I hold the theory that ideas are much more social than we think. When the time is ripe, they simply flourish in many minds at the same time. For example, one year ago, in this blog we published an entry discussing the possibility of a peer-review system on the ArXiv. Some hot debate ensued, and some people used very harsh language. But time was ripe, and people more influential than I have started a project, called epijournals. So far, the only source of information is Tim Gower’s blog and the comments on it. [Note added: it’s been pointed out to me that there is already an official link]

So, again, what is the problem? The problem is that basic research is funded, mainly, by public money. But the results of research are published in journals which are privately owned, and make a profit. Well, one might think that if it is a reasonable profit, and they add value to the publication, everything is fine. But the added value approaches zero asymptotically, and the profits are completely insane. Scientific journals get their stuff for free, since we scientists do not get any money for the articles. We do the research, typeset the articles, and make them camera ready. No charge. Also, we act as referees to judge the quality of other people’s papers, and very often also edit them. No charge! Yet, the journals are amazingly expensive. And sometimes they play harsh with the rights of the scientists to use their own work. E.g.: the dutch publisher Elsevier, which has become our bête noir, and against which we’re holding a boycott.

What value do the publishers give, anyway? Prestige. Their stamp of approval. If you publish an article in Nature, Physical Review Letters, or Cell, it means that it has gone through a careful selection process, it must be worth reading. But this revision process is done by… us scientists! And is done for free. So… if only the stamp matters, why not using it directly on the ArXiv?

The idea is simple: epijournals will work as regular journals, with editors and referees. The only difference will be that the journal webpage will not hold the articles themselves, only links to the ArXiv. So, you upload your paper to the ArXiv. Then, you write an email to the editor of your favourite epijournal, telling her your ArXiv number. She will send it to referees, make a decision and, if finally published, you’ll get a link in the epijournal webpage. Easy peasy.

As of now, it’s only mathematicians taking the step. It’s crucial that big guys send their papers to these epijournals, only then they will gather momentum and become prestigious. We, little people, need big publications to get tenure… so, please, please, if you have an important breakthrough in maths, consider submitting to one of them (when they start, of course).

What should the next step be? As of now, the “ranking of publications” is held by Thomson Reuters, a private company with a lot of power. For example, they compute (using a secret formula) the Libor and Euribor indices. They decide whether a journal is good enough or not to enter their Journal of Citation Reports, which charges the journals their revolutionary tax in order to stay in the list. Shouldn’t UNESCO take care of this task?

Other interesting issues related to epijournals:

  • Should we allow (signed) comments on the articles? I would like to.
  • Not only comments: I would love to chat with the authors of the papers I like. And, as an author, it would be so nice to chat with the readers. It will help reduce the feeling of irrelevance of theoretical work…
  • So, this would evolve in the direction of creating an on-line social network for scientists.
  • Is blind peer-review good enough? Why not double blind? (The referees do not know the authors either)
  • Make a two-stage refereeing system: a quick and dirty one, which would correspond to the current system, and another five years after publication, to assess what was the real impact.
  • About authorship: shouldn’t we allow for “partial” authorships? Or, more concretely: specifying what has everyone done in each paper. The current strategy of listing the authors in a certain “relevance order” is clearly not enough…

I hope more questions will come up in the comments… ;)

A personal dream: Journal of Physical Insight

Just a week after I published my post on the scientific publishing industry (#occupy_scientific_journals), the whole world seemed to explode. Tim Gowers started his personal crusade, and articles appeared even in mainstream media about how Elsevier and the strange world of scientific publishing. I was happy.

But complaining is not enough. I have had a dream for a long time: to create a scientific journal. A possible name would be “Journal of Physical Insight”, but others have been proposed by friends, such as “New Points of View in Physics”. Let me explain how it would look like.

Aim and scope. the journal would not aim at publishing original research. It would publish only original insight about known research. New ways of looking at old things. Conquering new territories is not more important than colonizing them.

Examples: revisiting old concepts using new tools, interesting conjectures, exposition of conceptual difficulties and possible ways out, more clever notations, unexpected connections between distant results… Do not misunderstand me, it would be a hard-core research journal, indexed in JCR. It would not be a teachers’ journal, although also teaching might be benefitted from it.

Publication style. I would like it to be a fully free journal, both for readers and authors. Authors would be required to typeset the paper carefully, in final form, check the references, etc. The editors would be volunteers, and they would be required to be young scientists, counting on the help of an advisory committee of senior scientists.

Special emphasis would be given to the writing style. The special aim of the journal suggests that editors and referees should encourage the authors to make a special effort to make concepts very clear. Also, evidently, to peruse the literature as deeply as possible, also outside your field: novel ideas in one field can be known concepts in another.

Peer-review process. That is one of the main novelties brought by the project. First of all, I want it to be double-blind, i.e.: the referees will not know the names of the authors or their affiliation. Also, I advocate for a two-stage peer-review process. The first one would be as quick as possible. Once the paper is published, its refereeing process would not be finished. It would start the second, community-driven process. Comments would be open for each article, and they would be collected for a reasonable amount of time, e.g. two years. It’s already time for scientific research to benefit from the 2.0 revolution! After that trial time, a second refereeing process would be carried out, to assess the impact of the work beyond its number of scitations. This second evaluation would be most beneficial to funding agencies, of course, because by then all scientists in the field would know the article.

Normally, the scientific edition procedure starts when the authors submit their finished work. Given its special scope, this journal would encourage authors to submit article proposals to the editors before embarking in the project, as it is done typically with review papers. The editorial board, if they consider the proposal interesting, will give support to the authors. This is a standard procedure in other areas, but not in science.

Of course, such a project will take a long time to bloom. It will require support from some scientific institution, although money is not an issue in this case: a few dedicated servers would be more than enough. Much more important is to convince a critical mass of colleagues, from all branches of physics, that this idea is worth trying.  Thus, I think time is ripe to ask for feedback… What are your thoughts?

(thanks to Silvia N. Santalla)


The main aim of this post is to propose a peer-review system on the ArXiv. We need a revolution in the scientific publication scheme.

1.- What is wrong?

Today I needed a scientific article for my research. My institution is not subscribed to the journal, but the publisher said “No problem, dude, just pay $33 and you can read the paper”. Seriously!?

Scientific publishing is a peculiar business model. Authors make no money from publication. Neither do referees. The typesetting of the articles is usually done by the authors themselves. Yet, the alleged cost per article is around $1.000-$10.000… Seriously!?

Work in fundamental science is usually paid by governmental funds, through taxes. And, even when the money comes from private hands, still their aim is to create knowledge and make it publicly available. But, as of now, the general public does not have free access to the results of the research they fund. Even professional scientists have frequent problems to obtain articles they need, thus making their research more difficult. This problem is getting worse with the economic crisis, and has always been a major issue in developing countries.

If authors do not make any money, why do they publish? For want of reputation and dissemination of their work. Funding agencies need some quality measurements in order to make decisions about which project to support. The accepted system, worldwide, is the number of publications and citations, and the prestige of the journals in which you publish. Journals are ranked by the JCR (journal citations report) index, which is itself… another private company (Thomson Reuters), which charges enormous amounts of money to universities and research institutes to pay for a faulty database.

Of course, some publishers are better than others. IOP and the DPG started New Journal of Physics, which is open. The problem is that publishing there is quite expensive. Other open journals can be found here.

2.- What do we want?

We want a cheap and open publication scheme. Most of the work is already done already by us.

We want a fair reputation system, which rewards high quality research, to serve as a guide for government agencies to direct their funding. And also as an internal guide to the relevant literature (too much to read, otherwise!)

3.- Ideas

The most promising point of departure is the ArXiv. It is free and open. It costs its maintainers (a board of worldwide research institutions) around $10 per article. Why not creating a peer-review system on the ArXiv? If authors so desire, they might ask for a “peer-review stamp” on their preprint. It wouldn’t be so difficult. A similar idea was already put forward by John Baez.

The peer-review process, as it stands today, is both too slow and too fast. It’s too slow because it takes months for a regular submission to see the light. By then, it is very often well known by the community, who had access to it through the ArXiv or otherwise. And it is also too fast because the referee process is not good enough to assess whether a paper will have impact or not. It takes time to know. So, why not making two “peer-review” processes? A quick-and-dirty one when the paper appears in the ArXiv. A second one, more detailed, after a few years, to evaluate its real importance.

Another nice idea would be to create an open discussion forum for each paper, where people might be able to make comments and ask questions. In the stack-exchange community style, reputation might be awarded for making questions and providing answers which the community approve. Of course, the forums need not be attached to papers only. The concept of paper as the “unit of research” may become outdated in such a structure. Papers were the natural medium for the exchange of information when the dead-tree technology was dominant… but, just like the mechanical loom, animal traction and congressmen, may be overthrown by history.