[femlisp-user] Femlisp-1.0.0

[Nicolas Neuss]

Hello,

a new Femlisp version is available.  Below is the relevant excerpt from
the CHANGES file.

Yours,
Nicolas.

--8<---------------cut here---------------start------------->8---
* From Femlisp-0.9.10 to Femlisp-1.0.0

Up to now, Femlisp has been almost completely autarchic.  This kind of
development is not adequate in a time when there are a lot of good
Common Lisp libraries easily reachable with the help of Quicklisp.
Therefore, Femlisp now includes dependencies on some libraries like
@library{bordeaux-threads}, @library{lparallel}, and
@library{cl-ppcre}.  This trend will continue in the future.

The Femlisp main repository switched from CVS to git.

Some changes by Fare (Francois-Rene Rideau): Improved file
@path{femlisp.asd}.  *read-default-float-format* is set dynamically to
double-float during Femlisp compilation.  People using Femlisp in
earnest should probably set it in their own init files to that value
as well.  The logical host #p"femlisp:" was removed.

Anisotropic refinement included.

There is blending support now in the sense of "transfinite interpolation"
(which means interpolating a cell map from arbitraty given maps of its
boundary).

There is now a Femlisp interface to Hang Si's Tetgen.  Domains can be
defined via their 2D boundaries.  Then 3D patches having such boundaries
are decomposed into tetrahedra.

Mesh graphic output is not redundant any more (before, vertices appeared
several times similar to plotting of discontinuous functions plots)

mesh/easymesh.lisp was removed - we do not support that format anymore

VTK graphics.  These are even interactive when you call Mayavi2 using the
command line "mayavi2 -x mayavi_poll.py" being in the #p"femlisp;images"
directory.

As a first step for the incorporation of complex-valued solutions, BLAS
routines may now depend on different matrix classes.
--8<---------------cut here---------------end--------------->8---


-- 
Nicolas Neuss    Email: address@hidden  Tel: 0049-9131-85-67208