Software at watzmann.net

You can find the following programs on this page:

g+f

Valentin Kamyshenko has written a nice tool to generate wrappers of FORTRAN routines for use with GUILE. If you have questions about it, please mail them directly to Valentin

find.scm

An implementation of the Unix favorite find(1) in Scheme (for GUILE, to be exact. Your mileage with other Scheme implementations might vary. You definitely need some POSIX system calls). Download it.

gtk-du

gtk-du shows disk usage. It helps you to get an overview of the sizes of directories and to clean them up by supplying a suitable GUI. Download it.

Here is a screen shot.

You need GUILE, GTK+ and the guile-gtk bindings in order to be able to use gtk-du. Be warned that you need the latest development versions of these programs. It is therefore not entirely trivial to get the right setup going (but it's not hard either if you know what a .tar.gz file is.)

bsbound

The MATLAB implementation of the bounds described in the paper Tight linear envelopes for splines, to appear in Numerische Mathematik, can be obtained as

The file README in this archive contains all necessary instructions on how to use the MATLAB routines.

The linear envelopes for splines computed by these routines give better piecewise linear enclosures of a spline than the traditional convex hull and are simpler to compute than the convex hull. They are simple offsets of the control polygon by a product of second differences and precomputed constants which only depend on the B-spline basis used, but not the control points.

repde

repde is a Maple program that takes an arbitrary linear partial differential equation (PDE) and a user defined reparametrization of the domain and makes the changes necessary to obtain a PDE over the new domain. For example, it can be used to automatically change a PDE to polar coordinates.

As it is written now, it produces output for Purdue's own ELLPACK PDE solver. It should not be too hard to change the code to output a problem description for other problem solving environments.

The code is tested for second order 2D problems, but can also handle higher-order and 3D problems. The 3D code is not very well tested.

You can download a report that describes some of the theoretical considerations and how to use repde or a tarball with the code, some examples and the report.