It may be cheating, but sometimes I have to go no further than the US/LHC front page to find something to blog about. This morning, this “headline” from Scientific American blogger John Matson caught my eye:
Reading the entry and the original Princetonian article, it turns out that Princeton undergraduate Xiaohang Quan found a bug in the CMS software for reconstructing the particles that come out of a proton-proton collision inside the detector. (Reconstruction is the process of interpreting the singals from the detector as particles. For example, an electron is identified when a charged track matches up with energy deposited in the electromagnetic calorimeter, a photon is identified when energy is deposited but no track matches, and so on. This picture helps.) My first reaction was that this isn’t news at all: if CMS software is anything like the software on ATLAS, then bugs are found in it all the time!
The software for our experiments is extremely complicated, trying to do myriad things at once, and it has (at a conservative guess) several hundred people working actively on it at the same time. All of them are adding features — for example, new and better ways to look for particles — and changing things around, which introduces a constant stream of incompatibilities and bugs. So we constantly have major glitches; then we (constantly) find them, and (constantly) create more. Very few of us are really “pros”; we’re physicists, not computer programmers, and it shows when you look at our code and how it’s organized.
And many of our collaborators are undergraduates, and they aren’t just working with us in order to “gain experience.” They are, and are expected to be, serious contributors to the overall experiment. They generally don’t take on extremely long-term or full-time projects, but they test equipment, investigate ways to explore new physics, and give talks in meetings. And yes, they introduce bugs in the software, and find them too — just as do graduate students, postdocs, and professors.
In short, Quan’s achievements don’t somehow imply that the rest of CMS was asleep at the wheel — they show that we all work together, and that every collaborator counts.
