See also Chapter 3.2 of Smashing Physics.
I haven’t been paying as much attention as I should to science policy in the election. Guilty, will re-engage. I have a good excuse, of course. Even after STFC cuts we’re still members of CERN, the LHC is running, and (volcanoes permitting) I am traveling, working and basically doing physics. And it’s exciting and absorbing.
Typical scientist. We just get the data and get happy. In STFC related politicking, I’ve heard our big mistake was that in economic good times, when science was getting roughly its share of the growth (and particle physics was getting somewhat less, but not being cut), we just shut up and got on with science.
It’s fair to say we should have said thanks (Thanks!), but this argument was that we should have been aggressively pushing for more & more cash, just so there was more fat to cut when bad times came. This shows a gulf in understanding between the worlds of science and politics, sadly.
Research is about finding good questions to ask, then answering them. Good questions which can actually be addressed don’t come along often, and the capacity to identify and answer them is grown slowly – you can’t just go out and double the size of a good research group overnight. There are some big, expensive things we could do if we had the money, but it would not take us long to run out of the skilled people to deliver them. This is something politicians often fail to grasp.
Scientists, on the other hand, mostly want to do science, and many of us only get noisy and politicised when our capacity to do science is threatened. As well as being a huge tactical mistake, this is an insult to the interested and intelligent lay public who foot the bill and deserve to hear what we are up to. It is also a recipe for political and social disaster – public life needs science, and scientists who engage consistently. (See for example the application of evidence-based thought to libel and quacks.)
Oh, the temptation of real data though! We drill on simulations for so long, and it helps, but it can’t fully prepare you.
Part of my PhD was writing an emulation (in FORTRAN 77) of code running on a network of transputers. This was part of the “trigger” system built to select data from electron-proton collisions in ZEUS. I had two years checking that code to death. Transputers were stupid little processors on a real-time network – massively parallel computing, very cutting edge for 1990 and very hard to debug. The idea was to feed simulated data into both the real transputer network and my FORTRAN, and check we got identical results. It worked great.
Then we got our first real data. And the transputers and the FORTRAN produced nonsense. It was not a good moment. All that careful preparation, and we couldn’t make head or tail of the real data. The only glimmer of hope was that the FORTRAN and the transputers were producing identical nonsense, so I could use the FORTRAN to investigate. A few stressful hours later, I had it. The detector had hundreds of little cells, each with 8 wires (numbered 1-8), and each wire received an electrical pulse if a charged particle passed nearby. In the simulated data, the pulse came out in order of wire number. In the real data they (facepalm!) came out in order of arrival time, which depended on where the particle was! Once we took this into account, all the crazy numbers lined up again. It was a minor epiphany – order out of chaos by thinking about it!
A good example of why simulation is good preparation, and why it is never enough. I ranted this story to my PhD student (hello!) over a beer in the CERN canteen last week. At the LHC we already had some of these little surprises, we need to watch for more. And for the big surprises too.
And perhaps in the election as well. Mmmm. This also is not a drill. I wonder if we can join the dots?