For some months now I have known of a very promising initiative that until recently I have been asked not to publicize too widely, because the people in charge of it did not have a good estimate for when it would actually come to fruition. But now those who know about it have been given the green light. The short version of what I want to say in this post is that a platform is to be created that will make it very easy to set up arXiv overlay journals.
What is an arXiv overlay journal? It is just like an electronic journal, except that instead of a website with lots of carefully formatted articles, all you get is a list of links to preprints on the arXiv. The idea is that the parts of the publication process that academics do voluntarily — editing and refereeing — are just as they are for traditional journals, and we do without the parts that cost money, such as copy-editing and typesetting.
The organization setting up this platform is called the Episciences Project, and they are referring to the journals as epijournals, which I’ll do here, though epijournals will probably not use the word “epijournal” in their titles (since they will want to make clear that the stamp of quality that they confer is every bit as legitimate as the stamp of quality conferred by a traditional journal). They aim to make the software good enough that the administrative burden on editorial boards is no greater than it is for a traditional journal. If they succeed in that aim, then it should be possible for epijournals to be “Diamond” open access — free to read and free to publish. Certainly the intention is that there should be no charges of any kind, with the costs of maintaining the site met, if I understand correctly, by an organization called Centre pour la Communication Scientifique Directe (CCSD) in collaboration with the Institut Fourier at Grenoble University.
One possibility being discussed, which I am very much in favour of, is each accepted article having not just a link to the arXiv but also a web page for (non-anonymous) comments and reviews. For example, the editor who accepts an article might wish to write a paragraph or two about why the article is interesting, a reader who spots a minor error might write explaining the error and how it can be fixed (if it can), and an expert in the area might write a review that could be very useful to hiring committees.
This may even go further, with comment pages being set up for other preprints and journal articles — not just the ones that have appeared in epijournals.
Apparently, the plan is for the whole thing to start this April. Because I have known about the project for some time, I have quietly sounded out a few people in additive combinatorics, and it seems that there is enough enthusiasm that we will be able to start an epijournal broadly in that area (with a title that is not yet decided, but that will definitely not be “The Epijournal of Additive Combinatorics”). I am also on a committee (actually, they call it an Epicommittee) that is discussing some of the details of what the platform should be like — any comments you might have will be read with interest. [Added later: now that he has said so on Google+, I feel I can add that Terence Tao is also on the Epicommittee, so he has joined the good guys too.]
One question that some people might have is why, when there are a number of initiatives out there, this one should be regarded as particularly promising and worth supporting. I don’t know enough to give a detailed answer to that, but my impression is that this initiative has significant institutional back-up, including funding, that makes it more likely to succeed. Also, it is being designed for mathematicians and with the needs of mathematicians very much in mind, though it may later expand into other subjects.
April is very soon, but I hope people reading this, especially people who are critical of FoM and would rather move straight to a more radically different publication model, will give serious thought to setting up epijournals or encouraging others to do so. Another possibility envisaged by the people running the project is that some existing journals might like to convert to epijournals, which would certainly be interesting if it happened. And finally, if and when people do start to set up epijournals, please support them: if an epijournal gets plenty of good papers, then it will be much easier for it to establish the kind of reputation that will impress hiring committees (though I hope that if post-publication comments and reviews take off, they will be seen to provide more useful information than what can be deduced from which journal a paper gets into).
The Episciences project will soon be releasing a statement about the project. When it has done so, I’ll provide the link here.
I’ve been slightly vague about who the people behind this project are, which is because I am not 100% sure. However, the initial approach came from Jean-Pierre Demailly, Ariane Rolland and Benoît Kloeckner and subsequent emails have come from Jean-Pierre Demailly, so I think it’s them — my uncertainty is over whether there are other people I should be mentioning too. If I discover that there are, then I’ll add their names.
Added later: Benoît Kloeckner makes the following comment below.
I can clarify a bit the “epi-team” composition. Jean-Pierre Demailly tried to launch a similar project some years ago, but it had much less institutional support and did not work out. More recently, Ariane Rolland heard about this tentative and, having contact at CCSD, made them meet with Jean-Pierre. That’s the real beginning of the episciences project, which I joined a bit later. The names you should add are the people involved in the CCSD: Christine Berthaud, head of CCSD, Laurent Capelli who is coding the software right now, and Agnès Magron who is working on the communication with Ariane.
January 16, 2013 at 6:18 pm |
Congratulations, let’s hope that it will work (however I don’t understand the secrecy behind the idea). For some time I try to push an idea which emerged from several discussions, described here Peer-review turned on its head has market value (also see Peer-review is Cinderella’s lost shoe ) with very valuable contributions from readers, showing that the model may be viable, as a sort of relative of the pico-publication idea.
January 16, 2013 at 6:43 pm |
[...] very recent post of Gowers “Why I’ve also joined the good guys” is good news! It is about a platform for “epijournals”, or in common (broken, in [...]
January 16, 2013 at 7:04 pm |
If such comment pages exists I think it’s important that:
1) Epijournals have the option of choosing not to allow such comment pages when setting up a journal. That way distaste for such comment pages won’t sink the whole project.
2) Such comment pages are moderated. In particular, no comments are publicly posted until they have at the very least been approved by the editor who handled that paper. (I also think the authors should have the chance to reply privately to the editor.)
I think this second point is completely necessary, but I also worry that the extra time commitment for editors might make it unpopular.
January 16, 2013 at 7:56 pm
I agree on both counts. I also think that there should be a general policy that negative comments are not allowed (apart from factual comments such as, “The following appears to be a counterexample to Lemma 2.1″).
January 16, 2013 at 8:28 pm
Actually, I’m not sure whether I agree with the first point after all. If the site becomes somewhere where anybody can post a (moderated) review of any paper, then publishing in a traditional journal, or even just posting it on the arXiv, won’t stop other people from reviewing your paper. So I think I would modify 1 to say that epijournals can choose whether to have a policy of always providing reviews (perhaps written by the editors) or just leaving the comment page to the whim of the mathematical community at large. In the second case, the epijournal would be just like a traditional journal.
January 16, 2013 at 8:40 pm
Wikipedia is a good example that it is possible to have quality results without over-regulating.
January 16, 2013 at 11:10 pm
For what it’s worth, and from my non-mathematician perspective …
First, I think epijournals are awesome, and I am really excited about this initiative (even though, as usual, we biologists are trailing years behind you).
Second, I think that having the commenting facility is absolutely crucial for making these epijournals live venues rather than just dead lists.
Third, I disagree that all comments should be moderated: at least, this should be left to the choice of the individual journals. In general, moderation destroys interactivity and prevents real discussions from getting up and running, and that is a real loss.
January 16, 2013 at 11:39 pm
@Mike Taylor, are you referring to pre-moderation or post-moderation? I think if there are clear policies about what kinds of comments are acceptable, and if comments are removed or edited when they are found to be in violation of those policies (rather than having to wait for moderation before they appear), then they shouldn’t kill interactivity.
January 16, 2013 at 11:44 pm
Agreed, post-moderation (i.e. removal of unacceptable comments) does not impede discussion. Much better than pre-moderation. Personally I like the approach of allowing all comments except spam and extreme personal abuse, but I can easily see that some journals will want to do things differently.
January 16, 2013 at 7:16 pm |
Congrats to the Epijournals team!
Scholastica (www.scholasticahq.com) makes it simple to create arXiv overlay journals too! There’s a video that serves as an example here: http://bit.ly/yuD2G2. We developed a Ruby gem to interact with arXiv as well that can be found in the blog post.
January 16, 2013 at 7:57 pm |
This looks like a brilliant solution, and not just for the disciplines covered by the arXiv – although having a single repository clearly makes the thing easier to manage. Subject to this repository point, the solution could equally be extended to other disciplines, including the humanities, where there are just as many arguments about how best to approach open access, and perhaps greater worries than in mathematics and the natural sciences about the willingness of governments to come up with funding.
You say that the title “The Epijournal of Additive Combinatorics” will not be used. This might be because “The” would be inappropriate. It is perfectly possible that more than one epijournal will be created, even for a fairly narrow subdiscipline – although I suppose that once you get down to some fairly modest number of practitioners, and they all know one another, they will realize that this would be silly and would not allow it to happen.
Multiple epijournals for a subdiscipline might be useful, if the different epijournals had different approaches to the acceptance of comments. That would allow for experimentation, and the improvement of approaches as the epijournals learnt from one another. On the other hand, it would become more trouble to track down all the recently published papers on a given topic, and commentators would create more work for moderators by leaving the same comments in several places.
Another consideration is that of what might happen if several epijournals included links to a single paper. One might get a wider range of comments on that paper than if only one epijournal linked to it. On the other hand, it would take longer to track down all comments on a paper.
Perhaps the best thing to do is to have one (and only one) epiepijournal for each subdiscipline, or perhaps one just needs an intelligent search engine, thereby separating the business of finding papers from the business of evaluation for quality.
January 16, 2013 at 8:03 pm
Whoops, my comment about creating more work for moderators belongs to the penultimate paragraph, about several epijournals linking to a single paper, not to the antepenultimate paragraph, where I put it.
January 16, 2013 at 8:46 pm |
great idea
January 16, 2013 at 9:14 pm |
Great! Let the epiphenomenon begin!
January 16, 2013 at 9:56 pm |
Professor Gowers, what do you think of this
http://mathbabe.org/2012/12/03/diophantus-and-the-math-arxiv/
?
January 16, 2013 at 10:49 pm |
“Epijournal” is an oddly awkward phrase to see in English. The juxtaposition of the i and the j looks Dutch.
Anyway, this sounds great!
January 16, 2013 at 11:15 pm |
If Epijournals are worried about authors not liking comment pages, they could allow authors to opt out of comment pages for their article.
January 16, 2013 at 11:45 pm |
This is very like the idea I’ve been blogging about for some time for astrophysics. See this latest post
http://telescoper.wordpress.com/2013/01/14/aaron-swartz-and-open-access/
and subsequent comments for a discussion.
January 17, 2013 at 2:07 am |
An important question here is whether it is possible to move already existing journals to the new platform. I imagine that if the new platform turns out to be functioning smoothly, editorial boards of traditional journals might want to “liberate” their journals and move all their operation there. This would raise an obvious continuity issue (the publisher won’t be delighted to see this sort of “betrayal”, so presumably the editorial board would have to assume a new name for the journal etc.), but the benefits would be so huge that it’s hard to imagine that the editors are not entertaining this possibility. Can anyone closer to people running respectable journals comment on this possibility?
January 17, 2013 at 8:35 am
I’ve often wondered about this: what do publishers own of journals? Does Elsevier own the name Cretaceous Research? If the editorial board of that journal all decided to move from Elsevier to an epijournal, what sanction would Elsevier have to prevent or impede that move?
January 17, 2013 at 9:03 am
Mike Taylor- this has happened once before to a mathematics journal Elsevier owned, at the instigation of Donald Knuth. In brief, Elsevier own the title, but not the board. The board resigned en masse and set up a competing journal with a different name.
January 17, 2013 at 2:14 am |
Interesting idea! But it is important to thing about it well, not to end up as ghost town, like http://www.scirate.com/.
Personally, I see such thing as a traditional review process plus:
- ability for others to comment globally, or parts;
Not to end in a mess, it will be important to distinguish:
- general impressions, comments on general value,
- “issues” (like in software development), e.g. “[bug]: in (5) there should be ‘-x’ instead of ‘x’”, “[notation]: using X for set and Y for its element looks misleading”, …
- cite or reference recommendations (this one hand side may be useful, but on the other also – bait people wanting to overly promote their work or line of research)
In general, some ideas from software development may be worth adopting (like version control (Git/Mercurial) and bug and issue tracking tools (like http://www.atlassian.com/software/jira/overview)).
January 17, 2013 at 2:56 am |
Hopefully everyone here is aware of http://arxaliv.org. The site has been up for a while and is a very capable “overlay” platform.
January 17, 2013 at 3:34 am |
This is a great initiative. For biologists, similar initiatives of repute are taking shape. For life sciences, http://www.elifesciences.org/ is a great new venue where you can publish not only your paper but all the crucial data. Nature setup a preprint server (http://precedings.nature.com/), arguably so that their traditional journals can become overlays. But, it was shutdown. Any one has an insight as to why nature decided to pull the plug on precedings? – surely they have enough money and manpower to run the server.
January 17, 2013 at 4:00 am |
http://f1000research.com/ is a closely related approach that i also think is quite promising. it is geared more for biology than math.
January 17, 2013 at 5:23 am |
That’s awesome news. I remember rooting for such projects several times. I’ll look at the link in comments here.
Also I think it’s about the first anniversary of your statement against Elsevier. Happy anniversary.
January 17, 2013 at 7:02 am |
[...] of Diamond OA (as mentioned in Tim Gowers very interesting “Why I’ve also joined the good guys“) I suggest that a better and inspiring name for this yet mysterious idea if epijournals [...]
January 17, 2013 at 8:39 am |
I can clarify a bit the “epi-team” composition. Jean-Pierre Demailly tried to launch a similar project some years ago, but it had much less institutional support and did not work out. More recently, Ariane Rolland heard about this tentative and, having contact at CCSD, made them meet with Jean-Pierre. That’s the real beginning of the episciences project, which I joined a bit later. The names you should add are the people involved in the CCSD: Christine Berthaud, head of CCSD, Laurent Capelli who is coding the software right now, and Agnès Magron who is working on the communication with Ariane.
January 17, 2013 at 8:58 am |
What I find rather elegant about this approach is that existing journals could be “cloned” — that is, comment templates could be applied to the arXiv versions of their existing articles. Potentially, since most articles are in arXiv anyway, any journal “liberation” might be made to appear almost seamless, at least within the epijournal system.
January 17, 2013 at 9:05 am
like!
January 17, 2013 at 12:27 pm
Yes, it’s a good idea. And then with CC BY license (http://creativecommons.org/licenses/by/3.0/).
It’s important to remember that arXiv by itself allow to use CC license, but a default practice is to avoid to (because usually it may conflict with a journal’s policy).
This would be good for opening science, with benefits for authors and society.
January 17, 2013 at 12:31 pm
BTW., one terrible mistake in arXiv is that (unless I am being very stupid and missing something) the pages for the articles don’t say what licence they are provided under. For example, my own article at http://arxiv.org/abs/1209.5439 was deposited there under CC BY but I can’t see any indication of this.
January 17, 2013 at 8:59 am |
The Emergency Physicians Journal (www.epijournal.com/) will be wondering why they are getting so many hits since yesterday.
January 17, 2013 at 9:13 am |
Reblogged this on In the Dark and commented:
Again, no time to post properly today but here’s another variation on the theme of Open Access. The idea described in this post sounds very familiar, actually…
January 17, 2013 at 10:10 am |
This idea has some relationship to a debate I am engaged in right now on publishing an academic journal as a blog, an initiative already represented by RECONFIGURATIONS: A Journal for Poetics & Poetry / Literature & Culture :http://reconfigurations.blogspot.co.uk
I wonder if anyone, including Gowers, would like to comment on using a blogging platform in publishing an academic journal.
January 17, 2013 at 10:46 am
Annotum is a WordPress theme for running academic journals. It sounds like it ought to be rather low-rent and fragile, but apparently not: the Public Library Of Science is using it as the platform for PLOS Currents, successfully. So it can be done.
BTW., I notice that Annotum is one of the themes freely available in the no-cost hosted WordPress-based blogs at wordpress.com. So anyone who’s interested can quickly trial their own Annotum-based journal at no financial cost and for very little investment of time.
January 17, 2013 at 11:24 am |
This is an exciting initiative and I hope it really takes off.
I have a quibble with using the term “diamond open access”. You have defined this to mean free for both author and reader, but open access is also about access being open to everyone. The word “diamond” makes to sound like the most open possible but arXiv is not open to everyone for submission.
The most open journal possible would also accept submissions from anyone. In fact most journals do accept submissions from anyone so this will be quite a significant restraint for the epijournals and should be reflected in the terminology.
January 17, 2013 at 11:31 am
Actually, I am not fond of the term “Diamond” either, but for a different reason: it implies something fundamentally different from Gold; whereas the rival term Platinum OA indicates “like Gold but even better”.
January 17, 2013 at 11:50 am
Interestingly, that’s exactly how I interpret “Diamond”, because a Diamond anniversary is like a Golden anniversary but even better.
January 17, 2013 at 11:37 am
Gold, diamonds and platinum are bad names for OA, because they show that the respective publication model is based on the old idea of SCARCITY OF GOODS. Instead, I propose AZ open access (open access from A to Z) as a possible name, see the ping-back which appeared earlier in the comments (is also a tribute to Aaron SwartZ).
January 17, 2013 at 12:00 pm
“AZ Open Access” is not a bad term, but I think Gold and Green are far too well-entrenched now to be supplanted.
I don’t have strong feelings about Platinum being a better term that Diamond; but I do wish that we as a community could agree on one or other of the terms rather than having these synonyms floating around. I fear that we’re eventually going to end up calling it “Platinum/Diamond” … at which point we’ll argue about whether “Diamond/Platinum” would be better!
January 17, 2013 at 12:19 pm
I don’t see any problem with using metals and gems for the terminology. This is common in many walks of life from weddings to credit cards so everyone will appreciate it. If diamond is now set to mean free access and submissions I suggest that platinum should mean free access and submission open to anyone without restraint. Platinum comes after diamond on the anniversary scale so this is fitting.
January 17, 2013 at 12:33 pm
For me using things with valuable gems or metals is a _bad_ analogy, as it suggest _high_ prices, not free, easily accessible and reusable material.
However, “not all that glitters is gold” – some “open access” means the same academic journal scam, just they charge authors, instead of readers.
It’s important to put the emphasis on making publishing (when it comes to copyright and non-profit standards) as possible (so, say, as for arXiv, but this time additionally CC BY).
A text of me and my friend on that matter, and others:
http://offtopicarium.wikidot.com/v1:open-science-2-0
January 17, 2013 at 12:34 pm
Aside from the nomenclature: Who is prohibited from submitting papers to arXiv? I thought anyone could, after registering.
January 17, 2013 at 12:38 pm
Wikipedia dixit: Artificial Scarcity.
January 17, 2013 at 12:47 pm
Andrew King asks: “Who is prohibited from submitting papers to arXiv? I thought anyone could, after registering.”
Almost. But before your registration is accepted, it has to be sponsored by someone who is already registered. For a legitimate researcher, finding someone to do that should not be hard. The idea is just to keep spammers and cranks out, I think, not to raise a barrier to researchers.
January 17, 2013 at 12:59 pm
People who cannot submit to arXiv submit to viXra (of which I am an administrator), The intake is growing exponentially. There will be many papers there that you will no doubt consider as crank, but many others have been accepted in peer review journals. If future journals become tied to arXiv they will find it much harder to publish. I don’t think it will be good for science if the ability to submit to journals becomes tied to the ability to make friendly contacts with established researchers in academia, not to mention the whim of the arXiv administrators who can still reject papers even when they are endorsed.
January 17, 2013 at 1:08 pm
There is in fact a technical point that I didn’t make in my post above, which is that the Episciences platform will be based on the HAL archive, but will allow links to other archives. So I think it is not part of the definition of an epijournal that it should be an arXiv overlay — that’s just what I imagine most of them would in practice be.
January 17, 2013 at 1:12 pm
That sounds more promising, but it would still be good to have a distinction in the terminology so that when the epijournals appear we can classify them according to whether or not there are any restrictions on who can submit.
January 17, 2013 at 12:36 pm |
Also, why we (i.e. scientists) cannot base research on a platform for open source, collaboration-friendly stuff?
Something like “GitHub for science”? (https://github.com/, see also: https://bitbucket.org/)
I have an impression that when it comes to open culture, scientist are _way_ behind programmers.
January 17, 2013 at 12:51 pm
That’s the future, clearly. It already proved its viability, from the economic point as well. But (a) nobody has found a viral idea about how to do it (yet), (b) there are obvious (but vague) interests in delaying the announcement that the patient is dead, (c) you can change anything if you awake people’s imagination, see for example the conjectures concerning why MOOC has more impetus than OA here: MOOCs teach OA a lesson by Eric Van de Velde.
January 17, 2013 at 4:06 pm |
I think that this is a great idea. Congratulations and good luck!
January 17, 2013 at 6:12 pm |
Related to this, I had an interesting discussion the other day on the possibility of a journal where you submit your research project BEFORE you collect data. As long as you do what you said you were going to do, the journal will publish your paper. This way, negative results won’t be buried and people will be evaluated on their IDEAS, not their results (which they have no control over).
January 17, 2013 at 6:55 pm
Well, in some fields (e.g. mathematics, theoretical physics) you are not guaranteed to get results at all.
IMHO something other is important:
- to credit scientists for publishing negative data,
- to credit scientists for repeating experiments
(seriously, now the system is flawed, in principle science _relies_ on repeatability, but now one get little to no credit for repeating someone’s else experiment),
- to publish continuously, in chunks smaller than a publication, e.g. like open source projects on GitHub (so paper only as “summary and final version of”, not the sole citable and recognizable way of communicating progress (or lack of it)); related – open notebooks.
January 17, 2013 at 8:41 pm
Helda’s suggestion that we should do something to ensure that non-results get published sounds like a very good one, especially given that in the age of the Internet, that need not crowd out reports of interesting results. One obvious area in which it matters is drug trials. Ben Goldacre has been very strong on this, telling us how many trials that don’t show that a drug is any good simply get ignored. Another nice example comes from Richard Feynman’s essay Cargo Cult Science. He talks about a Mr Young (I don’t know who this was), who in 1937 did experiments on rats running along corridors and going to particular doors. He didn’t discover anything interesting about how rats learn, so his work was apparently ignored. But he did discover that you have to place the corridors on sand, otherwise the rats respond to the different sounds that are made as they run along different stretches of corridor, and that is (according to Feynman) a really important thing to know when designing that kind of experiment. Thus the fact of a non-result can carry an important lesson.
January 18, 2013 at 12:43 am
Such an idea is also current in medical research (see Ben Goldacre)
January 17, 2013 at 6:12 pm |
OK, I got it! You have served us a specimen of second degree british dry humour and you performed an impersonation of Anakin Skywalker. Empire strikes back followed by the Return of the Jedi, in just two blog posts.
January 17, 2013 at 7:15 pm |
Here’s an example of arXiv overlay journal: http://about.eptcs.org/