Paul Ginsparg, the founder of the arxiv preprint server for physics, has a very nice article at Physics World reminiscing about the rise of the Internet, particularly in physics. This also serves as a nice counterpoint to his talk at the Science21 conference (video, microblogging), which included a wealth of fascinating information about the current operation of the arxiv.
In both of these, he mentions that the arxiv grew out of a pre-existing preprint culture in high-energy theoretical physics. People in the field would make copies of their manuscripts in progress, and send them to other people in the field, for comments entirely outside of the normal publication system.
It occurs to me that there's something really remarkable about this: physicists have the reputation of being arrogant and hard to deal with, and theoretical high-energy physicists are regarded as arrogant SOB's by other physicists. It's really kind of amazing that this community, of all the research communities in the world, was the one to develop the preprint culture and then the arxiv.
Looked at this way, it also makes me wonder: What's wrong with biologists, anyway?
Biology, after all, is widely regarded as the most open and welcoming of all the natural sciences. The only field with a kindlier reputation among students is psychology, and that's only half a science (the other half is social-science). If you were going to guess a field that would've been among the first to embrace the idea of sharing pre-publication results, you might very well have guessed that it would be a subdiscipline of biology.
And yet, even today, seventeen years after the launch of the arxiv, every attempt to set up a preprint service for biologists has been a dismal failure, as noted by both Ginsparg and Timo Hannay (whose Science21 talk notes are up at Nature Networks. You can also get video and microblogging). Contrary to what a naive outsider's opinion might suggest, biologists appear to be highly resistant to the whole idea of sharing pre-publication results.
So what's the deal? Why are the prickly bastards in high-energy theory so open to sharing, while the squishy and lovable biologists guard their research so jealously?
- Log in to post comments
Why are the prickly bastards in high-energy theory so open to sharing, while the squishy and lovable biologists guard their research so jealously?
Because biology has more potential for patentable or potentially money-making ideas?
My sister's a post-doc in molecular Bio. and from the stories she's told me the answer is that biologists are INSANE. Like seriously, literally almost setting fire to each other insane. (And I DO mean literally here)
Also, according to my sister, Biologists believe very strongly in the competition aspect of science. Where, scooping each other is a viable option to get ahead.
In computer science (at least the part I worked in), if you're working on something and might get scooped by another group, you pretty much just work WITH the other group and everyone gets their name on the paper. But no, Biology is a zero-sum game somehow...
I have no background in biology or physics, but it might be something to do with how easy ideas are to swipe. If any experiment is easily replicable by anyone with a little spare time and some common lab equipment, it would be easier to steal (and more anonymous) than if you needed to schedule time on a multi-billion dollar particle accelerator whose schedule is widely published.
I disagree entirely with brian's anecdote. I'm sure there are lots of insane people in biology, but it is beyond silly to say they are all insane, and I'm sure the distribution of insane people in the other sciences is similar. To me the real reason people researching biology haven't taken to the preprint culture is scale and inertia.
In a subfield like high energy theory, there are probably only a few thousand high energy theorists, and of those, probably only on the order of 100 who are truly influential and important to the progress of the field. Getting those people on board with something like a preprint server is hard, but maybe not terribly so. Any given subfield (okay, maybe not _all_ of them) of biology, however, has orders of magnitudes more scientists, many more subfields that range from pure science to clinical/pharma/etc, and I can only assume the inertia to get such a massive number of people to embrace a preprint culture is very difficult.
That being said, there is also an experimentalist/theorist divide on the arXiv (I feel like there are far fewer important experimental papers that make it onto the arXiv before publication, though it is admittedly only a guess), and there are also far fewer theorists working in biology.
The irony of it all, however, is that the only truly "high impact" open access journals are all biology journals (hello PLoS), so perhaps the real question is, what the hell is wrong with all those physicists?
I disagree entirely with brian's anecdote. I'm sure there are lots of insane people in biology, but it is beyond silly to say they are all insane, and I'm sure the distribution of insane people in the other sciences is similar. To me the real reason people researching biology haven't taken to the preprint culture is scale and inertia.
In a subfield like high energy theory, there are probably only a few thousand high energy theorists, and of those, probably only on the order of 100 who are truly influential and important to the progress of the field. Getting those people on board with something like a preprint server is hard, but maybe not terribly so. Any given subfield (okay, maybe not _all_ of them) of biology, however, has orders of magnitudes more scientists, many more subfields that range from pure science to clinical/pharma/etc, and I can only assume the inertia to get such a massive number of people to embrace a preprint culture is very difficult.
That being said, there is also an experimentalist/theorist divide on the arXiv (I feel like there are far fewer important experimental papers that make it onto the arXiv before publication, though it is admittedly only a guess), and there are also far fewer theorists working in biology.
The irony of it all, however, is that the only truly "high impact" open access journals are all biology journals (hello PLoS), so perhaps the real question is, what the hell is wrong with all those physicists?
RE: Phillip @4. All I can say is that the stories my sister has told me about the crazy in molecular biology have no equivalent in computer science, or any other field that I know of.
I'm sure there are lots of insane people in biology, but it is beyond silly to say they are all insane, and I'm sure the distribution of insane people in the other sciences is similar. To me the real reason people researching biology haven't taken to the preprint culture is scale and inertia.
There are probably only a few thousand high energy theorists, and of those, probably on the order of only 100 who are truly influential and important to the progress of the field. Getting those people on board with something like a preprint server is hard, but maybe not terribly so (particularly since it has happened). Any given subfield (okay, maybe not _all_ of them) of biology, however, has orders of magnitudes more scientists, many more subfields that range from pure science to clinical/pharma/etc, and I can only assume the inertia to get such a massive number of people to embrace a preprint culture is very difficult.
That being said, there is also an experimentalist/theorist divide on the arXiv (I feel like there are far fewer important experimental papers that make it onto the arXiv before publication, though it is admittedly only a guess), and there are far fewer theorists working in biology.
The irony of it all, however, is that the only truly "high impact" open access journals are all biology journals (hello PLoS), so perhaps the real question is, what the hell is wrong with all those physicists?
Is it really necessary to post the same comment 20 times?
There is a false premise in Chad's article that might be relevant to the biology discussion. Although ArXiv, as the successor to xxx.lanl.gov, is heavily loaded with particle theory, this is only partly reflective of the origin of the idea of electronic distribution of preprints in a part of hep-th, not the use of preprints themselves.
Paper distribution of preprints by mail was commonplace in nuclear and particle physics, both theory and experiment, prior to the mid 90s. Electronic distribution simply replaced mail and is now coupled into journal submission for some journals. Typical practice back in the day was to mail out a preprint coincident with journal submission, but there were cases where early reports were circulated for comments and suggestions (like peer pre-review). [I gather math uses it a lot for pre-review.] I regularly got experimental results this way, often six months before the printed article would appear in a journal, even earlier for some slow-moving journals.
To my mind, the question should be whether biologists send out preprints at all. If you never or rarely sent them out, you don't need an electronic way of doing it.
A separate, and interesting, question would be the current ratio of never published to published work (by subfield) in arXive. My impression is that there is a lot of volume from samizdat publishing of highly speculative ideas in particle theory.
Wow, that's awesome. Sorry for the spam.
I have intervened to remove some of the duplicate comments, because it was getting a little ridiculous.
We're having some problems with the commenting system, and it's throwing a lot of submission errors. Most of the time, these are just local time-outs, and the comment has actually been properly submitted and will appear when you re-load the original page.
PLEASE DO NOT RE-POST YOUR COMMENT WITHOUT CHECKING TO SEE IF IT WENT THROUGH.
This is speculation on my part, but I would guess that the Manhattan Project played some role in the development of a preprint culture in physics. Most of the important physicists of the 1940s (at least those who worked in the US, plus a few foreigners) had some connection to that project, which by its nature was a massive collaboration, so there was a need to communicate works-in-progress and other latest results. Many of the Manhattan Project scientists, such as Feynman and Bohr, made major contributions to quantum theory which were essential to the subsequent development of particle theory, and others, such as Wilson, would later specialize in particle physics.
I do not know of any other field which has had anything even remotely resembling a Manhattan Project, and certainly the field of biology is too big to have that degree of collaboration across the entire field. Probably physics itself is too big nowadays. But the practice of mailing out (p)reprints remained until the World Wide Web made this use of snail mail obsolete.
In addition to the ArXiv, and the preprint culture that preceded it, the theoretical high energy community is responsible for a generating a few other models of collaborative research, which have not been adopted widely elsewhere. These include workshops (small gatherings with no formal talks, geared towards generating collaborations), and research centers devoted to hosting those workshops, video conferencing to share seminars and graduate courses, open access peer-reviewed journals (e.g. JHEP), and a few other ideas I am probably missing now. The HEP theory centers (starting with the Rutgers "string palace" in the early 1990s) have come up with a series of innovative ideas on how to encourage collaboration, including clever design of the basic architecture of the work environment (you might have seen some of that reflected in the design of PI). In short, the HEP theory sub-field has a long and ongoing tradition of community building and collaboration, maybe the puzzle you point out is the result of one bad premise...
Preprint cultures grow up where there's value to the recipient of the preprint. Mathematicians had one. Mathematicians spend their working lives in a state of cluelessness, wondering what to try next. Maybe Fred's paper will supply a clue. It's better to get that clue early, when Fred submits to the journal (and sends you a preprint), than late, when the journal finally publishes it.
I assume the same sort of thing happens in physics. Theoreticians are better off hearing about an experimental result that their theory doesn't handle well earlier rather than later; experimentalists are better off hearing about a theoretical prediction before they assemble their experiment rather than after the grant request as gone in.
And if receiving preprints helps you, you have to send them out, too. They'll show you theirs if and only if you show them yours.
But maybe biologists aren't helped by hearing of others results early. There isn't quite the theoretician/experimentalist divide in biology.
Making preprints was tiresome. If the recipients' reaction is to throw them away, people won't make them.
I don't think it's a question of fear that ideas will be stolen. Date of submission to the journal establishes priority, and preprint circulation typically occurred concurrently with journal submission.
I'd say false dichotomy, except there are more than two things you can do with a blog (e.g., how about a blog as a teaching tool, targeted by set of topics and discussion level).
Ok, I'll bite into this apple as a "biologist." First, we do have a ton of sub-disciplines that we deal with. Depending on your take of my research (I used to do coastal fisheries biology and oceanography before I became a "manager"), any research I do could be published in two or three dozen different journals, in different disciplines. And, what I might publish in Marine Ecology Progress Series, for instance, might not appeal to the readers of Nature, Oecoligia, Modern Oceanography, or the Journal of Fisheries Management. As you can see, for "biologists" this can get out of hand quickly.
There is also a bit (or a lot) of "disciplinary tradition" involved. As has been noted, there is often a lot of very cut-throat competition for "biology" research grants, and as the funds dry up, the competition gets worse. "Our" tradition is to keep our ideas to ourselves, until we publish, so we can get established in the pecking order and increase the likelihood of future funding because we are "cutting edge." Sad - yes; true - also yes. Perhaps the Physicists can help us redeem ourselves with a 12 step program or something.
I'd say false dichotomy, except there are more than two things you can do with a blog (e.g., how about a blog as a teaching tool, targeted by set of topics and discussion level).
Geez, my comment appeared twice, both times in the wrong thread. Feel free to erase.
Paper distribution of preprints by mail was commonplace in nuclear and particle physics, both theory and experiment, prior to the mid 90s.
This was true of astronomy as well.
For what it's worth, Ginsparg doesn't seem to claim (in the Physics World article, at least) that high-energy physics invented the preprint culture -- though he does claim to be the first to attach an email address to a preprint. So I think Chad's suggestion that "this community, of all the research communities in the world, was the one to develop the preprint culture..." isn't really justified.
(I mean, it might turn out to be true that the preprint culture did start in high-energy physics, but no one's offered any evidence for this claim.)
Well, duh, it's that the fizzycysts are just ever so much more clever that they never have to worry about being scooped. Anyone can do biology, so you have to keep your ideas secret.
< /sarcasm >
"Biology, after all, is widely regarded as the most open and welcoming of all the natural sciences. " Ha. Hahahahahahahhahahahahahaha. HA!
Widely regarded by whom? Seriously, wtf mate?
I think your basic assumption, "prickly bastards" and "arrogant SOBs", is flawed. There are some of those, yes, but they're more likely to be prickly and arrogant toward outsiders than toward others in their field. (There's a reason the sneering you see from certain people on the interwebs is directed at, say, Garret Lisi or Peter Woit, not at even relatively incompetent people doing serious theoretical high energy physics.)
And if we judge the character of a field by blogs, then clearly PZ Myers is just as much a prickly bastard and arrogant SOB as anyone writing about physics. (And again, the prickliness and arrogance is directed toward people who arguably deserve it.)
What Moshe says is true: this is a field that strives to encourage collaboration, a field where people love to gather in interesting places and talk to each other about whatever they're doing and try to generate new ideas, a field where even before the preprint hits the arxiv, you're likely to know what persons X, Y, and Z are doing and what their results are.
From the biased perspective of someone in biology/related field:
We spend multiple years and hundreds of thousands of dollars setting up the experimental system to get out a paper. If it gets scooped, everyone involved in the project is out of luck. The next research grant or renewal won't be funded, because you don't have the publication background to prove that you are at the forefront of your research area. Your grad students won't get good postdocs, your postdocs are out of luck. So sad, maybe you shouldn't have shared your results early.
If it works, we can run a number of different experiments in that system, become established in that area, and send out papers at a fairly rapid rate while our next system develops.
The theoretical physicists and mathematicians I know don't seem to have that startup period, and send out papers once a month. If any single paper gets scooped - oh well, just write a new one. One more or fewer when you've got 20 to your name as a grad student just isn't a big deal.
Experimental physics would be a much closer comparison to biology; wasn't there a massive brouhaha recently resulting from someone analyzing someone else's data based on a conference presentation, complete with references to "if they'd only published on arXiv we wouldn't have this? With rebuttals of "well, the major journals won't accept for publication if it's gone on arXiv"? That is a lot closer to what would happen with biology preprints.
I have collaborators who don't even want me to talk about our results at closed conferences in order to maximize the impact of our results when we finally publish, and minimize the chance someone else will propose our ideas first.
Moshe @ 13:
In addition to the ArXiv, and the preprint culture that preceded it, the theoretical high energy community is responsible for a generating a few other models of collaborative research, which have not been adopted widely elsewhere...
I find it highly dubious that these were all invented first and only by high-energy physics.
Peter Erwin, you are absolutely right, some of those structures exist in other communities, but at least some of them really did originate in the HEP one. More to the point, these are not very common in Biology, I think, or even in most subfields in Physics. One possible cause is that theorists and experimentalists have very different needs when it comes to collaboration, and therefore it is not unexpected theorists put a higher premium on it, and design structures to encourage it. Biologists, for the most part, are tied to their labs and are in direct competition for resources, not unlike experimental physicists, so collaboration is not an absolute necessity. It is kind of unfortunate the "lone genius" model is associated with theoretical physics, nothing could be further from the truth.
RE: Reesei @22 I think that the difference is that in physics, for the most part, if there's 2 groups working on the same issue, then they're going to be co-authors on all of the relevant papers. Possibly because the equipment is SO expensive that if someone's already built the apparatus, then what's the point of building another one, if you can just use theirs?
Computer science seems to work the opposite. that everything's electronic to begin with, so you might as well just shove it up on your website, and let people read it...
Brian, if I may, I suspect the reason there are fewer crazies in CS (if that is indeed the case) is that with the Googles and Microsofts and what-have-yous, a Ph.D. in computer science pretty much guarantees you a $100,000+ job if you want it (even if it's not necessarily doing exactly what you want). This is in marked contrast to other fields, where a Ph.D. carries no guarantee of any job at all. That can be quite a bit of stress...
Also, being in CS myself, I am pretty careful about what I put on my web page and when. Conference reviews are double-blind, and putting papers up while they're in review is a big no-no (in my subfield at least). Putting them up after being reviewed for a conference but before actual publication in the proceedings or in a journal is indeed common, however.
I've got a background in two fields, organic chemistry and biology, and I think that neither field is going to embrace a pre-print culture anytime soon. It's interesting that some people mention publishing papers as co-authors instead of competing; the closest I've seen to that in biology is the coordination of simultaneous publication of a set of papers, back-to-back in the same journal. That's the most congenial a lot of biologists get, and even that wouldn't happen in (organic) chemistry.
I think the problem is that first, the fields are both gigantic, so relative anonymity or lack of personal contact with the person you're contending with breeds bad social behavior (this reminds me of large internet forums...hmm...). Next, both fields move quickly, and the next steps in many experimental situations is rather obvious. Sure, serendipity often results in a lot of interesting stuff, but there are a ton of churn-and-grind experiments that are important results, too, lending themselves to insane work hours and cut-throat competition. The latest experiments in quantitative proteomics via mass spec are one example that comes to mind. Finally, funding itself is fairly competitive. There's a lot of money at stake, but not nearly enough to go around, so it basically works as a zero-sum game. Especially with so few tenure-track faculty positions available, with so many people competing for those jobs and competing for tenure, it's kind of hard to see how the culture would end up in any other way. First-author publications are absolutely necessary, in both quality and quantity, to "make it" and get funded, so there are very few social incentives for being very "open" about the publication process.
I don't know the history of the field too well, but maybe high-energy particle physics was able to keep a culture going because it all started with a small community? In addition, if they need to pool their money into large collaborations, that kind of thing would force them to overcome their personal differences...
As a biologist, married to a geophysicist, I have to back eric.suh and others, and say that part of it has to do with size of the field. I am very aware of and very jealous of the pre-print culture. BUT biology is so big, and one by-product of that is the offshoot culture of journals, as noted by other posters. My geophysicist husband goes to AGU and presents everything there. Period. Most everything he does goes to...one of 2 journals. Most of what I do could be presented at one of 6 or 8 meetings, and published, arguably, in one of about 30 journals. A number of those are for-profit journals published by elsevier, etc., and while I'd love to avoid them as much as the next guy, I am pre-tenure, so my options are limited, especially after one of the (insanely hard to publish in, in my sub-discipline at least) society journals rejects it. And those for-profit journals are much pricklier about copyright.
The second thing I'll point out, from a behavior perspective, is that the smaller a community is, the more direct personal relationships individuals make, and the more trusting and collaborative the community is. It works in animal societies with apparent altruism (which isn't really, it's just, "You scratch my back now, and I'll be good for the promise to scratch yours later") and I think it works in these professional situations too. I have met in passing a number of people whose work is most closely related to mine, but an awful lot of them I've never laid eyes on, much less built a trusting back-and-forth relationship with. And that's true for almost all of us in biology; there's no culture that reinforces the trust, to nearly the same degree. It doesn't have to be because we're insane, it's simply a function of scale. I am not building trust and collegiality with 20 or 100, but rather with an order or magnitude, or two, more. (If this number seems like an exaggeration, consider that my closest colleagues are those who work on my species or those that are closely related, but also anyone who works on similar questions theoretically, in almost *any* species. There is simply no analogous situation in physics which rapidly balloons the number of colleagues whose work can be considered closely-related.) Since our community is so large, and our connections so much more diffuse, we are simply not knit together with the personal give-and-take trust, nor is it reinforced the same way by others around us.
Molecular biologists might be light-one-another-on-fire insane (I am not one of them, and admit to having my doubts at times about mol. biologist collaborators), BUT I would argue that is more likely to be an effect of the scale of the field and resultant lack of close collegiality, rather than the cause of a breakdown of collegiality. And personally, I have never smelled a burning lab coat in my building....