RJython? [was Re: [R-gui] [RGUI] Where do you start?]

[Philippe Grosjean]

>GTk is portable (win, macos, linux)...
>There are some binding (java, python, ecc..)
>Is fast a lot.

OK, that is just blablabla... Here is an example:
http://wiener.math.csi.cuny.edu/pmg Poor Man's GUI, a GUI for R that uses
GTk (I don't want to critisize it: it is a nice project, really). But... is
it faster than R Commander (http://socserv.mcmaster.ca/jfox/Misc/Rcmdr/),
which is similar but uses tcl/tk and runs under Unix/Linux, Windows and
MacOS? Not on my computer. Could it be improved to become faster? Probably.
Does it run on Windows? No. Could it run on Windows in the future? This is
possible.

I think the key is here: currently, there is a tcl/tk interface in R that is
fully operational (although, it can certainly be improved), and there are
several other possible interfaces (most of them are contributed by OmegaHat,
and in particular by Duncan Temple Lang). Are they completely operational?
Well, *almost*. And that is the problem. I really would enjoy to have the
same facility to access Gtk, Java or other than it is currently the case for
tcl/tk from R, on *all* platforms. DTL, could you update the various GTKxxx
packages and make them compatible with the latest versions of R on all
platforms? Is it a huge work? Or is it still not compatible with some
platforms?

Otherwise, I agree: tcl/tk is quite old, lack many features and widgets
compared to other graphical toolkits... But we already had such a discussion
TWO YEARS ago. And there is not much new today.

I have set up a developer's page in
http://www.sciviews.org/_rgui/rgui/devel.html, after an idea from Thomas
Friedrichmeier that was followed by a long discussion about wether it should
be possible, and how to make code independent from a particular graphical
toolkit (the idea is to describe the dialog boxes and then to use an
separate engine to implement it in Gtk or in tcl/tk, or wathever...).
Obviously a huge and ambitious task, but at least we could have tried it.
This is a good way to content everybody, because every particular user could
choose its favorite graphical toolkit. Well, Thomas and I started to draft
it,... and nobody followed. So, I gave up because I need time to develop
SciViews (and other projects), and I have very limited time remaining for
continuing endless discussions that lead to nothing concrete.

Best,

Philippe Grosjean

.......................................................<°}))><....
 ) ) ) ) )
( ( ( ( (   Prof. Philippe Grosjean
\  ___   )
 \/ECO\ (   Numerical Ecology of Aquatic Systems
 /\___/  )  Mons-Hainaut University, Pentagone
/ ___  /(   8, Av. du Champ de Mars, 7000 Mons, Belgium
 /NUM\/  )
 \___/\ (   phone: + 32.65.37.34.97, fax: + 32.65.37.33.12
       \ )  email: Philippe.Grosjean at umh.ac.be
 ) ) ) ) )  SciViews project coordinator (http://www.sciviews.org)
( ( ( ( (
...................................................................

-----Original Message-----
From: r-sig-gui-bounces at stat.math.ethz.ch
[mailto:r-sig-gui-bounces at stat.math.ethz.ch]On Behalf Of Daniele Medri
Sent: Friday, 13 February, 2004 13:20
To: r-sig-gui at stat.math.ethz.ch
Subject: Re: RJython? [was Re: [R-gui] [RGUI] Where do you start?]



> > I wonder if some combination of SJava and RSpython code, probably
> > implemented in Java,  could provide a library that gives R users access
> > to swing, via jython, from R.
>
> Please let's not pursue this as a GUI interface for R for replacing tcltk.
>  We need something fast, light, and that does not use a huge amount of
> memory.

Why don't devel the actually gtk interface?
GTk is portable (win, macos, linux)...
There are some binding (java, python, ecc..)
Is fast a lot.

Do you know Gretl? Take a look at gretl.sf.net for example and think about
GTK
for your scope.

Bye
--
Daniele Medri <madrid AT linuxmeeting.net>
homepage: http://www.linux.it/~madrid/

"Le cose più grandi, nel mondo, vengono compiute attraverso altre cui
non prestiamo nessuna attenzione, piccole cause su cui sorvoliamo e
che alla fine si accumulano." -- Georg Christoph Lichtenberg

_______________________________________________
R-SIG-GUI mailing list
R-SIG-GUI at stat.math.ethz.ch
https://www.stat.math.ethz.ch/mailman/listinfo/r-sig-gui