DOC update Ubuntu installation instructions

* python-numpy-dev was listed twice, but no longer exists
* OpenBLAS may give a speedup (but hedge the recommendation
  because it's not particularly stable)
* the metapackage build-essential will install g++ and some basic
  development tools and libraries

doc/install.rst

@@ -24,32 +24,55 @@ There are different ways to get scikit-learn installed:
     If you wish to contribute to the project, it's recommended you
     :ref:`install the latest development version<install_bleeding_edge>`.
 
+
 .. _install_official_release:
 
 Installing an official release
 ==============================
 
 
-Installing from source
-----------------------
+Getting the dependencies
+------------------------
 
-Installing from source requires you to have installed python (>= 2.6), numpy
-(>= 1.3), scipy (>= 0.7), setuptools, python development headers and a working
-C++ compiler. Under Debian-based systems you can get all this by executing with
-root privileges::
+Installing from source requires you to have installed Python (>= 2.6),
+NumPy (>= 1.3), SciPy (>= 0.7), setuptools, Python development headers
+and a working C++ compiler.
+Under Debian-based operating systems, which include Ubuntu,
+you can install all these requirements by issuing::
 
-    sudo apt-get install python-dev python-numpy python-numpy-dev python-setuptools python-numpy-dev python-scipy libatlas-dev g++
+    sudo apt-get install build-essential python-dev python-numpy python-setuptools python-scipy libatlas-dev
 
 .. note::
 
-    In Order to build the documentation and run the example code contains in
+    In order to build the documentation and run the example code contains in
     this documentation you will need matplotlib::
 
         sudo apt-get install python-matplotlib
 
 .. note::
 
-    On Ubuntu LTS (10.04) the package `libatlas-dev` is called `libatlas-headers`
+    On older versions of Ubuntu,
+    you might need to ``apt-get install python-numpy-dev``
+    to get the header files for NumPy.
+
+    On Ubuntu 10.04 LTS, the package `libatlas-dev` is called `libatlas-headers`.
+
+.. note::
+
+    The above installs the ATLAS implementation of BLAS
+    (the Basic Linear Algebra Subprograms library).
+    Ubuntu 11.10 and later, and recent (testing) versions of Debian,
+    offer an alternative implementation called OpenBLAS.
+    While this implementation has some issues
+    (please don't file bug reports about this),
+    it may offer a significant speedup to some modules of scikit-learn,
+    especially on multicore hardware.
+    Replacing ATLAS with OpenBLAS only requires two commands::
+
+        # NumPy may not run when both ATLAS and OpenBLAS are installed,
+        # so remove the former.
+        sudo apt-get remove libatlas3gf-base libatlas-dev
+        sudo apt-get install libopenblas-dev
 
 Easy install
 ~~~~~~~~~~~~