I've been working at CERN for 15 years as an experimental physicist. I'm currently on the Atlas experiment. The prevailing operating system now is SLC (Scientific Linux Cern), having gone through various flavours of Unix and even OS9 (anyone remember that?). It is stable, which necessarily means 'outdated' with respect to Ubuntu etc. Various installation configurations are available (google is your friend...).
For Data Acquisition, at the sharp end we have single board computers running homegrown software. The 'slow control' (measuring temperatures etc on a timescale of seconds rather than nanoseconds) we are using PVSS, but often in the lab (not on the main detectors) we run with LabView, using VME acquisition boards or PXI. Increasingly there is a move to USB for benchtop systems, but you still find plenty of IEEE (GPIB) instruments. For open source, maybe check out the python/C++ DAQ framework written by Carlos Lacasta.
For analysis and data presentation most people use ROOT (open source), which includes a C++ interpreter for scripting. This has recently been ported to python in the form of PyRoot. In my (minority) view, Root is a dog of a software package. It is largely monolithic in its design, and the intrinsic data format is not well documented, so once you adopt Root you are really stuck with it. It has permeated all our software, and we suffer for it. The programming model mixes content and presentation. Many people love it. Igor (proprietary) is in many ways superior for small (5Gb) datasets. The presentation is superb. There are other packages (e.g. Hippodraw, Gnuplot) for presentation. I understand that Mathematica and MatLab are popular for magnetic field calculations, but much of our software is homegrown using the (open source) CERNLib.
Slashdot Mirror
I've been working at CERN for 15 years as an experimental physicist. I'm currently on the Atlas experiment. The prevailing operating system now is SLC (Scientific Linux Cern), having gone through various flavours of Unix and even OS9 (anyone remember that?). It is stable, which necessarily means 'outdated' with respect to Ubuntu etc. Various installation configurations are available (google is your friend...). For Data Acquisition, at the sharp end we have single board computers running homegrown software. The 'slow control' (measuring temperatures etc on a timescale of seconds rather than nanoseconds) we are using PVSS, but often in the lab (not on the main detectors) we run with LabView, using VME acquisition boards or PXI. Increasingly there is a move to USB for benchtop systems, but you still find plenty of IEEE (GPIB) instruments. For open source, maybe check out the python/C++ DAQ framework written by Carlos Lacasta. For analysis and data presentation most people use ROOT (open source), which includes a C++ interpreter for scripting. This has recently been ported to python in the form of PyRoot. In my (minority) view, Root is a dog of a software package. It is largely monolithic in its design, and the intrinsic data format is not well documented, so once you adopt Root you are really stuck with it. It has permeated all our software, and we suffer for it. The programming model mixes content and presentation. Many people love it. Igor (proprietary) is in many ways superior for small (5Gb) datasets. The presentation is superb. There are other packages (e.g. Hippodraw, Gnuplot) for presentation. I understand that Mathematica and MatLab are popular for magnetic field calculations, but much of our software is homegrown using the (open source) CERNLib.