Michael Baudin

I created a thread to discuss this: https://openturns.discourse.group/t/normalization-of-the-input-sample-in-krigingalgorithm-in-ot1-16/101 Actually, the `IsoProbabilisitcTransformation` cannot be the way to go in general in my opinion, because this may involve a complex nonlinear transformation...

What if the g function is pre-processed, so that a new function, operating on a standard space, is created before using the algorithm? Suppose that t is the transformation which...

It seems to me that the current version of the documentation is unclear enough. The following is from the [theory doc](https://openturns.github.io/openturns/master/theory/reliability_sensitivity/ranking_src.html):  The sentence should make clear that this is...

> Why does your command line mention `KrigingAlgorithm.rst` and `GeneralLinearModelAlgorithm.rst`? What are these files? The corresponding help pages are in `KrigingAlgorithm_doc.i.in` and `GeneralLinearModelAlgorithm_doc.i.in`. But these files are unknown to Sphinx....

@josephmure , @m-balesdent : Please review this PR and tell me if that seems OK from your point of view.

@jschueller : I plan to proceed as follows: - remove the unwanted change in `BuildDistribution` and return a `Distribution`. - Implement the html pretty print of `ComposedDistribution` in C++ by...

> seems good Edit 30/10: - Created a pretty-print of `ComposedDistribution` at the new dedicated PR #2447. - Removed commit 74586396af5684c438765d1523f12bc52abb76be which contains the faulty `ComposedDistribution` changes.

@jschueller : I think this it ready.

I am sorry that the message wasn't clear enough. In the previous math text, the indices are from 1 o n. In Python/C++, the indices go from `0` to `n...

See e.g. $\{\boldsymbol{x}_1, ..., \boldsymbol{x}_{i-1}, \boldsymbol{x}_{i+1}, ..., \boldsymbol{x}_n\}$, which fails to render.