Richard, While I am not in scientific computing anymore, while working at the CFA (http://cfa-www.harvard.edu/) I can say our shop was probably 75%+ FOSS (Note: I'm counting solaris in this, even tho it's debatable) After I left, from what I understand it's gone even more (eg, solaris replaced w/ fedora, or other linux distros) although, just about any laptop which is purchased is a Mac. In every research institution I have ever worked at/with, FOSS has played a major role. However, most researchers I know will use the best tool for the job. If it happens to be Matlab, it's Matlab, if it happens to be Perl, it's Perl. They don't care, as long as it get's the job done, and doesn't break the bank on their research grants.
However, there is a difference in mindset I've found between the "hard" and "soft" sciences that tends to lead one towards FOSS and one towards proprietary applications, but YMMV. Good luck w/ Matlab! I'm not sorry to know there's a 99% chance I'll never have to use it again! Mike --- On Thu, 10/9/08, Richard Lawrence <[EMAIL PROTECTED]> wrote: > From: Richard Lawrence <[EMAIL PROTECTED]> > Subject: Anyone using FOSS for scientific computing? > To: [email protected] > Date: Thursday, October 9, 2008, 12:49 PM > Hi everyone, > > I have a topic to throw out for discussion, if anyone is > interested: > how many of you are using FOSS for academic or scientific > computing? > Have you been successful with it? Do you find yourselves > fighting a > tide of money, bureaucracy, etc., or is your environment > more receptive > to FOSS than perhaps the general public is? > > I'm thinking of Numpy/Scipy here, Octave, R, etc. as > alternatives to > things like Matlab and SPSS. > > I ask because I recently moved to Philadelphia to start a > job as a > programmer and research coordinator in a psychology lab. I > accepted the > job largely because I thought I would be working in Python, > writing FOSS > data analysis programs that would be used both in this lab > and > distributed freely. As it turns out, my lab is extremely > Matlab focused > (which means that any code I write can't be run by > anyone who doesn't > pay the $$$$ for the proprietary Matlab license, unless I > spend time > making it compatible with Octave), and the principal > investigator > consistently chides me for pushing for greater Python and > Octave use. > (Yesterday, he said, "Your first name is Richard...but > your last name's > not Stallman, right?") He also has what I consider to > be slightly > suspect ideas about what it means to be a steward of > taxpayer dollars. > > All of this is a little odd to me, because according to lab > lore, at one > time you weren't allowed to graduate if you had never > compiled a kernel. > Now, we're all on OS X. > > Have others encountered this scenario? Did you have any > success pushing > back? Do you have a persuasive counter-argument to the > "pragmatism, not > idealism" sentiment that has apparently won out here? > > Richard
