57

Many of you will know that, whether per subscription fees or via open access models, there is a ridiculously absurd amount of money flowing to many big (or medium-size) publishers for getting our research outputs (i.e. papers, etc.) published in high-quality venues in the field.

The problem: Content and peer reviews (I think, the most important artifacts from a scientific point of view) are delivered largely for free and voluntarily. In some fields, such as math, physics, CS, EE, and related, even proper camera-ready type setting (e.g. via LaTeX) is also very often done by the researchers themselves. Moreover, in fact, the downstream (i.e. after-acceptance) editorial process in traditional publishing companies is nowadays (for a number of reasons I don't want to elaborate on here) more likely to introduce faults through the various conversions and process stages that are involved. I'm speaking not only of my own experience but have even seen publications that have been corrected using post-prints, personal copies of authors, etc.

So, my question again: Can't we just extend well-known and widely-used archival platforms such as arXiv, biorXiv, zenodo.org, etc. with a light-weight infrastructure to perform peer review (e.g. via an integration with EasyChair or the like and building up editorial teams) and transfer the whole peer review life-cycle from established publishers (with over the years questionably evolved business models) to such platforms? A tiny fraction of the (mostly tax payers') money currently flowing to such publishers would be invested in running the mentioned infrastructure. My naive assumption is that no relevant quality loss (if not even a gain in quality) is to be expected in certain fields of science after an appropriate ramp-up phase.

The only relevant issue, I can see for now, is the issue of a lacking reputation or trustworthiness of such a platform as a publishing venue where, I suppose, it will be a matter of time for this issue to be overcome.

One of the reasons why I am asking is because I think that the mentioned amounts of money can be better invested in science for the public than it is now. A question (totally obvious to me) that I've pondered over for a long time but was afraid to ask. Thanks for your thoughts on these, I apologise, probably too progressive and naive ideas.

In response to some of the comments: Archival has very little to do with publishing. This question is about modern peer-reviewed publishing. Please, also note that with this question I very much value professional editorial work and am looking for solutions in support of EiCs, editorial committees, and reviewers. By no means is this question to be understood in any way as a critique of any of the work done by EiCs, editorial committees, and peer reviewers, not at all. If so then it was definitely not my intention.

31
  • 9
    There is a lot of administrative work associated with peer review for a large journal. It's one thing to be program chair for a large conference once (or once every several years). It's another thing to handle 10 times that volume every year on an ongoing basis. It really is a situation requiring a paid staff person. Aug 30 at 19:38
  • 20
    "even proper camera-ready type setting (e.g. via LaTeX) is also very often done by the researchers themselves." This is common in math, physics (?) CS and related fields. Over all of "science" it's not "very often." The vast majority of academics do not know how to use LaTeX. Aug 30 at 20:46
  • 24
    @jamesqf There's a trend on this site that people assume everyone knows LaTeX because this site's demography is fed by SO, Math.SE, etc. I can assure you, even if you take "in STEM fields" (and not the humanities, who have the same problem, so idk why we're excluding them) the vast majority do not know LaTeX and haven't heard of it. I love LaTeX, but I come from a programming-ish background. It is not easier than using Word ... If the field doesn't use LaTeX, where are they going to get URAs who know it? Aug 30 at 22:03
  • 19
    People suggest LaTeX all the time, but frankly, even though I like it, I don't want to spend all my time formatting my articles instead of analyzing and doing research. The answer to "Why don't we have scientists typeset their own articles?" is "Most scientists don't know any typesetting software." Aug 30 at 22:05
  • 16
    -1 because this question keeps getting asked by people who think papers are really interesting and they should be paid more to write, edit, and review them even though the average paper is read by no more than 10 people. If anyone thinks they can do it better, try setting up an arXiv overlay journal and maintain it till it becomes self-sustaining. Don't point at the arXiv overlay journals that already exist; that's like pointing at Messi and saying it's easy to earn £500,000/week playing soccer. Set up your own journal.
    – Allure
    Aug 31 at 5:14
62

There are Diamond open access journals (which are free to both author and reader) run as arXiv overlay journals. An example in my area is Logical Methods in Computer Science, which is generally well-regarded in the field (and indexed whereever it needs to be).

Besides the already existing journals, there also seem to be plans for additional journals of this type to better cover the various disciplines. However, founding a new journal takes a lot of effort from a team of respected (and thus very busy) academics. Early on, it will not be on all indices and thus less attractive to the unfortunate amongst us whose administrations force these issues. Taking an established journal away from a publisher to move to this framework is, even where possible, a radical act and usually pure inertia will stand in its way.

Nevertheless, my sentiment is that Diamond open access (with arXiv providing the long-term storage) is the way to go; and I believe that this is a rather popular perspective.

Despite being lightweight, an arXiv overlay journal still has some monetary costs. They seem to be enough to be an issue, but also to pale in comparison to journal subscriptions or open access charges by commercial publishers. My best guess why institutions/governments aren't more eager to support them is that traditional publishers have a lot of lobbying power, while arguing pro-Diamond open access is more of a hobby.

7
  • 1
    Thanks, Arno. Very good points, I haven't been aware of Diamond OA. So, there is also EPTCS (info.eptcs.org), which is well integrated with arXiv for peer-reviewed workshops and conferences run e.g. using EasyChair. EPTCS would indeed also follow that model while not being a journal, though. But your example and mine seem to be among the currently still very few exceptions. There are costs, indeed, but the question is how many orders of magnitude less ;)
    – Mario
    Aug 30 at 10:51
  • 3
    Another example: scipost.org Aug 30 at 14:48
  • Another example: The Open Journal of Astrophysics.
    – Vectornaut
    Aug 31 at 1:00
  • 20
    @And You seem to be under some kind of misapprehension. There is no reason for scare quotes here. The refereeing process works in exactly the same way as any other journal. Whether or not a journal accepts submissions by editors has nothing to do with whether it is Diamond OA/arXiv overlay/published by a society/published by a commercial publisher/printed/online only/etc.
    – Arno
    Aug 31 at 13:18
  • 2
    It is not at all clear that prepublication peer review adds value to the process. Perhaps it mainly serves as an entry barrier without which journal publishing would not be profitable. This is what the open access model suggests.
    – Dan Fox
    Sep 1 at 17:26
52

tl;dr: we could, but we won't, because we're in a Nash equilibrium.

It would be the better outcome for all involved parties (except the traditional publishers) if we would find ourselves in the situation that you sketch. However, if a few individual parties make a move towards the optimum, while others all stick with the current strategy, the few movers lose. Hence, the change is unlikely to happen.


Suppose that you're a PhD student, who seeks a decent venue for their next publication. You can either go for arXiv, where science is free, or you can go for a traditional venue, where it is not. In a few years, however, you might want to apply for a tenure-track position at some top university, and the competition will be fierce. Most of your competitors will still go for the traditional publication venues, even if you move on. Current practice at hiring committees is to look for candidates with many publications at the top venues, which is typically measured by traditional journals and conferences. Why would you run the risk of having your best work not recognized by future hiring committees? Surely, you shouldn't run the risk of kneecapping your own future employability, so submission to traditional venues it is.

Suppose that you're a tenure-tracker, who seeks tenure. You want to apply for grants that bolster your tenure application. To get those grants, it helps if you can show that the community considers you an expert in the field. How do you establish your name as an expert in the field? For instance, by serving on the program committees of the top conferences, or becoming a member of the editorial board of the top journals in your field. You may consider reviewing for arXiv, but there is only one tenured spot at your university; the four other tenure-trackers with whom you must compete will all review for the top journals/conferences at the traditional publishers. Surely, you shouldn't run the risk of not getting tenure, so reviewing for traditional venues it is.

Suppose that you're tenured. This is the moment! You have freedom, so now you can finally be the change you want to see in the world! But hold on, you're writing research papers together with a new batch of PhD students who all would like to be tenure-trackers someday, and you apply for funding jointly with tenure-trackers who all would like to be tenured today. Will you run the risk of kneecapping your direct colleagues for the sake of your principles?


In all these examples, it's not impossible to advance in a scientific career while moving along the arXiv path. In all these examples, it's probably over-the-top to characterize choosing the arXiv path as kneecapping yourself and your coworkers; any real damage may be subtle and small. The point is, though, that the academic career ladder is murderously competitive, and every little bit just might help edge out your competitors. It stands to reason that not many people would allow themselves to risk any disadvantages in this fierce competition, and hence I see no path to get from the current situation to the situation you suggest.

14
  • 5
    Valid analysis. Imporant point: "You have freedom, so now you can finally be the change you want to see in the world!" Replace [You] by [EiC of a top journal] (typically tenured) who takes their freedom to move their editorial committee (overnight) to arXiv (with a little bit of preparation), advertises for it, and within months everyone in the field WANTS to publish there, and hiring committees will quickly notice and value that. Not far-fetched, I'd say, but, again, your analysis is valid, though being pessimistic with respect to the inertia of the system. Sure, EiCs could be bribed to stay.
    – Mario
    Aug 30 at 15:02
  • 4
    "murderously competitive" is certainly true. As you mentioned that from a game-theoretic viewpoint, I wonder about the extent to which this is actually necessary and even caused/reinforced by the publishing industry...
    – Mario
    Aug 30 at 15:05
  • 12
    Very good analysis, in my opinion. This dynamic is what strongly inhibits many changes. And, yes, surely the for-profit publishing industry finds no reason to help jolt us out of this unfortunate equilibrium. Aug 30 at 19:35
  • 11
    @Mario the scenario you sketch of the EiC of a top journal going rogue is a potentially plausible one, and quite similar to the founding of the Journal of Machine Learning Research. This happened in 2001, when 40 members of the editorial board of Springer's journal "Machine Learning" resigned (letter can be seen here: sigir.org/files/forum/F2001/sigirFall01Letters.html ) in protest against the hefty fees Springer charged; they founded JMLR as an open-access alternative. So it's not impossible, but such change is rare and slow. Aug 30 at 19:37
  • 2
    Thanks, Wetenshaap, for another nice example, I've even heard of this journal, though not being active in that field. I very much believe such change is possible within the medium term, it just needs quite a few more nudges and, of course, platforms like arXiv need to provide further features/integrations with other tools. No change without effort and costs, for sure. But you can do hell of a lot with 100s of Mios of subscription fees and APCs.
    – Mario
    Aug 30 at 19:58
5

If you restrict attention to areas like pure math, where papers are mainly text with equations, then Arvix overlay journals can work really well. However, different areas of science have needs that are not so compatible with the Arxiv. Of course, other preprint servers might work.

One problem is with supplemental files. Consider the supplemental videos and how they appear one this webpage of the Proceedings of the National Academy of Science. Nice descriptions, links to forms for copyright permission. Now consider the supplemental videos in this preprint of the arxiv. All you get is a title.

Images on the arxiv are another issue. A few of my papers I did not post on the arxiv as the images do not compress well and I did not want to fiddle with creating low-resolution images just to get the figures small enough for the arxiv. (The size limits used to be smaller.)

Recently the Arxiv has added the "papers with code" feature, but this mostly assumes hosting the code elsewhere.

I think the needs of journals are too diverse to count on overlay journals to cover too much of math and science.

4
  • 1
    Recently (actually, apparently it was a little over a year ago) arXiv increased the default size limit to 50 MB, so having too large images (or too many) should be much rarer than it used to be.
    – Anyon
    Sep 5 at 0:35
  • @Anyon Good to know about the increase. Thanks. Sep 5 at 3:12
  • @TerryLoring Absolutely, arXiv would need a bunch of further features, such as support for supplemental material (videos, code, data) and further meta-data. I like zenodo.org in this regard. And the possibility to add comments/post-pub peer reviews. I like pubpeer.org in that regard. Overall, I don't think that the required implementation of such features in arXiv would be a showstopper for the suggested change/solutions. arXiv seems to be on its way to widen its applicability for Diamond OA journals and the like.
    – Mario
    Sep 6 at 14:27
  • In a software as a service world, one could organise peer reviews via EasyChair, forward papers to arXiv (like eptcs.org), upload supplemental materials to zenodo.org, code to bitbucket, and allow post-publication discussions on pubpeer. And arXiv or someone else could provide an interface that collects and distributes the data. Funding for such a project should be acquirable.
    – Mario
    Sep 6 at 14:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.