[ale] Object Model on Linux

[Douglas Todd]

Some ideas on implementation:

Brad Dixon has made a demo of IBM's Common User Access interface
available at http://r52h146.res.gatech.edu/~bdixon/ibmcua.zip. It's
a neat idea, and may be a worthwhile user interface to implement.

In the demo, the end user is able to record a macro which ties together
various programs, pipes the output from one program to another, and is
allowed to save the macro. Some extensions that would be nice are the
ability to branch and loop, define variables and call, or jump to other
macros. ( which in turn, could then call other programs ). To the user
the interface is drag and drop.

I wonder how much work it would be to collect the different parameters
for the binutils suite, and some of the common operations. I'm thinking
that we could write a drag and drop interface where the icons represent
utilities like grep, sed, awk, etc, and have a 'registry' that
identifies 
how the utilities are called, ( the command line arguments ), 
icons that represent the utilities, and a shell program that allows 
the user to graphically connect these programs together. 

For each utility, there would be an icon, and a set of data base entries
that
describes each utilities operations and functions. The GUI provides the 
'test' function, and may either execute the utilities directly, or build
a shell script and then execute the shell. Functionally, the GUI takes
the place
of BASH, and uses the 'registry' data base to format the command line 
arguments. 

I can see some AI, and quite a bit of X programming in this project.  

Any takers ? -)

Doug.

-- 
dwt at atlanta.com
doug.todd at sciatl.com