Monday, November 01, 2004

Science and hardware

I would be the first person to admit that I have an affinity for Linux. While I am not the "Linux bigot" that I used to be, I still think that for most scientific use, it is by far the most effective OS. Linux machines are fast, the Unix-style OS efficient, familiar and convenient, and the graphical capabilities, as issue in years past, are excellent. At least where I am concerned, I still can't do any serious modeling on Windows. The latter is great for certan applications, including scientific applications, but a 5 ns MD simulation is something best left to a Linux box or better still a cluster. Proprietary UNIX OS' have niche applications, but have probably seen their best days. I have fond memories of AIX and IRIX, two OS' that have served me well over the years, but today, I would pick Linux over those 9 times out of 10.

Where does leave the Mac, especially OSX. For many scientists, OSX and Apple can do no wrong. It is definitely an intriguing OS, with the front-end classic apple and behind the scenes, a UNIX box. I suspect that in academia and in some bioinformatics circle the Mac OS will always be the OS of choice, but I see some XServers being deployed in a number of places over the next few years. Scientists are a funny bunch. When they like something, they tend to hold on to it like their lives depended on that act, even when the object of their affection has long past its zenith. Then there are the IT groups that dictate hardware choices, pissing off many a research scientist in the process. With all the choices available, and a new generation of scientists, the hardware/OS wars are only just beginning.