Minor feedback on covariance matrixes in a calibration context

I have noticed that feeding a calibration algorithm a prior covariance or an observation covariance matrix in the following fashion:


will lead to a type error. OpenTURNS demands that the covariance matrixes should be a CovarianceMatrix object. I had to rewrite my code in these lines :

ot.CovarianceMatrix( ot.CovarianceMatrix(22)*2)

It is not a very big deal but it can be a bit inconvenient in my opinion.


the problem is the multiply by zero case as it has to be SPD