This weekend Trento hosted the Italian Free Software Conference
. I could attend only friday because then I went to Milano for sci.bzaar.net (report in next post).
There was a very interesting presentation by Emanuele Somma about Bank of Italy and their internal use and infiltration of Free Software.
He cited the paper “Reproducible Econometric Research (A Critical Review of the State of the Art)” in which the authors, Roger Koenker, Achim Zeileis, cite Buckheit and Donoho (2005) citing what de Leeuw (2001) has called Claerbout’s Principle:
An article about computational science in a scientific publication is not the scholarship itself, it is merely advertising of the scholarship. The actual scholarship is the complete software development environment and the complete set of instructions which generated the figures.
Koenker and Zeileis go on reporting about scholars in economics have somehow turned into programmers:
The transition of econometrics from a handicraft industry (Wilson, 1973, Goldberger, 2004) to the modern sweatshop of globally interconnected computers has been a boon to productivity and innovation, but sometimes seems to be a curse. Who among us expected to be in the “software developmentâ€ business? And yet many of us find ourselves precisely in this position, and those who are not, probably should be. As we will argue below, software development is no longer something that should be left to specialized commercial developers, but instead should be an integral part of the artisanal research process. Effective communication of research depends crucially on documentation and distribution of related software and data.
So their contribution:
Our main contention is that recent software developments, notably in the open-source community, make it much easier to achieve and distribute reproducible