Tuesday, August 15, 2017

Are Scientific Conferences Obsolete?

Many of you have seen this iconic photo, which hangs in physics departments all over the world:


It's the Solvay Conference of 1927, the formative era for quantum mechanics, when giants walked the Earth: Einstein, Heisenberg, Schrodinger, Curie. I've often wondered what it must have been like to be one of the two or three people in the photo that no one has ever heard of. At least you'd get your photo on lots of physics department walls.

Recently, I attended a physics conference myself, TeVPA 2017, hosted by Ohio State University. The conference covered the overlap between particle physics and astrophysics -- my main interest was dark matter. But the conference itself was something of a Rip Van Winkle experience for me. I've stepped down as department chair after 13+ years (hurray!), so I am just now getting back onto the conference circuit. And I've noticed one tremendous difference between the conferences of my youth and the one I just attended.

Let me first take a step back and talk about a quiet revolution in the way that physicists do their work. It's something that most people aren't even aware of, but it's had a profound effect on the way that physics gets conducted.

A scientist with new research results will generally write a paper about the results and send it off to a scientific journal. The editor sits on it for a few days, and then forwards it to one or two "referees" for review. The referees check the validity of the paper, judge whether it's interesting or not, and then send a report back to the editor, who decides whether or not to publish it. If it does get accepted for publication, then the journal typesets it, sends it back to the author for proofreading, and eventually publishes it.

Back in the pre-internet days (say, before 1995), this whole process, from submission to publication, could take as long as a year or more. But who wants to wait that long? So we physicists resorted to our own independent publication system. We would print up 50-100 "preprints" and send them out to top physics departments and colleagues working in our field. You can see the obvious disadvantages of this system -- scientists who were not at leading institutions or otherwise "out of the loop" would never get to see these advance publications. And the rest of us could never really be sure we weren't missing some sort of important paper. It was not unknown for nearly identical results to be submitted to journals, with the scientists unaware of each other's work.

And then came the revolution. In the mid-1990s, the physics preprint arXiv appeared. Now physicists had a central repository at which they could upload all of their papers, and their colleagues could be certain that they were missing nothing. Essentially all important new work in physics and astronomy gets posted to the arXiv before it's submitted to a journal. And if you peruse this website regularly, you will have the same access to the latest physics results as the top people at the best universities in the world.

Within the span of a couple of years, preprints disappeared. But the arXiv did more than replace preprints --  it also supplanted, to some extent, the role of the scientific journals themselves. We still publish our papers in journals -- they remain the gold standard for hiring, promotions, and annual reviews of scientists. But as a means of transmitting scientific information? Not so much. I keep up with the arXiv every day, but I almost never look at journal articles at all, and the same is true of most of my colleagues. I should mention that physics (especially theoretical physics) and astronomy are in the lead in this regard. Biology and biomedical research still rely much more on refereed journals (and a good thing! I don't want my doctor recommending a treatment based on someone's latest internet posting).

And what does this have to do with physics conferences? I used to attend conferences with the expectation that I would learn about all of the most exciting new results in my field. But at this recent conference, pretty much every talk was based on work that had already been posted on the arXiv. In many cases, I recalled reading the original papers myself. Like refereed journals, conferences may have outlived much of their usefulness as a venue at which to present new work.

So does this mean that physics conferences are obsolete? Absolutely not -- they still play a couple of very key roles in the scientific enterprise. First of all, they allow physicists to stay in touch with the zeitgeist of their field (or for younger readers out there, "what's trending"). Am I implying that we are influenced in our research decisions by the intellectual fashions of the day, by the topics that our colleagues are working on, rather than boldly striking out on our own in a search for Truth? Don't tell anyone -- it will be our little secret.

The other important function of conferences is simply to facilitate personal contact with our colleagues. It allows us to develop new collaborations, to discuss ideas back and forth, to identify potential people to work with or to hire in the future. I would put this under the general category of "schmoozing," and it's one thing the internet will never replace.

No comments: