Sunday, June 21, 2009

Not Worth the Paper ...

Science too often trivializes the profound, answering questions that are very different from the ones that were asked. To formulate a question suitable for scientific research too often requires us to forget what it was that we really wanted to know. ~Earon Davis

A couple of French astronomers with time on their hands and some computing power available have determined that it you tweak Mercury's orbit by very small amounts, chaos will ultimately occur in the solar system. Well, ok, and if I drop another planet in between the orbit of Earth and Mars, some bad things will also occur.

I'm not at all sure what the point of their exercise is supposed to be. Is this another one of those "had their been a slight variation in starting conditions, we wouldn't be here" scenarios? What is the mechanism for chaniging the semi-major axis of Mercury? And, since they were working in 5 billion year intervals, did they take into account that the sun will probably have gobbled up Mercury before it could pinballing around the inner solar system?

But the ultimate question is: What the heck was it they were trying to do in the first place?

There are lots of interesting questions about the solar system's orbital mechanics. For example, how did the asteroid belt come to be? Why didn't everything get kicked out of the belt by Jupiter's gravity? How did Uranus end up on it's side, and why is Triton orbiting in the wrong direction? Why do all the gas giants have rings while no rocky planet has them?

In other words, if you're going to spend a ton of time developing a program to model the solar system, why waste time fiddling with Mercury's orbit when you could be trying to determine the conditions that got the solar system to where it is now?

Well, maybe, just maybe because a couple of French astronomers were trying to get published in a prestigious magazine to enhance their own reputation. Of course, they submitted the information as a letter, not a paper, which I imagine avoids peer review.

Which brings us to a paper submitted by the Center for Research in Applied Phrenology (CRAP).

The paper, entitled "Deconstructing Access Points", was submitted for publication to The Open Information Science Journal by CRAP researchers David Phillips and Andrew Kent. And it was accepted for publication "after peer review", as long as Phillips and Kent supplied an $800 publication fee.

None of this would be hugely unusual except that Phillips and Kent were actually Phillip Davis and Kent Anderson, and their CRAP paper was actually a pile of nonsense generated by a computer program designed to generate phony research papers. The name of their "research" outfit was deliberately designed to send a huge hint to Bentham Publishing, the outfit responsible for the journal, that maybe their collective legs were being pulled.

Evidently, Bentham editors didn't care, as long as the check was good.

Bentham, of course, now claims that they knew it was a gag all along, and they were just stringing these guys along to find out who they were. Interestingly, the editor of the journal subsequently resigned. Evidently, he wasn't in on the investigation by his own staff.

Now, add these to the incident of the "missing link" fossil that was sort-of-but-not-very-peer-reviewed, and you have the makings of a disturbing trend.
The pressure to publish, as I've said before, is huge. There is a lot of competition for research dollars, and getting published is one way to get hold of them. The problem is that getting published may not require that what is published have any particular scientific merit. Or, as in the case of the fossil, it may have merit not but justify the hype-filled conclusions.

Then there is the reputation factor. Jorge Hirsch, self-proclaimed genius, has determined that scientific reputation is determined by where you get published and how often you get cited. Interestingly, the h-index, as Hirsch has dubbed it, gives Hirsch a very good rating. Even more interestingly, the h-index takes no account of the quality, validity, or originality of the publications.

What's frightening is that there appear to be scientists to take this nonsense seriously.

At least the attitude that publishing anything is all the matters goes a long way toward explaining some of the whacko theorizing that has become the hallmark of the 21st century so far.

It might even explain dark energy.


jj mollo said...

Orbital mechanics is not as intuitively obvious as you might imagine. There are, for instance, low energy ways of moving satellites around in surprising ways. We have used some of them, one at least for chasing an unexpected visitor from the Oort cloud, another for rescuing a Japanese satellite. The disconcerting implication is that there may well be relatively easy ways that large objects could be disrupted. If I were looking for proof of such a thing, I imagine Mercury would be the first place I'd look as well.

If there is any reasonable support for the idea that we should protect ourselves from rogue asteroids, then it seems to me we should also know where the sensitive spots in our solar system actually are. Such spots might also require protection.

It's interesting to me that the more we know, the more responsibility we seem to have.

I met one of NASA's orbital mechanics gurus once upon a time. He was different than the rest of us mortals.

The Gog said...

Actually, I do know a little about orbital mechanics, and I understand that it isn't simple. That being said, scientists and engineers have been tackling orbital questions for years quite successfully.

There are numerous suggestions from very reputable people about avoiding catastrophic impacts by inducing small variations into an object's orbit. Had these scientists studied the effect of some small perturbation of any of the many NEO's, that would have been sensible.

My point was that proving that a little change in orbital conditions can cause drastic effects ain't news. And, in the context of the planet Mercury, it is little more than an intellectual exercise, most likely done to get something into publication.

jj mollo said...

I certainly agree with your main point. Publish or perish entails some seriously counter-productive incentives.

Peer review isn't all it's cracked up to be either. My pet peeve is the Iraq Mortality Study as published in Lancet by an epidemiological survey team based at Johns Hopkins. It had a profound political effect, but was full of holes.

My comment was directed toward your specific choice of that article. I read what I think is the original letter from Nature. I found it pretty interesting. It seemed to be a response to an earlier set of simulations that imposed several corner cutting measures. The letter was emphasizing the importance of the relativistic effects, hence Mercury, and inclusion of Luna in the model.

I don't know how good these models are, but I suspect there will be a response. Someone will probably say the authors should have included something else, or the relativity concerns are overblown, or the granularity was the whole cause of the difference, or there was a rounding error in the zillionth iteration. I guess there is also a reasonable question as to whether the whole thing is worth the computer time. If you like simulation type research, as I do, then maybe you think it was worth it. The methodology question could ultimately have a bearing on other simulation investigations, such as the very important climate modeling work.

jj mollo said...

Early in my programming career, some of my friends developed a buzzword generator for filing their weekly progress reports. I thought this was scandalous, and it probably took more time than the report would have taken, but it was funny.

In 1996 there was a physicist who did the same sort of thing to generate a postmodern deconstructionist hoax. It was, amazingly, accepted for publication in one of these rather self-conscious journals. The guy claimed that such an oversight proved his point that the whole field was phony. I read it at the time and it was hilarious.

The Gog said...

I do believe you would have an appreciation for Isaac Asimov's famous paper, “The Endochronic Properties of Resublimated Thiotimoline”, which I discussed here.

I think computer simulations are a useful tool, but it does seem that they have become an end unto themselves, as well as an easy way to justify any theory by juggling initial conditions. Just because a simulation comes out a certain way does not "prove" anything; it merely shows plausibility under the constraints entered into the program.

To keep simulations in perspective, one should always keep in mind NOAA's hurricane path predictions.

jj mollo said...

Asimov was a remarkable writer. Years ago I tried to read everything he wrote. I enjoyed his Foundation series in particular. I still have some of his non-fiction books that I have planned to read for years. He could write faster than I could read.

Simulations are improving all the time. It's frustrating work, I know, because complexity increases geometrically while effectiveness improves arithmetically, but it does happen. Here is NOAA's self assessment on hurricane path forecasting.