scicoslab vs scilab

[c***@gmail.com]

I've just seen a release notice for scicoslab. Can anyone tell what
is going on here? A post from Delebecque Francois on this forum
states that scilab is no longer being developed at INRIA? scilab 5.0
was just released. And what is scilabGTK?

[ycollet]

The only things I can say:
- scilab-gtk doesn't exist anymore, now, it's scicoslab. scicoslab
seems to be a fork of scilab-4.1.2 with a gtk interface and an
improved version of scicos.
- scilab-5 is a new version of scilab-4.1.2 (with a lot of
improvements like a java interface and modularisation) with a real
opensource licence (CeCILL).

YC

[c***@gmail.com]

Post by ycollet
- scilab-gtk doesn't exist anymore, now, it's scicoslab. scicoslab
seems to be a fork of scilab-4.1.2 with a gtk interface and an
improved version of scicos.
- scilab-5 is a new version of scilab-4.1.2 (with a lot of
improvements like a java interface and modularisation) with a real
opensource licence (CeCILL).
YC

So, if I understand correctly:
scicoslab has the latest scicos but not the latest scilab features,
and
scilab has the lastest scilab but not the latest scicos?
And, both are being developed under separate organizations of INRIA?
Are there plans to merge them back together? Is this just a license
issue or is there some kind of split within INRIA? What does this
mean for the future of scilab / scicos / scicoslab? This seems like a
pretty serious schism.

It looks to me like you are on your way to developing a serious
competitor to Matlab. I've been a Matlab user for decades, but it is
very overpriced. And their Linux/Unix support has been weak. The
competition is welcome. I believe Matlab was originally an open
source like development. It is good to see the open source concept
return to quality mathematics tools.
I wish you the best of luck!

[ycollet]

Post by c***@gmail.com
scicoslab has the latest scicos but not the latest scilab features,
and
scilab has the lastest scilab but not the latest scicos?

Yes

Post by c***@gmail.com
And, both are being developed under separate organizations of INRIA?
Are there plans to merge them back together?  Is this just a license
issue or is there some kind of split within INRIA?  What does this
mean for the future of scilab / scicos / scicoslab?  This seems like a
pretty serious schism.

Not exactly. Here is an extract from http://www.scicoslab.org:

ScicosLab is the new name of ScilabGtk. This change of name has been
decided in order to avoid all confusion with Scilab, which is no
longer developed at INRIA. ScicosLab is developed by some of the
researchers who originally developed Scilab at INRIA and ENPC.
ScicosLab is used in particular for disributing new software
developments stemming from research activities of the Metalau team at
INRIA and ENPC, such as Scicos and the Maxplus algebra toolbox.

Post by c***@gmail.com
It looks to me like you are on your way to developing a serious
competitor to Matlab.  I've been a Matlab user for decades, but it is
very overpriced.  And their Linux/Unix support has been weak.  The
competition is welcome.  I believe Matlab was originally an open
source like development.  It is good to see the open source concept
return to quality mathematics tools.
I wish you the best of luck!

I totally agree with you.

YC

[ycollet]

If you want some more informations related to scilab 5. Have a look
at: http://www.scilab.org/changes_5/

YC

[Delebecque Francois]

Scilab 5 is developed at Digiteo now. ScicosLab is maintained by the
original developers of Scilab at INRIA and ENPC to be used as the vector
of distribution for their research results, in particular Scicos and
Maxplus algebra tools. For Scicos in particular, ScicosLab plays an
important role because it provides a stable environment for industrial
users.
FD

Post by c***@gmail.com
I've just seen a release notice for scicoslab. Can anyone tell what
is going on here? A post from Delebecque Francois on this forum
states that scilab is no longer being developed at INRIA? scilab 5.0
was just released. And what is scilabGTK?

[c***@gmail.com]

Post by Delebecque Francois
Scilab 5 is developed at Digiteo now. ScicosLab is maintained by the
original developers of Scilab at INRIA and ENPC to be used as the vector
of distribution for their research results, in particular Scicos and
Maxplus algebra tools. For Scicos in particular, ScicosLab plays an
important role because it provides a stable environment for industrial
users.
FD

This does not sound good. Although INRIA and Digiteo are both members
of the Scilab Consortium, it appears that there is a distinct break
between the development of Scilab and ScicosLab and therefore Scilab
and Scicos. Was the break over Java vs GTK? Or personalities? Or
money? (I can't imagine sex was involved.) I hope there is a way to
merge the efforts again. It was already too bad that energy was split
among Octave, FreeMat, and Scilab. While competition with Matlab is
good, I cannot see any good coming with the Scilab / Scicos split.

[ycollet]

Post by Delebecque Francois
Scilab 5 is developed at Digiteo now. ScicosLab is maintained by the
original developers of Scilab at INRIA and ENPC to be used as the vector
of distribution for their research results, in particular Scicos and
Maxplus algebra tools. For Scicos in particular, ScicosLab plays an
important role because it provides a stable environment for industrial
users.
FD

This does not sound good.  Although INRIA and Digiteo are both members
of the Scilab Consortium, it appears that there is a distinct break
between the development of Scilab and ScicosLab and therefore Scilab
and Scicos.  Was the break over Java vs GTK?  Or personalities?  Or
money?  (I can't imagine sex was involved.)  I hope there is a way to
merge the efforts again.  It was already too bad that energy was split
among Octave,  FreeMat, and Scilab.  While competition with Matlab is
good, I cannot see any good coming with the Scilab / Scicos split.

A lot of good comments. But, because I do not belong to any of these
organisations, I will not be able to answer.
(I hope that sex is the evil source of the split :-) )

YC

[Fixpir]

Post by Delebecque Francois
Scilab 5 is developed at Digiteo now. ScicosLab is maintained by the
original developers of Scilab at INRIA and ENPC to be used as the vector
of distribution for their research results, in particular Scicos and
Maxplus algebra tools. For Scicos in particular, ScicosLab plays an
important role because it provides a stable environment for industrial
users.
FD

This does not sound good.  Although INRIA and Digiteo are both members
of the Scilab Consortium, it appears that there is a distinct break
between the development of Scilab and ScicosLab and therefore Scilab
and Scicos.  Was the break over Java vs GTK?  Or personalities?  Or
money?  (I can't imagine sex was involved.)  I hope there is a way to
merge the efforts again.  It was already too bad that energy was split
among Octave,  FreeMat, and Scilab.  While competition with Matlab is
good, I cannot see any good coming with the Scilab / Scicos split.

I totally agree with you.
If I were a Matlab salesperson, and had to argue over the cost of
Matlab over the free Scilab/Scicos, I would just print this page and
give it to my customer with a comment about committing costly
engineering time in such a mess.
IMHO, this kind of decision is absolutely catastrophic in terms of
public relation for Scilab / Scicos/Scicosxxx.
Or maybe there is an advantage in this way of developing incompatible
softwares that I do not understand. I am probably not alone to be
puzzled.
All my respects anyway to all people fighting for the development of
this/these software(s).

(If only sex was the cause of this decision ! At least two people
would have got some benefit out of it ! )

[ycollet]

The thing that threaten me a little bit (with scicoslab) is that we
have no access to the source code (not really open source), there is
no bugzilla thing, there is no mailing list.
When scilab-gtk was open it was difficult to build from source (you
needed to use cvs to get the latest developpement code, and no
released source archive was accessible).
Scilab (the one accessible via www.scilab.org) is a lot more open
source friendly. It is supported by a consortium and I am really
confident in the future of this tool.

YC

[James Koch]

Post by c***@gmail.com

Post by Delebecque Francois
Scilab 5 is developed at Digiteo now. ScicosLab is maintained by the
original developers of Scilab at INRIA and ENPC to be used as the vector
of distribution for their research results, in particular Scicos and
Maxplus algebra tools. For Scicos in particular, ScicosLab plays an
important role because it provides a stable environment for industrial
users.
FD

This does not sound good. Although INRIA and Digiteo are both members
of the Scilab Consortium, it appears that there is a distinct break
between the development of Scilab and ScicosLab and therefore Scilab
and Scicos. Was the break over Java vs GTK? Or personalities? Or
money? (I can't imagine sex was involved.) I hope there is a way to
merge the efforts again. It was already too bad that energy was split
among Octave, FreeMat, and Scilab. While competition with Matlab is
good, I cannot see any good coming with the Scilab / Scicos split.

I'm an industrial user of both Scilab and Scicoslab. I tried Xcos but even the little demo simulation had problems I couldn't figure out. When I tried to run one of my Scilab programs in Scicoslab, it gave an error message about incompatibility relative to reading Excel files. So I have to keep both programs installed. I have had excellent results with Scicoslab simulating an adaptive constant on-time hysteretic switching voltage regulator. Both of these programs are important to me in my work. I'm OK with maintaining them separately, but it does seem inefficient, and less competitive.

[F. S. Farimani]

Hi guys,

As an update on this matter, Apparently ScicosLab is now replaced by Nsp aka Tumbi. I wrote about this over here:

https://www.reddit.com/r/math/comments/b6gn9r/nsp_is_a_gpl_scientific_software_package/

Best,
Foad

[F. S. Farimani]

There are two more siter projects now:

- Nelson https://github.com/Nelson-numerical-software/nelson by Allan Cornet which aims at
- more MATLAB compatibility
- Replacing Java with Qt/QML
- better performance

- Balisc https://github.com/rdbyk/balisc by Dirk Reusch which aims at better performance

I added these to the above post.