Outdated instructions in advanced installation instructions for Linux

[lesteve]

From http://scikit-learn.org/stable/developers/advanced_installation.html#linux

At this time scikit-learn does not provide official binary packages for Linux so you have to build from source if you want the latest version

This is no longer true with wheels. This is misleading and caused confusion recently in #2889 (comment).

[rth]

Also, the advanced installation instruction do not seem to mention conda at all, unlike http://scikit-learn.org/stable/install.html.

Using OpenBLAS can give speedups in some scikit-learn modules, but can freeze joblib/multiprocessing prior to OpenBLAS version 0.2.8-4, so using it is not recommended unless you know what you’re doing.

OpenBLAS 0.2.8-4 was released in August 2013, and oldest supported Ubuntu LTS (14.04LTS) seems to have 0.2.8-6. Maybe it could worth refactoring that part to use OpenBlas by default, only mentioning ATLAS for "legacy systems"? ATLAS vs OpenBlas only seemed to add to the confusion in the comment you mentioned.

The windows instruction also look outdated,
http://scikit-learn.org/stable/developers/advanced_installation.html#windows
particularly

if there are no binary packages matching your python, version you might to try to install scikit-learn and its dependencies from christoph gohlke unofficial windows installers or from a python distribution instead

this should say "use conda" IMO.

MacPorts for Mac OSX
sudo port install py26-scikit-learn

py36-scikit-learn ..

[narendravardi]

@jnothman @lesteve can i work on this?

[jnothman]

You would be welcome to, thanks

[cmarmo]

@lesteve, last fall the advanced installation page had been updated. Are you happy with the results? Should we close this one? Thanks!

[lesteve]

Yep, closing this one, thanks!