Posted by fabriziogiudici
on November 24, 2009 at 9:13 AM PST
The scoop of these days is the “climagate” - a hacker stole and then published on the web a bunch of e-mail messages and documents from a British university and those documents allegedly affirm, among other things, that some data sets used to proof the Anthropogenic Global Warming (AGW) have been “massaged” way beyond the legit terms. Depending on the country you live in, you'll find a widely variable coverage in the media: for instance, many newspapers in US and UK quoted the news while, mysteriously, in Italy almost no one did (with a few exceptions ) - but there are ongoing discussions on specialized blogs .
This blog is not competent on climatology, but there is a detail related to the software world that surprised and intrigued me. Part of the discussions focuses on a piece of software that was used to post-process some empirical data to produce one of the commonly used AGW-related datasets; in the source code of the software a few comments allegedly reveal some “massaging” bad practices.
Being those comments true or false, properly interpreted or not, the basic point is that - as far as I understand - this piece of software is available to public analysis for the first time today, and because it has been stolen. This is appalling: some of the basic principles of science are peer-review and the fact that everybody should be able to reproduce an experiment or a process used to get to some assertions. Scientific data should be open as well as the software used to process them - as everyone with a bit of math experience knows, the wrong processing technique (in bona or mala fide) can easily transform data into garbage. The need for transparency is even more evident in a debate that is taking the world by storm and is conditioning huge parts of the world development and environmental policies.
So, why that software hasn't been open-sourced from the beginning (since when the data it produced have been used in published papers)? Universities and public research centers are some common places where the open source culture is spread - it has been partly invented there. After all, even we poor software engineers are accustomed in making the source code for a benchmark publicly available...