To Do list:

-Make an example for MuNCut.
-For the bml make sure it works for clsutering variables only first. Since, if we let R=1 it should simplify to that.
-bml gives a weird error message sometimes.
-I think you have a problem with the intercept when doing glmnet in MuNCut.
-Extend this example to bml.
-I am still getting the error message for the muncut function 'Error in if (J2 > J) { : missing value where TRUE/FALSE needed'
 This error tends to happen when p=100,200; that is when the dimension is small.I think it is also a problem of the gaussian kernel getting really small values. 
-I get erros message when I use dist='correlation' in muncut.
-Make sure ncut with CME works for K>2.
-Make sure spawn2 works for K>2.
-Currently you have spawn and spawn2. Join them into one function with option between SA and CEM.
-OOP for the plot() and summary().
-Get ride and coalesce the C++ functions. They are too many. 