By Kate Metropolis
Robert N. Cahn (LBNL) is the current chair of the BABAR
Publications Board. Known for the clarity of his prose as well as the
clarity of his calculations, Cahn is co-author, with Gerson Goldhaber, of
The Experimental Foundations of Particle Physics (Cambridge University
Press, 1989). He has been a member of the Theory Group of LBNL since 1979.
Cahn took time recently to answer some questions about publishing papers
courtesy of SLAC)
Q: What makes a good experimental physics paper?
A: Of course, a good experimental paper must have an important result.
It needs to explain the significance of that result for our understanding
of fundamental physical laws. It needs to explain how the experiment was
actually conducted. It should explain how the data were analyzed and
should identify the critical aspects of the experiment. What was special
about the data or the analysis? What was novel here?
A good paper provides enough information for the reader to assess the
experiment. Unfortunately, this makes it very hard to write a good
experimental paper for Physical Review Letters. In the few pages that are
allowed, it is nearly impossible to explain thoroughly analyses as
complicated as those that are typical in high energy physics these days.
The discussion of systematic errors plays an especially important role,
for it is here that the authors need to demonstrate their competence in
identifying possible problems and quantifying them. When so many analyses
are based on similar datasets and similar analysis techniques, the quality
of the systematic error discussion provides guidance to the quality of the
Of course, a good experimental paper is written clearly. It is
particularly hard to avoid the jargon that pervades a 600-person
collaboration. It becomes the lingua franca and you forget that it is
incomprehensible to the rest of the world. We need that jargon and the
acronyms to communicate efficiently, but it is hard to shed them when it
is time to write.
A good paper should show some craftsmanship. The quality of the
presentation inevitably affects the readersí judgment of the result. If
the paper is sloppily written, you worry that the analysis was sloppy as
well. Good writing is simple and to the point. I am happy to report that
you can find Strunk and White [authors of the definitive writing guide,
The Elements of Style] in the BABAR
graduate student room at LBNL. Actually, it belongs to a Russian whose
knowledge of English grammar is far too good to be that of an American,
Canadian or Brit.
Figures and tables are critical parts of a paper. They express results
that cannot be conveyed in words. Designing an effective figure is not
easy. You need to worry about proportions, about the scales for the axes,
about labels. You need to distinguish real data from Monte Carlo, signal
from background. It is important that the reader quickly grasp the meaning
of the figure. Tables are not so subtle, but it is still important to make
them attractive and consistent. Have you used the correct number of
significant figures? Are the units correct?
Q: How does BABAR
ensure that its papers are good?
works very hard to make sure its papers are correct and well presented.
These efforts are demanding enough to generate some controversy. Are we
putting too much effort into our internal review process? Do we worry too
much about questions of style and grammar?
Papers are written by small groups of authors or, occasionally, a
single author. The authors function within Analysis Working Groups (AWG).
The AWGs provide the first critical audience for a new analysis. When the
analysis is far enough along and well enough documented, the Physics
Analysis Coordinator signals to the Chair of the Pub Board that a Review
Committee is required. The three-person review committee ideally includes
one member of the Publications Board, one person from the AWG (though not
an author) and one other, an outsider.
When the review committee believes the analysis is complete, it informs
the Pub Board Chair, who announces the start of a collaboration-wide
review. During this time, every member of the collaboration is invited to
critique the paper, which by now is a full draft of a journal. To make
sure the review is thorough, a number of institutions (now typically 12 of
them!) are asked to review the draft paper formally, posting their
critique on-line using hypernews.
After the collaboration-wide review, the authors respond to the
critiques, posting responses again on hypernews. When the review committee
believes that the authors have responded adequately and made the
corresponding changes in the draft, they notify the Pub Board Chair. At
this point, the Pub Board Chair announces that the draft is in Final
Notice and that, if no problems are identified within a week, the draft
will be submitted to a journal.
In parallel with the final notice, two members of the Pub Board do a
final reading of the draft. At a minimum, they catch remaining grammatical
problems, typos, etc. They may find other problems, and occasionally there
are real issues of physics that are identified at this stage. Once the
Final Notice and final readings are done, the paper is submitted.
Q: What is your role?
A: I have primary responsibility for forming the review committees. I
sit in on a large fraction of the practice talks for conferences (which I
havenít discussed here). I am involved in formulating policy on
publications. I serve on a few review committees myself. I try to be one
of the final readers on every journal paper.
I want BABAR
papers to be something we can be proud of. This takes real
work. At the same time, we need to keep some sense of balance. We need to
get results out and we canít really take the time to make each of them a
literary gem. What we can and should do is make sure they are clearly
written. This is hard enough.
Q: How significant a metric is the number of papers produced by a
collaboration? When does the difference in the number of papers produced
and Belle become significant?
A: The number of papers written by the two collaborations is an
important measure for people who canít actually understand the papers
themselves. It allows them to judge productivity. What the members of the
collaborations know, and that other knowledgeable people in the field
know, is that both of these collaborations are making real contributions
that are significant. Both collaborations have done a great job measuring
sin 2 beta with the J/psi Ks final state. Weóespecially Antimo Palanoódid
a great job in discovering the Ds(2317), and Belle did a great job in
following up on that discovery.
The real competition is for great results in pi pi, in rho rho, in phi
Ks, and in many other channels. This is what really counts, not the number
of papers. Still, we want the agencies who support us to see that we are
productive. Making sure we get our results into print also helps them
understand we are doing a good job.