I am working on some changes to cm to make it work properly with meshes that have collapsed elements, but there is a problem with my code which I need some advice on.


My test mesh has two collapsed elements adjacent to each other:


.. image:: snapshot1-mesh.png


When updating derivatives using ``fem update node derivatitive n`` the default behaviour is to use the average of all versions of a node and when updated with this command the mesh looks like this:


.. image:: snapshot5-averageversions.png


For some meshes this updating distorts the initial mesh such that fitting becomes impossible so I have added a command ``fem update node derivative n versions individual`` which has the following, desired, effect:


.. image:: snapshot2-sideview.png


Also I am using a modified version of fe02/DLSE.f which sets some scale factors to zero for collapsed elements to ensure that the correct interpolation formula is used.


The problem with the command that I have added is that the scale factors for one of the lines on the collapsed elements are not updated, as shown in these images:


.. image:: snapshot3-scalefactors1.png

.. image:: snapshot4-scalefactors2.png


The scale factors for one of the versions of the collapsed line have not been updated.


Where should I be looking to find out what might be causing this?


From karl Mon Feb 5 14:12:46 +1300 2007
From: karl
Date: Mon, 05 Feb 2007 14:12:46 +1300
Subject: 
Message-ID: <20070205141246+1300@www.cmiss.org>

This looks like the same problem as mentioned in comments #4 and #7 in bug `#240
<http://www.cmiss.org/cm/tracker/240/pcng_issue_view>`_?


From glennr Mon Feb 5 14:58:46 +1300 2007
From: glennr
Date: Mon, 05 Feb 2007 14:58:46 +1300
Subject: 
Message-ID: <20070205145846+1300@www.cmiss.org>

Thanks, I now understand what you were talking about.