Monday, July 9, 2012

In Defense of the Scientific Paper

http://www.flickr.com/photos/openeye/5428831/
Since leaving academia, I still find the time to serve on scientific program committees (recently NSDI, MobiSys, and SOCC) and have plenty of opportunity to read both good and bad scientific papers in various states of preparation. And although I am not required to publish papers in my current job, I certainly hope to do so -- a lot of the work we are doing at Google is imminently publishable -- it's just a matter of finding the time to sit down and write them!

Although I've blogged about how the scientific publication process needs fixing, I still feel that the process of writing a scientific paper is a hugely rewarding experience. Arguably, the primary value of scientific papers isn't in reading them, but writing them. You learn so much in the process.

Writing a paper sharpens your mental focus like nothing else. Like Japanese bonsai art or building a ship in a bottle, paper writing forces you to obsess over every meticulous detail -- word choice, overall tone, readability of graphs -- and of course more mundane details like font size and line spacing. This microscopic attention to every aspect of your work brings out a wonderful, if somewhat exhausting, intellectual rapture. I have never thought so clearly about a piece of research than when I'm in the throes of putting together a paper against a deadline.

You start with nothing, a blank editor window and some LaTeX boilerplate, some half-baked ideas, a few axes to grind and a tremendous apprehension at how much your life is going to suck between now and the deadline. You throw in all of the raw ingredients, the rough ideas, the broken implementation, the confusing data, the missing citations. Over a period of days or weeks you grind it and refine it and throw it out and start over and eventually hone the paper to a razor-sharp, articulate, polished fourteen pages of scientific beauty, and then just hope like hell that you didn't screw up the margins or forget to cite some important piece of related work.

I used to think that writing a paper was something you did after the research was over, but now I realize you should sit down to write the paper as early as possible -- sometimes before even starting the "research work" itself. On a few occasions, it wasn't until I started writing a paper that I knew what the hell the research project was really about. Case in point: Our SenSys 2009 paper on the Mercury wearable sensor platform came out of a project that had been running for nearly two years without a clear set of goals or any real insight into what the interesting research problems were. We had built a prototype and had some stuff working, but we didn't know what was publishable about it, and most of the problems we had to solve seemed mundane.


In a last-ditch measure to revive the project, I got the students together and said, fuck it, let's write a SenSys paper on this. As we started piecing together the story that we wanted to tell in the paper, we realized that none of our work to that point tackled the most important problem: how to ensure that the sensors produced good, and useful, data when there was a hard limit on battery lifetime. With the deadline just weeks away, the students pulled together and reimplemented the system from scratch and cranked out a ton of new measurements. The process of writing the paper resulted in a flood of new ideas, many of which bled over into my other projects, ultimately resulting in a half dozen papers and three PhD theses. It was awesome.


And even if a paper does not get accepted, crystallizing the ideas through the process of putting together the submission can be really energizing. I never assumed any paper I wrote would actually get accepted, so submitting the paper was often the start of a new line of work, riding on that clarity of thought that would emerge post-deadline (and a much-needed break of course).

19 comments:

  1. Simon Peyton Jones also emphasises the importance of writing the paper early on as part of the research exercise. http://research.microsoft.com/en-us/um/people/simonpj/papers/giving-a-talk/giving-a-talk.htm

    ReplyDelete
  2. I think it's hard to overestimate how valuable it is to be in the habit of regularly writing and speaking about what you're doing. It's far too easy to spend all our time in our internal dialogue without ever trying to express those ideas to others; when we do, things generally come out muddled, and as it's generally hard to get people excited about poorly communicated ideas, you're likely to continue on with an audience of one.

    ReplyDelete
  3. On a few occasions, it wasn't until I started writing a paper that I knew what the hell the research project was really about. Case in point: Our SenSys 2009 paper on the Mercury wearable sensor platform came out of a project that had been running for nearly two years without a clear set of goals or any real insight into what the interesting research problems were. We had built a prototype and had some stuff working, but we didn't know what was publishable about it, and most of the problems we had to solve seemed mundane.

    I guess being able to talk about stuff like this is one of the luxuries of not being in academia any longer, but this is a serious dereliction of duty to your students. Thankfully, it seems like it all worked out at the end.

    ReplyDelete
  4. I can resonate a lot with this post. I've had multiple research projects where we wander around a bit, not exactly knowing where we're going, but knowing we're getting good stuff and moving in the right direction. Then, we sit down and start working on a paper, and as we try to "write the story", we find the gaps in our knowledge that we have to go back and fill in order to present our complete picture. I think most projects work this way!

    Even for theory papers, where you supposedly have proofs "figured out" when you start writing, I find that writing things down usually makes a huge difference. You realize there are special cases you haven't covered that you need to take care of, or recognize a tangent that suddenly seems important and worth following. Unfortunately, sometimes in the act of writing it down, you realize your proof is broken -- and then, back to the drawing board.

    Definitely it's good to start writing early in the process. But maybe not too early -- the "wandering time" in the beginning is also important, as you don't want to get locked in to a specific path before you've had a chance to explore.

    ReplyDelete
  5. @Anon Re: "Dereliction of duty": This depends very much on your attitude towards student advising and mentoring. My approach was generally to let the student be in the driver's seat, which means that some projects would indeed flounder. I think Michael gets it right when he says that the "wandering time" can be just as important as the paper-writing phase of the project. In that case, the wandering time took a lot longer than it should have.

    ReplyDelete
  6. while writing may be necessary for the publishing enterprise, it is important to recognize that writing is not a "sufficient" condition for doing good science. it is sad if at the end of the day, your research problems and results are built mainly because of writing, and if you tend to see the rest of your project as "wandering aimlessly" until writing the final 14 pages. a bad analogy - it was not as if newton sat down to write about apples and realized he could "package" or sell the whole thing as a "theory of gravitation".

    ReplyDelete
  7. Matt (9:10) : I was definitely trying to subtlely make the point that it was not (from what you've said) a "dereliction of duty" on your part, with my statement on wandering time. We agree on this one!

    Anon (9:11) : Just to be clear, "aimlessly" is your word. My research wanderings are rarely aimless. The point is that in research (unlike, say, many cases of product development), you don't know exactly what the end goal is, even if you have a direction or a sense of the destination. You need to give yourself time to learn, explore, and possibly even have that Eureka moment, that allows you to the chance to find something that maybe is as exciting as gravity.

    ReplyDelete
  8. I once heard Vern Paxson advising someone that essentially you needed to do something twice: the first to really understand it and what the real problems were, and the 2nd time to do the research and do it right. One could think of the "wandering around" as the first step.

    ReplyDelete
  9. Matt, re “it's just a matter of finding the time to sit down and write them,” I believe this a fundamental problem with trying to do research in a company, not a superficial issue that could be addressed by, say, getting a little more organized.

    When I finished my PhD, I partnered with some technology-loving geeks who had a small but very profitable company. They set me up with my own lab, a large discretionary budget, the ability to hire some great people, and even an amazing view of the Charles River! I figured I had everything I needed.

    But I found that it was difficult to find the time to write up (and present) the cool stuff we were doing. I eventually concluded that it was because research is not the company’s ultimate purpose---making money is. If your efforts are aligned with the company’s goals, writing and presenting papers is often not the best possible use of your time. (And even if it is actually a good use of your time, it may not be perceived that way by others in the company, including whoever does your performance review.)

    I’ve been moving in the opposite direction from you along the academia-industry gradient. After the small co, I joined Intel, where it was easier to justify publishing, and I am now a prof at UW, where research and publishing are fundamental. I find it very satisfying that my desire to do research and publish is 100% aligned with the University’s mission, and not something I have to explain or think hard about.

    So, my claim is that the difficulty in finding time to write papers in industry is a symptom of the fact that the ultimate mission of companies is to make money, rather than knowledge.

    ReplyDelete
  10. My guess is that nearly all advisors would be guilty of this sort of "dereliction of duty" if a hard look were taken at our research programs. Even when we're not overworked or otherwise busy doing things that aren't research/advising, we often don't know what we're doing. Many of the good student outcomes are partially or largely dependent on good luck. Fundamentally, we put ourselves and our students in a position where our intuition says something good might happen and we work hard and hope. I'm not defending any kind of malfeasance, just trying to characterize the process as I see it existing in the real world.

    ReplyDelete
  11. Josh - I don't think it reduces to something as simplistic as "making money" versus "knowledge". Google, Microsoft, IBM, AT&T, and many other companies place a high value on producing published research results. This has tangible benefits to these companies in terms of moving technology forward, as well as recruiting some of the smartest people. See my recent blog post on this:

    http://matt-welsh.blogspot.com/2012/06/googles-hybrid-approach-to-research.html

    In my case, it is true that my priority is building products, but publishing papers is valued and something that we certainly hope to do. I think it depends very much on the company you work for and what value they place on publishing research results. It varies a lot, but I don't agree you can make a black-or-white comparison between academia and industry in this way.

    ReplyDelete
  12. How will I know what I think until I see what I write?

    ReplyDelete
  13. Thank you for these few words. I am looking forward to have my first paper written after I finish my internship and I am a bit a bit excited but nervous because it seems to be a difficult task.

    A valuable post that deserves every scientific related engineer reads.

    And also it is nice to know that big companies as Google takes care and recognizes paper publications.

    ReplyDelete
  14. "You start with nothing, a blank editor window ...", that para is the most passionate, amorous, piece of writing I have seen defending publications. Just brilliant !

    ReplyDelete
  15. There is a vast difference here between writing when you *can*, and writing when you *have to*.

    If your job depends on N number of "publication points", you have to consider the trade-offs between writing the paper you *want to* write, but will get little to no credit for, vs. some bastardized POS that will satisfy the editorial egos of an A-grade journal.

    The writing of a paper is certainly a cathartic process, but the current publishing system often makes it meaningless.

    The writing of a paper is indeed a helpful process; what happens afterwards however is horribly broken.

    ReplyDelete
  16. "imminently publishable" or "eminently publishable"?

    ReplyDelete
  17. “I used to think that writing a paper was something you did after the research was over, but now I realize you should sit down to write the paper as early as possible” -- can't agree more! For some wired reason people tend to be so lazy that we hardly think deep until several weeks before the deadline, regardless how far away is the deadline... I'm always wondering, if a PhD can be constantly as productive as near deadline, how successful this guy will be

    ReplyDelete
  18. Writing the paper early is another example of Test Driven Development: The first test case is the conclusion of the paper. The test fails, so you have to do some work to make the test succeed...

    ReplyDelete
  19. My experience show me that the writing of a scientific work must be done after completing the research project. During the writing period, we might modify some parts of the work, say, some implementations approaches or extending some parts of it. In my own side, the writing period is a major complement phase toward the performed research work.

    ReplyDelete

Startup Life: Three Months In

I've posted a story to Medium on what it's been like to work at a startup, after years at Google. Check it out here.