Tom,
At 01:32 PM 11/27/2002, Tom Whittaker wrote:
To that end, I'm always open to suggestions for changes, additions, etc.
That has been my experience in the 4-5 months I've been using VisAD, and I
do appreciate it. Everyone associated with the project has been very
generous with their time and advice.
Like I (and others) have mentioned, it is sometimes practical to interface
VisAD Data objects to other graphics libraries...but I'm curious to know
what basic functions are missing or awkward to use...aside from the
comments on aesthetics...? I put things like drawLine and drawString in
there to get started, and would be very happy to add more interfaces
and/or options to existing ones.
Off the top of my head, I think Maohai's (apologies if I mangled your name
:) suggestions are good: the generic 2D toolkit is not so high level. 2D
graphs are much more common among scientists-in-general than 3D. It's true
that 3D can convey much more information on a screen, but on paper, for
publication in greyscale, 2D is often better. I'm sure that VisAD can do
things like legends, histograms, error bars, etc. But it's just not
available easily from Jython, and it's not that clear how to get it from
Java either. Most of the Java examples (and there are plenty, which is
great) require 3D even to compile, which I don't have installed on the
machine where I do most of my development (a Linux cluster where I do not
have root privileges). I understand that 3D is the main target of VisAD,
but I think 2D is the on-ramp for most people. It takes a long time for
casual users (as most scientists are) to learn to use a software tool
(think LaTeX or Word, and ask yourself how many people you know who
*really* know how to use it). You start with the simple things, you do them
over and over, to reinforce learning. If the simple things are too
difficult, you stop and look for another way.
Are there things in there that might be a candidate to add to
JPythonMethods? I've run across some iterative tasks that when done in
Jython were way too slow, so I recast them in Java.
The Jython interface will only evolve as the people using it push it.
Of course, I'm prepared to admit that perhaps VisAD is just not the right
tool for what I'm trying to do right now. Perhaps it is overkill. It's a
view I expressed when I first joined the list. I'm going to revisit my
dormant visualization project. The conversation of the last week has
rekindled my determination to get the most out of VisAD. I'll get back to
you with more specific ideas about what I'd like to see in VisADJython.
Someone last week said something along the lines of "...if it seems like
we're whining, it's not that we're ungrateful, it's just that we can see
the potential of VisAD, and want to get at it more easily..." That's
certainly how I feel.
-Frank
PhD, Computational Biologist,
Harvard Medical School BCMP/SGM-322, 250 Longwood Ave, Boston MA 02115, USA.
Tel: 617-432-3555 Fax:
617-432-3557 http://llama.med.harvard.edu/~fgibbons
