jump to navigation

Cosmic-ray studies of the CMS tracker January 28, 2009

Posted by dorigo in news, personal, physics.
Tags: , ,

It is always nice to open the web browser in the morning, check the arxiv for new interesting preprints, and be surprised to find one’s own name in the author list. That is what happened to me today, while browsing lazily the list of new hep-ex papers, as my eyes set on “Performance studies of the CMS strip tracker before installation”.

The paper describes the full testing of a sector of the CMS tracker. The tracker (see picture on the right, showing a detail of its inner barrels) is a daring device made of many concentric barrels of silicon strip sensors. During the summer of 2007 a quarter of the device was fully instrumented, cooled, and read out while it was being exposed to cosmic rays, and a total of 4.5 million tracks were reconstructed, allowing to gain critical experience with its operation, and detailed studies of its tracking capabilities, the tuning of a simulation of the detector, and the development of advanced tools.

Of course I knew the paper was being prepared -the submitter is Patrizia Azzi, a member of my group in Padova (although she’s full-time at CERN)- but no, I did not contribute to it in any significant way and no, I had not even read the draft!

To be fair, the author list includes over 400 names, the members of the CMS Tracker Collaboration (people who were somehow involved in the construction of the tracker), so you should not run out screaming “Dorigo is a parasite!” -at least, I am not the only one! This is how things work in large collaborations: you focus on one or two studies at a time, on the time-scale of two-three years, but you do not just sign your papers: you sign all of them.

In retrospect, I should be even less severe with myself.  Although the paper contains no results of mine, I did work on the analysis of the data. I did a study of multi-track events, trying to figure out how the presence of large amounts of hits close together could affect the tracking (a matter of relevance for LHC, where dozens of tracks will pack together within small volumes), and I studied the extraction of the angle of incidence of tracks from the width of clusters of charge in the silicon strips (tracks crossing a layer of silicon at normal incidence leave a ionization trail which gets collected in few strips, while tracks crossing with a large angle leave a signal in many adjoining strips).

Those studies did not end up providing a valuable addition to the paper -mostly because I did not conclude them- and they were left out of it, but I invested at least one month of work in them. Not much, but I do not feel a parasite after all: the paper is maybe the result of 20 or 30 man-years of studies, so each of the 400 authors contributed an average of less than one month of full-time work!


1. Xylem - January 28, 2009

T- actually, the model i think you’re describing of ‘opt-out’ author list in a HEP expt is *not* universal — it is the one we used on Babar, and i think is pretty common, but e.g. i know on Belle they used ‘opt-in’ where you had to say for each paper that you explicitly wanted to be on it. I prefer that model, as one is forced to minimally read the title, at least.


2. dorigo - January 28, 2009

Hi M.,
sure, I know… Not everywhere things work the way I described them. I think an opt-in method also has its problems, because it favors people with low moral standards: those who shamelessly sign every paper regardless of their contribution get the additional advantage that those papers have fewer signatures!

3. jeffwyss - January 28, 2009

First my HEP experience, then my 5 cents worth.

Many years ago I quickly converged to formally using, for career advancement, only a very VERY small subset of the hundreds of papers I signed when on SLD and CDF. At the end, when it really counted, I used only about a dozen, those where I felt that my contribution was non-zero (either hardware or analysis). In particular I never used the many papers that kept coming out with my name after I quit actively doing HEP (about 10 years ago).

The real trouble with HEP these days (last 20 years) and other very complicated forms of experimental physics is that it is not at all clear what to do. Indeed there are people that are immediately thrust, for various reasons, into the spot-lights and steal the show (some humbly, others with little or no humility) by doing analysis, while there are many others that work very VERY hard for years behind the scenes, almost in the dark, to build the apparatus and the software packages working properly. If one DOES want to keep track of all those involved in producing a quality paper then the list of signature does get very long. I do think it should include those that do hardware (keeping the detectors up and running), triggers, montecarlos, data taking and analysis, to the managers and group leaders that created the conditions for the various types of people to work in synergy. Sometimes I can concieve that once a detector system has been working for many years in a relatively stable configuration then the recongition of past hard working path-breakers, and even standard service workers, might be considerably lowered. But then I lapse into the feeling that it is impossible to decide, without injustices, when historical debts can be considered extinguished or what is to be considered standard non-mentionable slavery.

I conclusion I truely and vigorously feel that the active opt-IN method (where one has to explictely ask each time that his name appear on the author list) should be the standard method in science. The ethical aspects would be continuously refreshed as the decision to step forward would have to be taken every time. Of course there are people that are unethical and would like to sign everything, but with a healthy opt-IN method in place these people must be VERY cautious else they risk ridicule. It takes a long time to rebuild a reputation once you lose face.


4. Andrea Giammanco - January 29, 2009

I agree with this analysis, and I’m in favour of opt-in too.
I’ve read a public document by a working group entirely dedicated to the issue of how to sign papers in HEP, and I learnt that the introduction of opt-in in the Belle collaboration halved the number of average signatures per paper. This means still too many (~300 instead of ~600 signatures per paper) but it could be possible that ~290 per paper are signatures by people who actually built the detector or did some essential “service work” since the beginning, and so could be fairly considered authors of whatever comes out from the high-level analyses. In any case, I consider it a progress.

5. dorigo - January 29, 2009

Hi Andrea,

I find that unrealistic. The analysis performed by those 10 authors rests on the work of all the people around them. Not just detector builders, but software developers, shift takers, group organizers, etcetera, etcetera. The idea is that an experiment works because everybody does their share of work.

Any collaborator is expected to provide some time of service work before they can sign papers. Once they’ve done it, they should sign all papers, right ? Then we get back to the 600 number easily, and have only to discard the new entries, fresh PhD students who are those who need to sign papers the most.

I think the opt-in method works well to reduce the number of signatures because it excludes those senior members who do not need more papers. However, if anybody who is young and in search of a position wants to “play it fair”, they will decide to not sign some fraction of the papers, and they are thus going to be at a disadvantage with respect to those who do not play a fair game and sign everything. I think this is bad enough to make the system a bad choice.

Look at it this way: what do YOU care about the fact that a CMS paper will have 2500 authors ? Does it really bother you, or do you feel your career is damaged by this mechanism ? I really doubt it does. The search committees know how the system works, and are capable of doing the math. They will, in fact, look at the resume of candidates to find out what papers they really did contribute to. The list of authorship will count for nothing, and in fact a long time ago I started doing what Jeff mentions: when I apply for a position I include only 20 publications.


6. dorigo - January 29, 2009


besides what I answered to Andrea above, I think the issue of “losing face” you mention at the end of your comment is a non-issue. People find all sorts of justifications for their behavior, and in large groups they cannot easily be singled out. Imagine in CMS, with 2500 potential authors, papers coming out with 1500 signatures, 200 of these by people who sign every single paper. We can debate the likelihood of those figures, but they are plausible numbers. Now, are those 200 “immoral” people going to lose their reputation ? I seriously doubt it. If they were 10 or 20, maybe. But they easily hide in the additional order of magnitude.


7. Andrea Giammanco - January 29, 2009

Well, exactly the fact that expHEP people are forced to rely *entirely* on alternative ways to judge a CV should be counted as one of the big problems of the system.
Theoreticians have several ways to take into account objective figures when judging a candidate (number of publications, number of citations, h-index, etc.). None of those are perfect indicators, but just reasonable numbers which everybody agrees being at least correlated with the very poorly defined quantity that you would like to “measure” (i.e., how “good” is the candidate). Indeed, people who are part of selection boards and who are theoreticians tell me that these objective indicators are not relied 100% upon, nevertheless they are useful at least in the early stages of selection (when you get ~200 applications and want to reduce this number to something more manageable, that you will then carefully consider case per case).
I think that this is a fair balance between objectivity and subjectivity.
But in experimental HEP, the fact that everything is left to human judgement scares me.
I’m not scared by the weight of “reputation” on one’s own career (although, this can have its risks), rather by the fact that whenever there are no objective indicators to rely upon, biases from unnoble motivations become more prominent (like personal relationships, or struggles for power both at the inter- and intra-group levels, etc.)

8. CMS and extensive air showers: ideas for an experiment « A Quantum Diaries Survivor - February 6, 2009

[…] CDF, cosmic rays, DELPHI, LHC, Pamela trackback The paper by Thomas Gehrmann and collaborators I cited a few days ago has inspired me to have a closer look at the problem of understanding the features of extensive air […]

Sorry comments are closed for this entry

%d bloggers like this: