Interest in a "scientific" Linux distro is becoming increasingly evident as scientific users are looking for both a single source for scientific software and a desktop environment that is easy to get used to. The previous Linux in Science report described how one SEUL-sci developer, Matias Mutchinick, was involved with an effort at the academic level to produce a distro of software relevant to academic work. Many people feel that having a distro specifically for scientific users in general is appropriate. Danny Gray expresses the sentiment well:

I am currently working in a lab that does Immunology, Molecular and Biochemistry. We have been investigating Linux in the lab. I have played with several distro's and as of now have decided on Corel Linux due to the ease of getting Windows networking integration up and the easy interface. This decision was necessitated by a boss who feels that Linux should be as easy as Windows for him to use it. I agree and was hoping SEUL would come up with a GNU alternative to the current distributions. I'll be honest and say that I have not tried "Independence Linux" yet but will do so soon. I want to try it on a Thinkpad 755CX that I have in the lab.

I also think that if a GNU distribution targeted specifically at science is not made, the linux community will be doing itself a disservice. I like the way Debian and others do installation options. I would love to see a distribution that had science package groups something like this:

• General Science (Unit converters etc.)
• Math
• Physics
• General/Organic Chemistry
• Biochemistry
• Anatomy/Physiology
• Molecular
• General Biology
• Medical (study helps & electronic patient records [Circare, FreeMed, Littlefish, LinuDent,LAMDI])

This would be a very useful project to undertake for our own needs and those of the scientific community in general. A similar project hosted at SEUL, The Independence distribution, is led by Jean Francois Martinez and focuses on the end user. Amongst other considerations, Indy provides numerous software packages particularly appropriate to dial-up users. The Indy archives provide a wealth of information on the issues that the creators of a distribution must deal with.

Such a project, however, would require the involvement of a number of individuals in various roles:

• Installer hackers
• Program testers
• Program packagers
• Documentation writers
• Install testers

Here are some more links to scientific software that we've found...

Moldy

Moldy is particularly useful because in addition to calculating radial distributions and thermodynamic averages, it can interface to at least two popular ray tracers, Rayshade and PovRay to produce amazing and complex molecular images, as seen on the Moldy homepage.

Keith indicates in his email that Moldy has been used on a number of systems, even on a Cray T3E! The Moldy home page can be found at http://www.earth.ox.ac.uk/~keith/moldy.html. (GPL)

NetCDF

Another link suggestion for the libraries section:

NetCDF (network Common Data Form) is an interface for array-oriented data access and a library that provides an implementation of the interface. The netCDF library also defines a machine-independent, self-describing, format for representing scientific data.

http://www.unidata.ucar.edu/packages/netcdf/

netCDF interfaces are available for many languages (including C,C++, and Python) and for many packages (including MATLAB and IDL).

NetCDF is published under a "free licence"

IRAF

Astronomers and astrophysicists most commonly use the packages SM (Super Mongo) and IDL. These are mostly plotting packages, but either can do much more. Neither are free, but both run on linux. For image analysis, IRAF and IDL are most common. Iraf is free and runs on linux. For image file formats, we use FITS (Flexible Image Transport System). FITS is very old, so I'm sure it could be improved upon, however it was definitely withstood the test of time.

IRAF can be found at http://iraf.noao.edu/iraf-homepage.html and is published under a free licence.

Newmat

I thought this package might be useful - Newmat, a C++ library for extensive matrix manipulation. I've been using it since 1997. It's not the fastest, but it really helps me to write code quickly.

Remembering only too well the importance of getting a piece of code 'working well enough' soon rather than it 'working optimally' later, I can certainly understand the utility of this library.

Newmat's home page can be found at http://webnz.com/robert/cpp_lib.htm, and the Freshmeat Appindex link can be found at http://www.freshmeat.net/appindex/2000/03/24/953875728.html

Scilab

Scilab is a scientific software package for numerical computations in a user-friendly environment. It features:

Elaborate data structures (polynomial, rational and string matrices, lists, multivariable linear systems,...). Sophisticated interpreter and programming language with Matlab-like syntax. Hundreds of built-in math functions (new primitives can easily be added). Stunning graphics (2d, 3d, animation). Open structure (easy interfacing with Fortran and C via online dynamic link). Many built-in libraries:

• Linear Algebra (including sparse matrices, Kronecker form, ordered Schur,...).
• Control (Classical, LQG, H-infinity,...).
• Package for LMI (Linear Matrix Inequalities) optimization.
• Signal processing.
• Simulation (various ode's, dassl,...).
• Optimization (differentiable and non-differentiable, LQ solver).
• Scicos, an interactive environment for modeling and simulation of hybrid systems.
• Metanet (network analysis and optimization).

Symbolic capabilities through Maple interface.

Scilab can be found at http://www-rocq.inria.fr/scilab/ and is distributed under a free licence.

XPP

In his email, Felix Mic suggests that XPP would be useful for mathematicians, chemists, biologists and others involved in simulation work.

XPP is a system for the analysis and simulation of dynamic and probabilistic phenomena. It solves differential equations, difference equations, delay equations, functional equations, iterative equations, boundary value problems, and stochastic equations, all combined with probabilistic models. It is now available for free as a program running under X11 and UNIX. A variety of numerical methods are employed so that the user can generally be sure that the solutions are accurate. Examples include connectionist type neural nets, biophysical models, models with memory, and models of cells with random inputs or with random transitions. A graphical interface using X Windows as well as numerous plotting options are provided. PostScript output is supported.

XPP is a reliable software. Timothy S. Gardner used it to obtain results that were published in the high repute scientific magazine Nature, vol 403, 20 january 2000.

The current version is located on sparse web page at http://www.cnbc.cmu.edu/~bard/files.html and additional information can be found at http://www.cs.cmu.edu/afs/cs.cmu.edu/project/ai-repository/ai/areas/neural/systems/xpp/

Thanks to everyone who emailed me about these and other apps. I look forward to reading about useful scientific applications for Linux, and especially HOW this software is being used.

-- Pete St. Onge (pete@seul.org)