pairwise to work on rows instead of columns.

[musm]

Say I have a large matrix X in memory
I would like to use pairwise operating on the rows instead of the columns, without having to pass the transpose of X (so to not allocate another object in memory).

[rawls238]

I think this is a good idea, I'll put up an implementation of this unless @KristofferC has any objections about adding this

[nalimilan]

transpose now returns a lazy wrapper, so this may no longer be an issue.

OTOH it could be clearer for users to be able to specify the dimension over which variables are stored, since we differ from the convention used by other languages (see also JuliaStats/Clustering.jl#79). Another reason is that cov supports such an argument (and expects variables to be stored by column by default...).

[nalimilan]

I've filed #121 taking advantage of transpose to support a dims argument. Please comment there.