This is one of the things most confusing to folks outside of high energy particle physics. Our papers have 3000 authors. The ATLAS author list is about 17 pages long, depending on the formatting. Sometimes there are even fewer than 3000 words in the paper – surely we aren’t suggesting that different people cross t’s and dot i’s?
Well, what does it mean to be an “author” of a paper? In our case, it means that you made a contribution to the work described in the paper. Of course, if you built part of the detector, and that part of the detector is used in the analysis, then you should be a co-author on the paper! And if you were responsible for running that part of the detector during one of the critical times, or calibrating it so it would work, then you should be a co-author! And if you wrote some of the software (ATLAS software, at least) that was used during the analysis, then you should be a co-author on the paper! And if you derived some of the numbers that they used, even though you haven’t written a paper on your numbers yet, you should be a co-author on the paper! ….. The list grows pretty fast! In practice, it’d be almost impossible to try to select the correct several hundred people for an authors list. So instead, ATLAS keeps a running list of all the people who have contributed substantially to the experiment (roughly the list of people who have been members for more than one year) and uses that as our authors list. There are some stickier points to this – trying to draw a line between software we “develop” and software we “use,” for example.
Here I should add one category that is, perhaps, less obvious: graduate student advisors. As a graduate student, you don’t have a reputation in the field yet. Your advisor’s signature asserts that they vouch for the correctness and appeal of the work you’ve done. So I would say an advisor should always sign their students’ work – even if they have not been deeply involved in the paper.
This mess can lead to all kinds of confusing conversations. You keep a list of papers on which you were listed as an author and another list of papers to which you “significantly contributed.” And if you show those lists to anyone outside physics, they’ll say, “Wait, you didn’t contribute to these papers, but you’re an author????” I must admit that I have not read every word of all the papers that I’m a “co-author” of!
Because there are different practices in different fields (or even different experiments!), we can get into tough spots as well. Some fields list “major authors” first. Or they list graduate students first. Or they rotate the authors’ list alphabetically, so that one person is the “first author” on every paper. We don’t tend to do that in ATLAS, so it’s tough to tell who really wrote the paper.
It can be frustrating when you’re applying for a job – imagine having to explain which of the papers that you “co-authored” you actually wrote! – but I confess it’s kinda fun to have a list of publications as long as your arm…
One other note. Everyone has probably heard of the “six degrees of separation” game. You can play the same game with paper co-authors, and thanks to this notion of authorship the connections spread fast! In mathematics, it’s called the Erdos number after Paul Erdos. With physicists, there’s some debate over who should be the Erdos equivalent (my nominee is John Ellis). You can also play the game with academic genealogy. In fact, I can trace the Italian side of my academic genealogy further back (1860′s) than the Italian side of my family genealogy (1880′s), and they both go though Pisa! I leave you with the XKCD take: