change to residue.m

[Ben Abbott]

Doug Stewart and I invested some time and effort improving the  
accuracy of the residue.m script when pole multiplicity exists.

Doug's suggestion, which we've each invested several hours testing,  
was to set the value of each unique pole to that average of the  
multiple poles associated with it. The result has been found to be a  
few orders of magnitude more accurate.

To capture the improvement, the test scripts were also modified to be  
more stringent.

In the event, for those interested in verifying our work,  I've  
attached the patch and ChangeLog.

Ben

residue.patch
Description: Binary data

ChangeLog
Description: Binary data