[Developers] solve function test for singularity

[Mollie Brooks]

Hi Developers,
You may remember me complaining about my pond model in the past. With some help, I finally got it working, but it only works with the old version of the solve function which Dave still uses on his own computer. When I try to run the exact same code with the SVN version of the solve function, it works for a while and then goes to a bad region of parameter space and I get an error about the matrix being singular. I think the new version is defined in ludcmp_solve.cpp The code leading up to 
     cerr <<
	 "Error in matrix inverse -- matrix singular in solve(dvar_matrix)\n";
looks logically different to me, but I'm not an expert. 

Before I go messing with someone else's code...Is it legally possible to go back to using Dave's version or would the person who did the conversion be willing to reexamine the test for singularity? Dave would rather not look at the new version.
thanks,
Mollie


Mollie Brooks
Ph.D. Candidate
NSF IGERT Fellow
Biology Department
University of Florida
mbrooks at ufl.edu
www.zoology.ufl.edu/mbrooks



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.admb-project.org/pipermail/developers/attachments/20110823/2467618d/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: dave's solve.cpp
Type: application/octet-stream
Size: 8165 bytes
Desc: not available
URL: <http://lists.admb-project.org/pipermail/developers/attachments/20110823/2467618d/attachment.obj>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.admb-project.org/pipermail/developers/attachments/20110823/2467618d/attachment-0001.html>