linking back to

My lab:
I'm sure nobody ending up on this obscure blog could have missed the current frenzy about a Nature news article by Declan Butler attacking PLoS. In the meatime, there has been a follow-up by Nature publishing director Timo Hannay, also with comments and a reply by Timo. I think what we see here are the labor pains of a new scientific publishing model. People realize that things are not working effectively, some would maybe even claim that the entire system is broken and needs to be replaced. A good overview can be had from Coturnix in his post, but there are also comments worth noting individually such as Pedro Beltrao, Greg Laden, Lars Jensen, Mario Pineda-Krch, DrugMonkey or Bill Hooker.
Clearly, whatever the system will be that is coming out of this, at the very minimum, it needs to get rid of the following idiosyncracies (to put it mildly) of current scientific publishing:

  1. First and foremost, it is a waste of time and effort to submit and re-submit manuscripts with the same data but different formats until it finally is published. One single, exhaustive and critical instance (with X numbers of rounds) of peer review should be enough to find out if the manuscript under consideration has any scientific value at all. Once this minimum standard has passed, the manuscript should go into whatever form of database will then exist. Of course, all the papers in this database should be freely accessible to anyone. Personally, I care mostly for the access to scientists, but of course these are included in 'anyone' blush.png
  2. The quality of an aricle is not a one-dimensional value. Quality is not assessed by a certain elite group of gatekeepers who shunned professional science for whatever reasons. Quality is not assessed by a single number. Quality is definitely not assessed by where something is published. Some measure of quality is in the eye of the beholder, some measure of quality can be more easily quantified. Therefore, any new system will incorporate the capacity for a multi-dimensional quality assessment that works for scientists, administratotrs, the media and the general public, as each of these groups have different demands and different expectations of scientific quality.
  3. UPDATE: What definitely also needs to go is the requirement of searching in at least three different search engines with different syntax and strategies to be able to find really all the relevant papers on a given topic. What we need is one single search for all papers with an elaborate search mask and the possibility to search for citations in both directions.
Maybe there are more points (probably), but these are the most important ones. At the same time, any new system needs to retain the following two traits:

  1. Peer-review (Duh!)
  2. Trained, experienced professionals who provide their best estimate of what kind of relevance new research is likely to have for all of science or any field in particular. There is too much research going on to be able to follow everything. Human sorting is still required.
There probably are more good traits in the current system worthy of keeping, but right now I can't think of any smallgrin.png
You tell me!

Conflict of interest statement: I have published in Science and PLoS One; I volunteer as academic editor for PLoS One.
Posted on Monday 07 July 2008 - 14:41:05 comment: 0

Render time: 0.0797 sec, 0.0060 of that for queries.