I received word yesterday that the last paper from my dissertation work has been accepted for publication pending some "minor" changes (I use the quotes, because some of the required changes don't seem so minor). It will be nice to put a bow on this -- I worked pretty hard on my dissertation, so it's nice to see it more or less completely published in some form or another, it's good closure -- but it might very well be the last academic paper I write. I'm not in academia anymore. I don't have any intentions to get back into it, and I'm probably not going to be solving a whole lot of operations research / applied math problems in my spare time, because it's my spare time, and I'd rather do other things.
In a way, it's a shame. I had started to build up a decent resume, and establish myself in the field a bit (not enough to garner interest from any of the academic positions I applied for in the past, but still). In another way, it's a relief, because the publishing process in my field isn't very enjoyable. It's a) slow and outdated, b) competitive in the wrong way, and c) too amenable to self-serving interests.
On the first point, you submit a paper to a journal, they give it to referees, it sits in the refs' inboxes for several months, they give it to their grad students, it sits in the grad students' inboxes for a few weeks, they eventually read it, and if you're fortunate they recommend it be accepted, with changes of course, nothing gets accepted straightaway. So it comes back to you, you make the changes, resubmit and the process repeats. By the time it finally appears in print it is about three years after you finished the initial version, literally. By that time everybody else has move on to other things. It's insane. Here's a better way. Each institution should have its own online paper portal. Once you finish, you put it up on that (you can have internal reviewers to make sure papers meet a certain standard) and let other people look at it. Then the journals can browse these portals and grab the papers that have generated the most interest or those that they otherwise like and publish them. This way researchers can still get by only reading the journals, but things would get done in a much more efficient and timely fashion.
On the second point, the thing I hate(d) most about publishing in my field is the need to "improve" upon the best solutions in the literature by some miniscule percentage. For example, I worked on a vehicle routing problem where the objective was to get the most efficient routes (i.e., the shortest routes) possible. I came up with a new procedure and applied it to some standard benchmark problems in the literature. My routes were about 0.5% above the routes from the best-known procedure (not 5%, 0.5%). I then spent the next few months tweaking my parameters, so that I "beat" the best-known procedure. I was doing almost nothing creative, just running my algorithm with different initial settings. It felt like such a waste of time, but if I didn't do it, it's unlikely that a referee would have accepted my work (and it's certain my advisor wouldn't have accepted it). As a whole, our time as researchers is not well-spent shaving fractions of percentage points off solutions that are already extremely efficient. We should be coming up with new ideas to solve new problems.
On the third point, the people who are refereeing your papers are the same people who you are competing against. Numerous times I've been asked to review a paper where it being published will have consequences (sometimes positively, sometimes negatively) on a paper I'm working on, or one that I've already submitted for review. I tried to be objective, but I don't think it's completely possible. Also, the name on the paper can have a big influence. I think all reviews should be double-blind.
Anyway, that's all I have to say about that.
If this latest paper is, in fact, published it will lower my Erdos Number to 2. Your Erdos Number is the number of degrees away, through publishing, you are from the late legendary mathematician Paul Erdos. If you've published directly with Erdos your Erdos Number is 1. If you've publish with somebody who's directly published with Erdos it's 2, etc. It's like the Kevin Bacon game.
One of my coauthors on this paper is a man who published directly with Erdos, which is cool, but it's mostly superficially cool, because we didn't really directly collaborate. In fact, I provided very little of the "smarts" for this paper. What happened is, while working on something, I came upon an interesting theoretical problem. I couldn't answer it very well and was busy with other things (finishing my dissertation and graduating), so my advisor showed it to a guy who's very sharp, and he answered it about 75%. For the last 25%, my advisor asked his old advisor, basically the smartest guy he knows and the one who published with Erdos directly, and he solved it. Then, based on their notes, I wrote the paper. So, I certainly contributed. It was my problem. I got the ball rolling with some initial conjectures. I did (almost) all the writing. I also filled in many logical gaps -- it wasn't just grunt work -- but the "eureka moments" of the paper, the impressive parts, came from other people. Still, it's something to be proud of I suppose. Although it's weird, I'm not really all that proud of my academic achievements. I would much rather have a crossword puzzle published than an academic paper. I'm not sure why this is.
OK. Until next time...
[I'm not sure if this is from "The New Yorker" or not, but it's very New Yorker-esque. It's not particularly funny, but it's not particular not funny either.]
In a way, it's a shame. I had started to build up a decent resume, and establish myself in the field a bit (not enough to garner interest from any of the academic positions I applied for in the past, but still). In another way, it's a relief, because the publishing process in my field isn't very enjoyable. It's a) slow and outdated, b) competitive in the wrong way, and c) too amenable to self-serving interests.
On the first point, you submit a paper to a journal, they give it to referees, it sits in the refs' inboxes for several months, they give it to their grad students, it sits in the grad students' inboxes for a few weeks, they eventually read it, and if you're fortunate they recommend it be accepted, with changes of course, nothing gets accepted straightaway. So it comes back to you, you make the changes, resubmit and the process repeats. By the time it finally appears in print it is about three years after you finished the initial version, literally. By that time everybody else has move on to other things. It's insane. Here's a better way. Each institution should have its own online paper portal. Once you finish, you put it up on that (you can have internal reviewers to make sure papers meet a certain standard) and let other people look at it. Then the journals can browse these portals and grab the papers that have generated the most interest or those that they otherwise like and publish them. This way researchers can still get by only reading the journals, but things would get done in a much more efficient and timely fashion.
On the second point, the thing I hate(d) most about publishing in my field is the need to "improve" upon the best solutions in the literature by some miniscule percentage. For example, I worked on a vehicle routing problem where the objective was to get the most efficient routes (i.e., the shortest routes) possible. I came up with a new procedure and applied it to some standard benchmark problems in the literature. My routes were about 0.5% above the routes from the best-known procedure (not 5%, 0.5%). I then spent the next few months tweaking my parameters, so that I "beat" the best-known procedure. I was doing almost nothing creative, just running my algorithm with different initial settings. It felt like such a waste of time, but if I didn't do it, it's unlikely that a referee would have accepted my work (and it's certain my advisor wouldn't have accepted it). As a whole, our time as researchers is not well-spent shaving fractions of percentage points off solutions that are already extremely efficient. We should be coming up with new ideas to solve new problems.
On the third point, the people who are refereeing your papers are the same people who you are competing against. Numerous times I've been asked to review a paper where it being published will have consequences (sometimes positively, sometimes negatively) on a paper I'm working on, or one that I've already submitted for review. I tried to be objective, but I don't think it's completely possible. Also, the name on the paper can have a big influence. I think all reviews should be double-blind.
Anyway, that's all I have to say about that.
If this latest paper is, in fact, published it will lower my Erdos Number to 2. Your Erdos Number is the number of degrees away, through publishing, you are from the late legendary mathematician Paul Erdos. If you've published directly with Erdos your Erdos Number is 1. If you've publish with somebody who's directly published with Erdos it's 2, etc. It's like the Kevin Bacon game.
[Paul Erdos, a very interesting man. I recommend a biography of him, "My Brain is Open".]
One of my coauthors on this paper is a man who published directly with Erdos, which is cool, but it's mostly superficially cool, because we didn't really directly collaborate. In fact, I provided very little of the "smarts" for this paper. What happened is, while working on something, I came upon an interesting theoretical problem. I couldn't answer it very well and was busy with other things (finishing my dissertation and graduating), so my advisor showed it to a guy who's very sharp, and he answered it about 75%. For the last 25%, my advisor asked his old advisor, basically the smartest guy he knows and the one who published with Erdos directly, and he solved it. Then, based on their notes, I wrote the paper. So, I certainly contributed. It was my problem. I got the ball rolling with some initial conjectures. I did (almost) all the writing. I also filled in many logical gaps -- it wasn't just grunt work -- but the "eureka moments" of the paper, the impressive parts, came from other people. Still, it's something to be proud of I suppose. Although it's weird, I'm not really all that proud of my academic achievements. I would much rather have a crossword puzzle published than an academic paper. I'm not sure why this is.
OK. Until next time...
No comments:
Post a Comment