Mikhail Kagalenko

@devmotion The function ```demean_col()``` should be considered broken and replaced because of [this](https://github.com/JuliaLang/Statistics.jl/issues/22) .

> Thanks, but you don't seem to have answered all my questions/comments. I have not resolved [this](https://github.com/JuliaStats/StatsBase.jl/pull/742#discussion_r762436508) and [this](https://github.com/JuliaStats/StatsBase.jl/pull/742#discussion_r762437629) conversations, because I am not sure how to proceed. Those are...

> Thanks, but you don't seem to have answered all my questions/comments. In the latest revision I have removed all parts that caused questions. Now this PR is streamlined -...

At present, I must define the model to pass to curve fit as: ``` function model(xdata, alpha) ... end ``` and then call curve_fit as ``` curve_fit(model, xdata, ydata, p0)...

I do not understand this objection at all, the simplified form requires exactly zero additional knowledge about optimization. Implementation of the the old interface in terms of the new one...

> * other packages use the current form "Other packages", meaning MATLAB, use the verbose form because that language does not have closures. There's no reason to keep legacy crud...

Да, правильно. Не очень удивительно, учитывая что у них больше 500 открытых issues https://github.com/jgm/pandoc/issues/5964

Rebased on the latest master, any chance of having this merged?

Can every subtype of AbstractMatrix be updated in place?

So GenericSVD does not implement not in-place ```svd()```. To enable it, probably it would be necessary to rewrite it into something like ``` function bidiagonalize_tall(A_in::Matrix{T}, A_out::Matrix{T}, B::Bidiagonal) ... end bidiagonalize_tall!(A::Matrix{T},B::Bidiagonal)...