History for What is the difference between classes vs regions in CMISS?

changed:
-
Well, this is a fairly mirky area. People generally agree that you define a separate class for each type of problem you are solving. Eg: 3D finite elasticity is different to 2D membrane theory. So you need two classes. 

Now, a region is meant to be two distinct regions in space. I think that this definition has become mirky in the past because people have used these data structures in ways that help solve their problems. If someone who knows more about this stuff is reading this wiki, please feel free to add/change things I have said. Hopefully this will help people in the long run.

A good explanation of the difference between classes and regions can be found by 
searching for "BIDOMAIN ITERATION" in the 'cm Lookup Frontend'.  I found this to
provide a crystal clear distinction between the two.

From nash Wed Nov 2 23:05:57 +1300 2005
From: nash
Date: Wed, 02 Nov 2005 23:05:57 +1300
Subject: historical perspective
Message-ID: <20051102230557+1300@www.cmiss.org>

If I remember correctly, classes were originally intended to be used to set up multiple problem types (ie. different governing equations) for the same model - for example class 1 could be used to set up/solve governing equations for ventricular wall mechanics, and class 2 to solve electrical activity. 

On the other hand, regions were intended to be used for distinct geometries - for example, region 1 could be ventricular wall, and region 2 could the the ventricular cavity. 

Since inception, these concepts have certainly been hijacked for various purposes, eg. gotta luv that BIDOMAIN ITERATION model description - one of the more obtuse uses, but does the trick I'm sure ;-)