat access to where I want to go.
JAWS uses an interesting schemma to give quick access to word
documents and internet explorer, setting letters to go to thje next
and previous heading, next and previous combo box and etc.
This accelerates pretty much the navegatiom process when you now the
site structure your are using, while there are no enourmous keystrokes
combinations like control + alt + shift + ins + winKey + home + a to
access the third variable in the program.
Quick keys can be turned om or off and are in its majority simple keys
in the keyboard and intuitive, so memorizing them is not a big issue.
Other than this, I suggest that you take a look at the navegation
schemma used in the homerow utility in JAWS. It is pretty usable to
navegate structures. For example, I suggest that once you got yourself
in a variable either using a quick key or the arrows than pressing tab
or ssiift tab lets you navegate through its history values and that
pressing the space bar will open a html interface or text file with
the history written down om it.
Thanks,
Marlon

2009/10/14, R. Haynie <rhaynie@xxxxxxxxxxx>:
> Ah, Ok, thanks for the breakdown.
> You are really getting some of our juices going with this development.
> Thanks.
>
>
> Andreas Stefik wrote:
>> Rodney,
>>
>>     Well does your debugger have a window that shows all of the
>>     current variables that are in scope?  I know you mentioned the
>>     watch window, but I assume that is only watching variables that
>>     you identify.
>>
>>
>> Yes, it has such a window. I must have mis-spoke, we actually have the
>> opposite right now, it shows only those variables in scope, we do not
>> have the self identified variables yet (e.g., watch window). I suppose
>> our window would really be called a local variables window.
>> Eventually, this window will be combined with a watch window, just
>> like in NetBeans (but with history), although we haven't finished that
>> yet. We want to get the history stuff in there first, as the watch
>> window actually requires some extra parsing routines. It's not hard,
>> but it is time consuming.
>>
>> That's why we're asking for ideas about history, at least. We've been
>> batting around a number of designs, but haven't quite coalesced on one
>> we find really hip, yet.
>>
>>
>>
>>     If there is a window that shows all variables in scope, then I
>>     would suggest spawning from this window.
>>     Alex had a great suggestion on the resulting user interface from a
>>     keypress.  That interface could be called by pressing enter on the
>>     variable in the list.
>>
>>
>> Sounds good, I'll put another vote for this type of interface in place.
>>
>>
>>
>>     Of course this would work only for variables that are currently in
>>     scope.  Are there situations where you would want to see variables
>>     that no longer exist?
>>     -Rodney
>>
>>
>>
>> Yes, there are actually. Lets take what is, by far, the most trivial
>> example I can think of. Suppose you had the following code in our
>> language:
>>
>> integer a = 5
>> a = 10
>> a = 15
>>
>>
>>
>> You then run the debugger to the third line (a = 15), but do not
>> execute it. In our variables history window, you can get at any old
>> values of that value: a, which in this case would be only 5. The
>> current value is 10. Similarly, if you stepped over a = 15, you could
>> get 15, 10, and 5.
>>
>> Now, it gets much more complicated once you add in custom functions,
>> scopes, objects, packages, and whatever else, but that's the very
>> basic idea. And of course, all of this is designed with auditory
>> feedback as the debugger executes. In this case, our debugger would
>> say, as you step over, the following:
>>
>> a to 5
>> a to 10
>> a to 15
>>
>> Anyway, that's the idea.
>>
>> Stefik
>
>


-- 
When you say "I wrote a program that crashed Windows," people just
stare at you blankly and say "Hey, I got those with the system, for
free."
Linus Torvalds
__________
View the list's information and change your settings at 
//www.freelists.org/list/programmingblind

Other related posts: