An article by Steven Bachrach in the Journal of Cheminformatics has an excellent disscusion on Open Access journals. He notes that one of the problems with current scientific publishing is the plethora of journals. He also points out the huge amount of publications being generated. He succinctly states it as
We simply publish way too much
No one can keep up with a literature like this. It is time for the scientific community to rethink the role of publication. Should every little idea, every minor work receive the same treatment as the great discovery?
Most articles barely get cited – many never get cited; they are simply maintained within this ever-growing collective of scientific work, with the chaff growing at a rate exceeding the production of the wheat
He then proposes a two pronged approach in which 1) we reduce the number of full service journals, restricting it to just the top ranked journals and 2) the remaining journals publishing the “remaining 80% of the scientific literature” be replaced with institutional respositories.
I think this is a very sensible approach. But I see one major question that needs to be considered, before this approach might take off. And that is, why do we publish so much? As opposed to, say, putting papers up on web pages.
I think it wouldn’t be far from the mark, to say that a major push for publishing is credit. From an idealistic point of view, one can say that scientific results need to be disseminated. Therefore, by publishing in journals which everyone reads, we achieve this result. But the fact is, publications are academic currency – they help academics get tenure and get money. Ideally, committees and funders would be able to differentiate between the really significant publications and the “chaff” and maybe even penalize the latter. That doesn’t seem to be the case – hence publish whatever you can, whenever you can.
But, if this is the current atmosphere, what are the chances of success for Bachrachs proposal? Clearly his first point would be supported – publications in Tier 1 journals would be a bundle of currency. But that is the case, even now. The problem is that the bulk of scientific output would end up in repositories. Personally, that would be fine by me – easy access to papers and data, ability to interact with authors and so on. But, would anything and everything by an institutions scientists be deposited in such a repository? If so, how does one differentiate between the good and the bad? Maybe good papers would get a lot of comments and feedback? Bachrach already notes that participation in journal supported forums is generally low. Why would participation in forums maintained by institutional repositories be any greater?
Personally, I’d love to be in a universe where I could simply write up a study and publish it on the web along with code and data and then move on. Unfortunately, in this universe, resources are limited and I must compete. Hence some form of external validation (or at least the appearance thereof) comes into play. And as a result I have to play the publishing game.
Bachrach mentions that Open Notebook Science as a publishing model is far too radical for near term adoption. I agree with this. But it seems that pushing 80% of the literature into institutional repositories also requires a fundamental (even radical?) rethinking of how academia rewards acheivement.
Update: As noted by Egon, the world is a better place if one links to DOI’s rather than PDF’s. In that spirit I updated the link to the Bachrach paper to point to the DOI.