GUI tookit for science and education

[Mateusz £oskot]

Hi,

I'd like to ask some scientists or students
which GUI toolkit they would recommend
to develop scientific prototypes (for education and
testing some theories).
I think such toolkit should fill a bit different
needs and requirements:
- very simple to learn
- easy to install
- beautyfiers and advanced features are not required like OpenGL,
direct access to Windows GDI subsystem, beauty look and skinning
- multiplatform required

Let's say someone has big amount of algorithms and
statistical models implemented in Pascal
(not well designed console apps).
Now he want to move on using better and modern language
and GUI toolkit.
Python is seleceted as user friendly and simple
language, Pascal successor.

Thank you for any piece of advice in advance.
Regards
--
Mateusz £oskot, mateusz (at) loskot (dot) net
Registered Linux User #220771

[Paul Rubin]

Mateusz £oskot <se****@signature.net> writes:

Thank you for any piece of advice in advance.

Ask yourself why you want a GUI toolkit. Maybe you can write a web
application instead, and use a browser as the GUI. That's a lot
easier to write (just use html), and makes it trivial to run the
application and the browser on separate machines.

[Alessandro Bottoni]

Mateusz ?oskot wrote:

I'd like to ask some scientists or students
which GUI toolkit they would recommend
to develop scientific prototypes (for education and
testing some theories).
I think such toolkit should fill a bit different
needs and requirements:
- very simple to learn
- easy to install
- beautyfiers and advanced features are not required like OpenGL,
direct access to Windows GDI subsystem, beauty look and skinning
- multiplatform required
First, have a look at:
- wxWidgets (http://www.wxwidgets.org)
- wxPython (http://www.wxpython.org)

Or, as a second choice:
- FLTK (http://www.fltk.org/)
- PyFLTK (http://pyfltk.sourceforge.net)

There are many others GUI toolkit around. Just search Google for
"multiplatform Python GUI toolkit" or something like that.
Let's say someone has big amount of algorithms and
statistical models implemented in Pascal
(not well designed console apps).
Now he want to move on using better and modern language
and GUI toolkit.
Python is seleceted as user friendly and simple
language, Pascal successor.

There was a python to pascal automatic converter at:

http://no.spam.ee/~andreie/software/...ish-index.html

(Now offline)

CU
-----------------------------------
Alessandro Bottoni

[TPJ]

Recently I was considering the choice of PyGTK or wxPython. They are
both rich GUI libraries, and they both are cross-platform ones (well...
they work on GNU/Linux and on Windows).

I chose PyGTK, because it has *much* better documentation (I wasn't
very happy when I had to look for information in documentation of
wxPython - and lose a couple of hours - when I wanted to do something
really simple...) and it seems to work stable on Windows.

Let's say someone has big amount of algorithms and
statistical models implemented in Pascal
(not well designed console apps).

OT: I would recommend the use of Python + Numeric and, eventually, C
(when the performance becomes really crucial). I'm working on genetic
algorithms and - for now - the combination of Python and C was the best
I got ever. In a few days I will run the first version of GAs
implemented with Numeric.

[phil hunt]

On Mon, 15 Aug 2005 04:59:29 +0200, Mateusz £oskot <se****@signature.net> wrote:

Hi,

I'd like to ask some scientists or students
which GUI toolkit they would recommend
to develop scientific prototypes (for education and
testing some theories).
I think such toolkit should fill a bit different
needs and requirements:
- very simple to learn
- easy to install
- beautyfiers and advanced features are not required like OpenGL,
direct access to Windows GDI subsystem, beauty look and skinning
- multiplatform required

Tkinter is the default GUI for Python, it runs on lots of platforms
and often comes already installed (on Linux or Unix distributions).

Is it "very simple to learn"? Well I suppose that depends on what
you mean by "very simple".

--
Email: zen19725 at zen dot co dot uk

[phil]

Tkinter is the default GUI for Python, it runs on lots of platforms
and often comes already installed (on Linux or Unix distributions).

I use Tkinter for a geometry course. I think it is fairly easy to
learn, much easier than say VB. VERY portable and pretty well
documented and very well supported on this list.
It doesn't have much math built in. For functions you have to
plot points.

[Grant Edwards]

On 2005-08-15, phil <ph***********@anvilcom.com> wrote:

Tkinter is the default GUI for Python, it runs on lots of platforms
and often comes already installed (on Linux or Unix distributions).

I use Tkinter for a geometry course. I think it is fairly easy to
learn, much easier than say VB. VERY portable and pretty well
documented and very well supported on this list.
It doesn't have much math built in. For functions you have to
plot points.

If you want to plot stuff, the gnuplot-py module is very easy
to use. http://sourceforge.net/projects/gnuplot-py/

The one feature that I'd really like to add is the ability to
plot a python function object. Currently you can plot a
function specified by a string (e.g. "sin(x) * sin(x)**2"), or
a sequence of data points. It would be nice to be able to pass
an actual function.

--
Grant Edwards grante Yow! hubub, hubub, HUBUB,
at hubub, hubub, hubub, HUBUB,
visi.com hubub, hubub, hubub.

[Mateusz £oskot]

Thank you all for valuable responses.
I think I will stick to Tk and Tkinter.

Cheers
--
Mateusz £oskot, mateusz (at) loskot (dot) net
Registered Linux User #220771

[Markus Rosenstihl]

>> It doesn't have much math built in. For functions you have to

plot points.

If you want to plot stuff, the gnuplot-py module is very easy
to use. http://sourceforge.net/projects/gnuplot-py/

The one feature that I'd really like to add is the ability to
plot a python function object. Currently you can plot a
function specified by a string (e.g. "sin(x) * sin(x)**2"), or
a sequence of data points. It would be nice to be able to pass
an actual function.

matplotlib is also ver good possibility

[Mateusz �oskot]

Markus Rosenstihl napisał(a):

It doesn't have much math built in. For functions you have to
plot points.

If you want to plot stuff, the gnuplot-py module is very easy
to use. http://sourceforge.net/projects/gnuplot-py/
[...]

matplotlib is also ver good possibility

Thanks, I'll try it.

Cheers

--
Mateusz �oskot, mateusz (at) loskot (dot) net
Registered Linux User #220771

[Thomas Bartkus]

"Paul Rubin" <http://ph****@NOSPAM.invalid> wrote in message
news:7x************@ruckus.brouhaha.com...

Mateusz £oskot <se****@signature.net> writes:

Thank you for any piece of advice in advance.
Ask yourself why you want a GUI toolkit. Maybe you can write a web
application instead, and use a browser as the GUI. That's a lot
easier to write (just use html), and makes it trivial to run the
application and the browser on separate machines.

Wow Paul!
I just couldn't help zeroing on that comment. a lot easier to write (just use html),
I would have used adjectives like "clunky" and "limited" when talking about
using an html in a browser app. Particularly if we are talking about high
powered math/graphs as we often are in the science apps indicated in the
original post.

I would take MS Excel/VBA as the premier fat client prototyping tool/GUI
toolkit for science & education. How would one go about replicating any of
that in an HTML/browser app? How do we get to "easier to write"?
Ask yourself why you want a GUI toolkit.

I just did. The answer is that I don't *think* you can do much of that with
html.

Then again - I was wrong once :-)
-Tom

[Mateusz £oskot]

Thomas Bartkus napisa³(a):

"Paul Rubin" <http://ph****@NOSPAM.invalid> wrote in message

I would take MS Excel/VBA as the premier fat client prototyping tool/GUI
toolkit for science & education. How would one go about replicating any of
that in an HTML/browser app? How do we get to "easier to write"?

As I said, I'm looking for multiplatform solution.
So, I will go with Python, Tk and C++ (for algorithm and critical parts
of applications). Tk is simple, very simple, Python I like, C++ I love,
as I'm professional C++ programmer.
Students and profs will use Python with Tk mainly.
I will convert some of libraries with algorithms they will prototype in
python (and exsiting in Pascal) to C/C++ and bind them with Python.

There is also C++/Tk by Maciej Sobczak, great tool.

Cheers

--
Mateusz £oskot, mateusz (at) loskot (dot) net
Registered Linux User #220771

[phil]

As I said, I'm looking for multiplatform solution.
So, I will go with Python, Tk and C++ (for algorithm and critical parts
of applications). Tk is simple, very simple, Python I like, C++ I love,
as I'm professional C++ programmer.

Good choice!