[SciPy-dev] numpy.ma.mstats in scipy ?

[Pierre GM]

All,

I've been working on the adaptation of scipy.stats.stats to support 
MaskedArrays. This is mostly an extension of numpy.ma.mstats, with most of 
the functions of scipy.stats.stats and a couple of corrections (ties 
correction in friedmanchisquare, tweaking ks_2samp to match results with R).

I was considering adding the package to scipy: where should I put it ? As an 
additional module in scipy.stats (it would then be called 
scipy.stats.mstats) ? As its own package (scipy.mstats) ? Once the module is 
in scipy, we could then get rid of numpy.ma.mstats and numpy.ma.morestats, 
that don't look like they fit in numpy.

Thanks a lot in advance.
P.