README

SGLib is toolbox developed for Matlab/Octave to implement and investigate
into stochastic Galerkin methods with a special focus on tensor product 
methods. 

INSTALLATION
============

There really is no installation. Just put the files into some directory of 
your own and start Matlab or Octave from *this* directory. The last point 
is necessary so that the initialization files startup.m (for Matlab) or 
.octaverc for Octave are called. You should see some lines saying something 
like

  SGLIB v0.9
  Type SGLIB_HELP to get help.
  Type SGLIB_SETTINGS for changing the settings.

The best thing to do first is to run the unit tests first. For this simply 
type 'testsuite' one the command line and after a while and some screens 
scrolled by it should say:

  Module "sglib": 0 of ??? assertions failed.


COMPATIBILITY
=============

The code is tested with Matlab version 7.2 to 7.8 and Octave since version 
3.0.0. It won't work with Octave version 2.x and probably not with Matlab 
6.x versions. While latter can probably be accomplished without too much 
work, the former is quite difficult since there is just too much missing
functionality in Octave. Especially support for sparse matrices was only 
added to the mainline Octave version since 3.x and still is not "complete". 
Further quite a few functions that use debugging/introspection features 
from Matlab have only limited functionality under Ocatve (i.e. directly 
jumping into the debugger to the correct line if a unit test failed). 
Further speed is for many example about 1/10 of that of the Matlab version.

CONTENTS
========

SGlib contains methods in the following areas 

* Polynomial chaose (PCE) methods
* Generalized polynomial chaos (GPC)
* Statistics functions
* Unit testing
* Tensor methods
* Parameter studies
* Option management
* Lots of utilities