🚀 go-pugleaf

I am considering using interviews for buiding the GUI for a project I
am working on. My experience with interviews is solely from reading
this newsgroup to see what goes on and having read some papers on IV.
I have just finished reading the 'Tale of Two Toolkits' paper by
Bashir and Sternlicht.

The project I am working on deals with signal monitoring. The whole
system is supposed to work in real-time (soft), therefore speed in
graphics is an issue. In the present application the signals
do not exceed frequencies of a few hundred hertz and most signals
will be an order of magnitude slower than that. Assume a sparc10
platform.

One of the things that I would like to try is to have a programmed
reconfiguration of the display to draw the attention of the user to
significant events in the signals when they occur (amongst other
things, I am trying to address the problem of information overload for
the user).

The monitoring/processing part and the GUI will be separate UNIX
processes using message passing for control and having a shared memory
segment for data. The programming of the system so far has been done
in C++.

As I see it the advantages of using IV for me are:
1) continuation of the OO support of C++ throughout the development
of the GUI
2) high level concepts in IV will hopefully help me in developing the
reconfiguation capability.

My main concern though is whether those same high level concepts that
make interviews a nice development tool get in the way of achieving
fast graphics. The Bashir/Sternlicht paper notes a performance
disadvantage of IV compared to Xt in event handling. 

I would appreciate some comments, ideas or past experiences related to
this, and opinions as to what extent IV is suited for developing a GUI in a
real-time application.

Thanks,

-- Eric.