A peer-review system for the ArXiv?

[jrlaguna]

The scientific publication scheme is clearly faulty. Although most research in fundamental science is funded by the taxpayer, the average citizen does not have access to the results. Scientific journals are typically quite expensive, although they offer little-to-nothing added value: authoring, typesetting and reviewing is done by the scientists themselves, typically, for free. The database used for impact evaluation is also a private company (Thomson Reuters).

So... why not, as a first step, create a peer-review system on the ArXiv? There are more ideas in this link

http://physicsnapkins.wordpress.com/2012/01/16/occupy_scientific_journals/

What are your thoughts?

 

[ZapperZ]

The scientific publication scheme is clearly faulty. Although most research in fundamental science is funded by the taxpayer, the average citizen does not have access to the results. Scientific journals are typically quite expensive, although they offer little-to-nothing added value: authoring, typesetting and reviewing is done by the scientists themselves, typically, for free. The database used for impact evaluation is also a private company (Thomson Reuters).

So... why not, as a first step, create a peer-review system on the ArXiv? There are more ideas in this link

http://physicsnapkins.wordpress.com/2012/01/16/occupy_scientific_journals/

What are your thoughts?

Why is it faulty, and why is your suggestion devoid of such faults?

Note that most of the major physics journals allow for the authors to upload the published paper on the authors website, allowing for the free distribution of such papers (read Phys. Rev. policy on this, for example). So your argument that the general public do not have access to such sources is itself faulty. In fact, all one needs to do is contact one of the authors, and ASK for a copy of the paper! Try it!

So considering that free access was the ONLY argument you put forth for the journals to be faulty, and I've already stated how one can get such papers without paying anything, what fault is left?

Zz.

 

[cgk]

Why is it faulty, and why is your suggestion devoid of such faults?

Note that most of the major physics journals allow for the authors to upload the published paper on the authors website, allowing for the free distribution of such papers (read Phys. Rev. policy on this, for example). So your argument that the general public do not have access to such sources is itself faulty. In fact, all one needs to do is contact one of the authors, and ASK for a copy of the paper! Try it!

Most authors don't actually upload their papers to their websites, however, and there are several branches of science where preprint services are rarely used (e.g., the whole of chemistry, but also some subfields of physics) .

So considering that free access was the ONLY argument you put forth for the journals to be faulty, and I've already stated how one can get such papers without paying anything, what fault is left?

I don't actually agree with the OP[1], but it can hardly be denied that some journals are very expensive, and sometimes the actual value created by the publishers is diminishingly small. Of course as scientists we often do not see that directly, but when I first looked up the budget our librarian had to spend on (online-)journal subscriptions, my jaw dropped to the floor...
I would not mind the cost of the journals if they actually did their job; e.g., did proper editing for language and would at least tell the authors if some of their graphs are completely unintelligible. But as it stands, that is not common practice.


[1]: The journals do provide more: they do typesetting (a short comparison of arxiv papers with the published papers makes the difference obvious), they sometimes offer very basic editing, and, most importantly: they offer a platform for exchange.

 

[dcpo]

Universities currently spend a great deal on subscriptions to journals, with the profits generally not going to the people who do most of the work. Now that paper copies are virtually redundant the only reason to continue with this model is tradition. All that journals provide is a 'seal of approval' from the name recognition, and a focus point for the efforts of those upon whose integrity the good name of the journal is actually based (i.e. the editors and reviewers).

I see only two reasons for submitting to a peer-reviewed journal these days, the first is to spread your research beyond your circle of colleagues, and the second is the recognition that publication in 'name' journals brings (which isn't just about personal glory, if you want to keep an academic position you'd better publish in journals your department recognizes). The question is, could arXiv provide those things if it were peer-reviewed? My thoughts are no, at least initially. It strikes me it would be a tremendous effort to add on some kind review system to arXiv, and I imagine attempting to do so would encounter a great deal of opposition. I can however see Internet based journals with rigorous peer-review bypassing the publisher system gaining significantly in prominence over the next 20 or so years.

 

[jrlaguna]

Many authors do not upload free versions of their papers, to start with. Of course, you can ask for a copy from the authors as a personal favour, no doubt. Nonetheless, I get trouble getting papers quite often. In theoretical physics this is less noted, but in other fields it can be worse. Normally, we have to recourse to friends and colleagues in other universities.

Scientific institutes and universities pay huge amounts of money for their subscription to journals, and for access to the Web of Science. The crisis has made some universities cancel subscriptions, and this is a permanent problem in developing countries.

Alleged cost of a scientific journal is around $1000 to $10000 per article... For what?! The cost of the ArXiv is $10 per article [http://people.ccmr.cornell.edu/~ginsparg/blurb/pg02pr.html, slightly outdated values given by P. Ginsparg]... What added value is given by the journals? Basically, their prestige. Authors make no money, referees make no money. Typesetting is done by authors themselves... Why should universities subsidize this business model?

 

[Andy Resnick]

<snip>Scientific journals are typically quite expensive, although they offer little-to-nothing added value: authoring, typesetting and reviewing is done by the scientists themselves, typically, for free. <snip>
What are your thoughts?

Why do you think the cost associated with peer-review to be zero? A typical first-rate Journal has to process >10k manuscripts every year: 100 *per day*. How many people are required to do this? How much time is spent just on assigning reviewers and managing the process?

Your linked essay has another major flaw which it tries to ignore- the cost is either borne by the subscriber or the author. I have to pay a considerable amount of money to publish in open access Journals. Which is fine- I build that cost into my grants.

The demand for free, fair, high-quality scientific publications is unreasonable.

 

[Vanadium 50]

Typesetting is most assuredly not done by the authors themselves.

 

[jrlaguna]

@Andy Resnick: I do not advocate for free journals, that makes no sense. But most of the work is done by us, scientists, nonetheless. They should be substantially cheaper, their cost being assumed by the same funding agencies that make science possible. You are talking about the edition process: obtaining referees and managing correspondence with them. Do you think this job really explains the high costs of $1000 - $10000 <b>per article</b>!?

@Vanadium50, in theoretical physics, we do typset our own papers. We're usually required to use the journal macro system for LaTex, which we do willingly.

 

[ZapperZ]

Most authors don't actually upload their papers to their websites, however, and there are several branches of science where preprint services are rarely used (e.g., the whole of chemistry, but also some subfields of physics) .

Many authors do not upload free versions of their papers, to start with. Of course, you can ask for a copy from the authors as a personal favour, no doubt. Nonetheless, I get trouble getting papers quite often. In theoretical physics this is less noted, but in other fields it can be worse. Normally, we have to recourse to friends and colleagues in other universities.

But this is besides the point! The fact here is that authors ARE allowed by most journals to do that! Just because they don't do it isn't the journal's fault! Not only that, many journals nowadays allow for the authors to "pay extra" to make the paper open access. Most, if not all, of the LHC papers are that way. Furthermore, PRST-AB, PRST-PE, and PRX are all open access! So what's the problem there?

Scientific institutes and universities pay huge amounts of money for their subscription to journals, and for access to the Web of Science. The crisis has made some universities cancel subscriptions, and this is a permanent problem in developing countries.

For many journals, including Phys. Rev., they not only provide different-tier pricing for their journals, in many cases, they also provide access for FREE. Phys. Rev. journals, in fact, gives public libraries and high schools free access to their journals! In fact, in most cases, going to your public library and requesting a paper from the librarian will typically get you such papers, even if they don't have free access to them.

Again, I really do not see this "I can't get access" excuse anymore, certainly not in this day-and-age. It is a very weak excuse. If this is all the "fault" that you have for the peer-reviewed journals, I am not in your corner.

Zz.

 

[ZapperZ]

@Vanadium50, in theoretical physics, we do typset our own papers. We're usually required to use the journal macro system for LaTex, which we do willingly.

Actually, you don't. Just compare your preprint typeset with the final version that actually appeared in the journal. If you submitted to Science/Nature, I will bet you 100% that your "typesetting" appears nothing like what will appear in those journals.

In PRL/Phys. Rev. journals, the same could be said. Your figure, etc. will not appear where you put them. In fact, the journal instructions specifically mention that your figures should not be imbedded in your text.

So no, you do NOT do the typesetting. The LaTex template is there to allow you to have an idea of the layout, especially page length. The journal still has to employ someone to do the final typesetting.

Zz.

 

[dcpo]

Why is typesetting so important? If papers are published in electronic form only then papers don't need to fit into a compilation. Beyond the minimum of typesetting required to create readable work, something easily achievable by the author, what is the purpose of the additional work? Creating a unified stylistic 'feel' for a journal? How important is that? For me, not very.

 

[ZapperZ]

Why is typesetting so important? If papers are published in electronic form only then papers don't need to fit into a compilation. Beyond the minimum of typesetting required to create readable work, something easily achievable by the author, what is the purpose of the additional work? Creating a unified stylistic 'feel' for a journal? How important is that? For me, not very.

Still, that is a the journal's requirement. You are welcome to question each of the different journals why such-and-such is required. You will also note that there are books and books on style manual, etc. You may not appreciate it, but you inherently depend on it each time you open a journal and automatically know where to look for stuff and automatically can make guess where things are located.

Journals such as Science and Nature has to make huge type-setting decisions because in many cases, figures often cross more than single columns, and articles don't normally start on its own page! Just because you don't see it doesn't mean a lot of things do not get done behind the scenes. Who do you think looked at the labels on the figures and noticed that they are just way too small to be seen clearly when inserted into the document?

Zz.

 

[jrlaguna]

@ZapperZ: I get most of my papers easily, that's *not* the point. The point is the money that universities and research institutes spend on journals. You should take a look at the prices, they're amazingly high. As I tell you, the cost they allege is $1000 to $10000 per article... So that's what they charge (summing up all institutions paying, I guess).

Referee assignation is a nearly-automatical procedure. Journals keep databases of possible reviewers, with field of expertise. Typesetting, as I tell you, is a scientists' task. Proofreading is something that they do, but it's not scientific proofreading, they only search for English mistakes and typos.

My proposal can be made more concrete. Without entering in direct competition, ArXiv might create an experimental "peer-review" stamp, to be asked for voluntarily by submitters. It might also ask for voluntary reviewers. In order to ensure quality, let us say that only endorsers may be eligible. Each reviewer would give his "keywords". Without much cost, reviewers might be selected at random, according to the keywords, and be asked to referee.

The proposal goes far beyond, extending to scientific publishing 2.0. For example: a way of discussing papers and making comments and questions. I like the idea of http://cosmocoffee.info/index.php, but it has almost no activity.

 

[jrlaguna]

In my opinion, universities and research institutes are subsidizing heavily private companies (publishers + Thomson Reuters) in exchange for a nearly negligible contribution to science. They only provide their prestige. But their prestige is OURS.

 

[dcpo]

Still, that is a the journal's requirement. You are welcome to question each of the different journals why such-and-such is required. You will also note that there are books and books on style manual, etc. You may not appreciate it, but you inherently depend on it each time you open a journal and automatically know where to look for stuff and automatically can make guess where things are located.

Journals such as Science and Nature has to make huge type-setting decisions because in many cases, figures often cross more than single columns, and articles don't normally start on its own page! Just because you don't see it doesn't mean a lot of things do not get done behind the scenes. Who do you think looked at the labels on the figures and noticed that they are just way too small to be seen clearly when inserted into the document?

Zz.

I'm sure these things are done behind the scenes, and I'm sure it costs money. My question is whether what is achieved is worth the premium that is paid for it. I read a lot of preprint articles, and they are often different to the journal version, but I've not found that the journal versions are superior, just more stylistically aligned with the journal. I doubt that inconsistency in typesetting would be a serious bar to the readability of a journal, anymore than differences in typesetting between journals is a problem. It's really about the look of the thing, which is less important as the role of print versions diminishes.

p.s. I'm talking about maths here, maybe the situation is different in physics.

 

[jrlaguna]

@dcpo, I am a physicist and I agree with you. I read either the ArXiv preprint or the final paper, I don't really care which one. "Professional typesetting" is negligible in most journals, but for Science and Nature, which are very fond of very pretty and ellaborate pics (which add nothing relevant to understanding), and one of my profs at SISSA called (in jest) "scientific pornography".

 

[ZapperZ]

I'm sure these things are done behind the scenes, and I'm sure it costs money. My question is whether what is achieved is worth the premium that is paid for it. I read a lot of preprint articles, and they are often different to the journal version, but I've not found that the journal versions are superior, just more stylistically aligned with the journal. I doubt that inconsistency in typesetting would be a serious bar to the readability of a journal, anymore than differences in typesetting between journals is a problem. It's really about the look of the thing, which is less important as the role of print versions diminishes.

p.s. I'm talking about maths here, maybe the situation is different in physics.

Actually, look up mathematics typesetting. Again, whole books are on these, especially on formatting mathematical equations. So it isn't trivial, and you are directly benefiting from the uniformity on how such things were typeset.

And yes, I was referring to physics papers, since it is a physics issue. I hate to think that we are talking about including all peer-reviewed journals, especially medical journals which have a whole different set of criteria.

Zz.

 

[jrlaguna]

ZapperZ, I agree with you that mathematical typesetting is a very interesting issue... that is tackled by mathematicians itself, not by editors. TeX was not developed by someone in the publishing industry, but by Donald Knuth, an outstanding computer scientist and mathematician. There are hundreds of useful packages which are developed by... us, scientists! Even I have contributed myself to the development of TeX and LaTeX. The journals do nothing relevant in this issue, you must concede.

 

[dcpo]

Actually, look up mathematics typesetting. Again, whole books are on these, especially on formatting mathematical equations. So it isn't trivial, and you are directly benefiting from the uniformity on how such things were typeset.

And yes, I was referring to physics papers, since it is a physics issue. I hate to think that we are talking about including all peer-reviewed journals, especially medical journals which have a whole different set of criteria.

Zz.

It's not just a physics issue, arXiv serves maths papers too. I'm not saying typesetting is trivial, I'm saying it's not worth the expense. I don't think it's possible to get through peer review without producing work presented with decent quality. In maths at least almost everyone uses latex, which gives the author the power to produce material with a high degree of professionalism. The typesetting done by journals goes much further than readability, and is only strictly necessary for the production of print journals. Think about how much the editors and reviewers of a journal get paid, and how much it costs for an institutional subscription. The typesetters may work hard, but are they doing work that needs to be done? How much do we collectively want to pay for the production of print journals?

 

[jrlaguna]

$1000 - $10000 for a "pretty printed" article...? That's too much, dude! :)

 

[ZapperZ]

It's not just a physics issue, arXiv serves maths papers too. I'm not saying typesetting is trivial, I'm saying it's not worth the expense. I don't think it's possible to get through peer review without producing work presented with decent quality. In maths at least almost everyone uses latex, which gives the author the power to produce material with a high degree of professionalism. The typesetting done by journals goes much further than readability, and is only strictly necessary for the production of print journals. Think about how much the editors and reviewers of a journal get paid, and how much it costs for an institutional subscription. The typesetters may work hard, but are they doing work that needs to be done? How much do we collectively want to pay for the production of print journals?

This is getting sidetracked. The point here is that a claim was made that scientists did ALL of the typesetting. This is FALSE!

Whether you agree that such type setting is needed or not is irrelevant. The rebuttal here is against the point being made above. The typesetting that is done as it appears in the final form is NOT the one that the authors did! You can argue till you're blue that it isn't needed, but that's a different thread and a different topic.

Zz.

 

[ZapperZ]

$1000 - $10000 for a "pretty printed" article...? That's too much, dude! :)

And dude, you ASKED for our "thoughts". And now you're dismissing it using such a flaky argument. I'm sorry I've wasted my time for nothing.

Zz.

 

[dcpo]

This is getting sidetracked. The point here is that a claim was made that scientists did ALL of the typesetting. This is FALSE!

Whether you agree that such type setting is needed or not is irrelevant. The rebuttal here is against the point being made above. The typesetting that is done as it appears in the final form is NOT the one that the authors did! You can argue till you're blue that it isn't needed, but that's a different thread and a different topic.

Zz.

I must say I disagree with your interpretation of the purpose of the thread. The argument that authors do all the typesetting may be flawed, though I say the modified argument that authors do all the typesetting necessary for electronic dissemination stands, but it is only one amongst a number of arguments against the current journal system. You may be talking about this one issue alone but it is not the only issue relevant to the thread.

 

[jrlaguna]

@ZapperZ, scientists do not do ALL the typesetting. But they do all the RELEVANT typesetting. Is the difference worth $1000-$10000?

I feel you're quite susceptible. I thank you for your comments and sharing your thoughts with me, but you should also be open to criticism. You dismissed my proposal in harsh terms, and you do not provide arguments for that. Please, consider improving your courtesy when discussing.

 

[Choppy]

I don't think I've ever seen anyone charge $10000 for a journal article. Maybe the journals are vastly different from field to field. I've seen costs in the $20 - $100 range, but $10000? Do those articles come with a secretary that reads it to you?

Or have I misread and you're claiming that a journal makes $1000-$10000 per article? Even that I would surprised at. In my, admittedly ignorant view, I would suspect a typical jornal might make that much per issue.

 

[ZapperZ]

I must say I disagree with your interpretation of the purpose of the thread. The argument that authors do all the typesetting may be flawed, though I say the modified argument that authors do all the typesetting necessary for electronic dissemination stands, but it is only one amongst a number of arguments against the current journal system. You may be talking about this one issue alone but it is not the only issue relevant to the thread.

I'm sure there are other issues in here, but the issue on "typesetting" was very clear based on the OP:

The scientific publication scheme is clearly faulty. Although most research in fundamental science is funded by the taxpayer, the average citizen does not have access to the results. Scientific journals are typically quite expensive, although they offer little-to-nothing added value: authoring, typesetting and reviewing is done by the scientists themselves, typically, for free. The database used for impact evaluation is also a private company (Thomson Reuters).

So... why not, as a first step, create a peer-review system on the ArXiv? There are more ideas in this link

http://physicsnapkins.wordpress.com/...ific_journals/

What are your thoughts?

And...

@Vanadium50, in theoretical physics, we do typset our own papers. We're usually required to use the journal macro system for LaTex, which we do willingly.

My argument here is that, no, you do NOT do the typesetting for the journal. The LaTex template is a GUIDE for the author to judge the length and layout, NOT for the authors to do typesetting for the publication! You are more than welcome not to use those template if you do not care. In fact, many people submit their manuscript using WORD documents!

So if you want to argue about the usefulness, non-usefulness of typesetting in journals, open your own thread and I'll argue with you on that. But if you look at my first response on this particular issue, I specifically showed why such a claim of authors doing the typesetting for the journal to be completely FALSE. Proof: the final version of the publication does NOT look like the one submitted by the authors. Do you disagree with this?

Zz.

 

[ZapperZ]

@ZapperZ, scientists do not do ALL the typesetting. But they do all the RELEVANT typesetting. Is the difference worth $1000-$10000?

And you think the ONLY expense in producing a journal is all the typesetting? Really?

I feel you're quite susceptible. I thank you for your comments and sharing your thoughts with me, but you should also be open to criticism. You dismissed my proposal in harsh terms, and you do not provide arguments for that. Please, consider improving your courtesy when discussing.

Susceptible? What does that mean?

I did not dismiss your proposal. I spent considerable amount of time in typing ALL those responses. Yet, you seem fit to simply reply "$1000 - $10000 for a "pretty printed" article...? That's too much, dude! :)" I find that to be even MORE dismissive! Your "criticism" is based simply on a matter of TASTES, not based on any facts. Think about it! You disliked something because you just did like it for a personal preference, not because you think there's something inherently wrong with it. You didn't say why such and such a cost is beyond the cost of doing business or maintaining a journal.

And yes, I would like to know where did you actually submit to be charged THAT high of a publication fee.

Zz.

 

[jrlaguna]

Apparently, I didn't explain well. $1000 to $10000 is, according to Paul Ginsparg (very relevant high energy theoretician, father of the ArXiv) the cost of production, which is charged to the subscribing institutions (to all of them, not to each!). Pls, read the source :) So it is payed by the scientific community and, in the end, by the taxpayers. Isn't it just too much?

ArXiv costs $10 per paper. I see no relevant difference in look, thanks to Knuth and Lamport. The extra cost would be to create a peer-review system, so... maybe $50?

 

[ZapperZ]

Apparently, I didn't explain well. $1000 to $10000 is, according to Paul Ginsparg (very relevant high energy theoretician, father of the ArXiv) the cost of production, which is charged to the subscribing institutions (to all of them, not to each!). Pls, read the source :) So it is payed by the scientific community and, in the end, by the taxpayers. Isn't it just too much?

ArXiv costs $10 per paper. I see no relevant difference in look, thanks to Knuth and Lamport. The extra cost would be to create a peer-review system, so... maybe $50?

The AUTHORS are NOT charged that much. The cost of production is different than the authors cost! Have you never published anything yet? If you have, how much did YOU or your institution have to pay? Go by first-hand knowledge rather than by hearsay! I have NEVER, in all the years that I've published papers (Nature, PRL, PRB, PRST-AB, JAP, APL, Physica-B) paid THAT much!

$50? How can you tell? If ArXiv becomes a peer-reviewed journal, they will have hire full-time editors in each of the relevant field of study to (i) weed out incoming manuscripts (ii) figure out what is "important" versus what is merely "interesting" (iii) find relevant referees (iv) keep a record in referees and their activities/complaints/compliments/etc. (iv) go to conferences to be up-to-date on the development in each of the fields, etc... etc. If you think publishing journals is simply dumping stuff into a publication and that's that, you should do yourself a service and talk to a journal editor and figure out what he/she has to do! They have a thankless job where, if they do their job very well, they are invisible to the rest of us, and, apparently, very much to you. And yet, what they do takes a lot of money, because these people themselves have to be experts in these various areas (Phys Rev. hires PhDs in these various fields to be their associate editors!).

And oh, do yourself another favor, and look at the VOLUME of papers published just in PRB alone in just ONE month! 2 very thick volumes at the 1st of the month, and another 2 very thick volumes on the 15th of the month! How many editors do you think is required to weed out ALL of the submitted papers? How much work do you think to deal with handling the remaining manuscript that got through the editors and then have to go through the refereeing process? Do you think all these manpower is cheap? It isn't! ArXiv going into the SAME standard will incur such similar expenses!

Zz.

 

[dimensionless]

The scientific publication scheme is clearly faulty. Although most research in fundamental science is funded by the taxpayer, the average citizen does not have access to the results. Scientific journals are typically quite expensive, although they offer little-to-nothing added value: authoring, typesetting and reviewing is done by the scientists themselves, typically, for free. The database used for impact evaluation is also a private company (Thomson Reuters).

So... why not, as a first step, create a peer-review system on the ArXiv? There are more ideas in this link

http://physicsnapkins.wordpress.com/2012/01/16/occupy_scientific_journals/

What are your thoughts?

I thought part of the idea of ArXiv was that it was not peer-reviewed, and that also many good ideas don't always get through peer-review. There's slightly more to the process than simply producing valid and useful scientific research.

Researchers need to use their best discretion with ArXiv, but they have to other sources as well. I would expect the authors of most peer reviewed research papers to have performed a Google search and to have skimmed the relevant Wikipedia pages (if they had not already done so).

If you need to contact an author for a paper, you can usually find an email address. If none is available, you can often find them on FaceBook and LinkedIn.