Many programming tools are available to the applied econometrician. There are high-level matrix languages mostly dedicated to econometrics (like GAUSS and Ox), to statistics (like Splus, R, or Stata) or to scientific computing broadly speaking (like Matlab, Octave, or Scilab). All these languages provide sufficient routines (in particular, related to matrix calculus and numerical optimization) to implement virtually any algorithm required by an econometric analysis. Why should we then suggest Python as yet another possibility when it is not even designed for scientific computing? Let us make a parallel with what happened in applied statistics: Lisp-Stat, an extension of the general-purpose language Lisp, was gradually given up at the end of the 1990s in favor of the free and open-source R. In 2005 Jan de Leeuw, a prominent actor in the R community, wrote: