From octave-graphics-request at bevo dot che dot wisc dot edu  Fri Feb 25 12:01:26 2000
Subject: Re: gnuplot (Was: Re: New on the list : new plotting tool) 
From: Teemu Ikonen <tpikonen at pcu dot helsinki dot fi>
To: "John W. Eaton" <jwe at bevo dot che dot wisc dot edu>
cc: Tim Piessens <Tim dot Piessens at esat dot kuleuven dot ac dot be>,         Discussion list octave-graphics <octave-graphics@bevo.che.wisc.edu>
Date: Fri, 25 Feb 2000 20:01:24 +0200 (EET)

On Fri, 25 Feb 2000, John W. Eaton wrote:

> My point was that it is bad design to have the user interface closely
> tied to the functionality of the program.  If a plotting package
> *requires* a particular GUI to operate at all, I would consider it to
> be a problem.  If there is some separation between the funtionality of

Ok, what we have here is perhaps a slight dichotomy of the concept
'plotting'. What I mean is that I want to (1) to visualize data, zoom in
and out and check some regions more closely and yes, I'd like to be able
to do it with a pointy clicky interface. When I'm done with my analysis I
want to (2) make some publication quality figures which preferably should
resemble closely the ones I see on the screen.

Now the second function doesn't explicitly need a gui, but the
visualization part IMHO definitely does need one. Using a free and popular
gui toolkit should not be considered extra baggage in this case, on the
contrary, it will reduce the wheel reinventing.

But, as tastes seem to differ quite alot in this matter, I think the
smartest thing to do is to try to make Octave as independent as possible
from a particular plotting/visualization package. 

t.