From octave-graphics-request at bevo dot che dot wisc dot edu  Mon Nov 11 12:36:25 1996
Subject: Re: GUI interface for Octave
From: "John W. Eaton" <jwe at bevo dot che dot wisc dot edu>
To: octave-graphics at bevo dot che dot wisc dot edu
Date: Mon, 11 Nov 1996 12:36:16 -0600

On 28-Oct-1996, Carl Leitner <bigcarl at email dot unc dot edu> wrote:

: 1) The idea is to make the tcl a text function, and when it is called the
: first time, it will initialize the tcl/tk interpreter.  In order to allow,
: as full as possible access to tcl/tk,

>From my perspective, the problem with this is that it will encourage
people to write GUI stuff for Octave directly in Tcl/Tk commands.
Then when Tcl/Tk is no longer the best GUI toolkit for the job, it
will be almost impossible to switch.  Not only will there be Tcl/Tk
commands in the glue that Octave uses for communicating with the
toolkit, but there will likely be a large amount of user code that
would have to be changed too.

I think it is important to make Octave's user-level code (or even very
much of the internals) independent of any given toolkit.  I think this
can be managed by writing an appropriate set of C++ classes, similar
to the way things are done for the new user-defined type system.

After spending a little time looking at Tcl/Tk and thinking about how
to integrate it with Octave (and leave open the possibility of using
other toolkits in the future) I was left wondering about the best way
to handle events.  I suspect that it will be best for Octave to
include everything necessary to collect and respond to events.  I
doubt that a simple call to TkMain will work, but maybe the situation
is not as bad as I think, because Octave's parser only has to respond
to complete input lines that are returned from readline, and not
individual keystrokes.

Do you have any experience working with the event mechanism of Tcl/Tk?

jwe