Stevan Harnad, University of Southampton, May 30, 1999

> Christophe Pallier wrote:
>
> Dan Ellis posted your advertisement for Cogprints on the AUDITORY list.
> Two of us had an argument about the idea that authors should make their
> work freely available on the net. I include part of the exchange at the
> end of this message.
>
> This led me to the 2 following ideas:
>
> 1. If not already done, one could maintain a web page containing a list
> of the journals which accept free dissemination of papers on the web.
> (I can volunteer to create and maintain such a page). We would
> recommend authors to send their papers to these journals. But how many
> of these journals are there?

A natural idea, and a benign one, but it is doomed to fail, and so it should. Authors will (and should) continue to prefer to submit their papers to the highest quality, most prestigious and highest-impact journal for which it is eligible. Hence a black-list is likely to have very little effect (though it will do no harm and might do a little good).

The much better and simpler thing to do, which will succeed, and could succeed very quickly, is simply to encourage all authors to publicly self-archive all their papers (both unrefereed preprints AND refereed reprints). The attempt to block self-archiving is so completely in conflict with the interests of research and researchers, and so unenforceable, that it is certainly doomed to fail -- and has already failed in Physics, because of a de facto "class action" (in the form of massive self-archiving) that is irreversible, and has now led to the most progressive and enlightened copyright policy of all on the part of the American Physical Society, publisher of the most prestigious and highest impact journals in Physics, a model for all future learned journal copyright policies:

http://xxx.lanl.gov/cgi-bin/show_monthly_submissions

http://www.cogsci.soton.ac.uk/~harnad/Hypermail/Author.Eprint.Archives/0006.html

> 2. Why not try to write a generic copyright notice for scientific work,
> in the same vein as the GNU Public License from the Free Software
> Foundation (www.fsf.org).
> It could one or two paragraphs, stating that the work we want to
> publish must be made freely available to reproduce by anybody. We could
> then, as authors, insist on having these paragraphs inserted in the
> copyright transfer agreements we sign with publishers.
> I don't expect this to work easily, but hey, why not try? We coudl try
> to launch a campaign like the one for "free speech".

APS are already well on the way to providing this model. As a start, see:

http://www.cogsci.soton.ac.uk/~harnad/Hypermail/Author.Eprint.Archives/0006.html

> The best would be if some publishers endorsed the idea of free
> scientiific work, and have the authors pay the copyediting & formatting
> job.

That is indeed the target, but the hope is that these will be the SAME publishers that now publish the established journals, but restructured for this new online world.

http://www.cogsci.soton.ac.uk/~harnad/nature.html

> A compromise may be to leave the exclusivity of publication to the
> publishers for a short period, say one year.

Absolutely not! What nonsense! What researcher would or should agree to a needless one-year embargo on research findings, particularly in the critical initial year! Don't under any circumstances ever accept a Trojan Horse like that.

> This may be very naive. You have a lot more experience than me with
> issues. What do you think we can do as authors and reviewers?

It is indeed naive, though well-intentioned. See further comments below.

> Christophe Pallier
> http://www.ehess.fr/centres/lscp/persons/pallier/
>
> --------------------------------------
> Christophe Pallier wrote:
>
> Pierre Divenyi wrote:
>
> How beautiful and Platonic an idea: an electronic preprint archive where
> everybody could post his/her new opus within minutes, to be read by tens of
> thousands of pairs of interested eyes!

Don't confuse preprints with reprints. Unrefereed preprints can be posted at once; but refereed reprints will still first have to undergo peer review, which can be accelerated a little online, but will continue to be a retardant for as long as referees (donating their services graciously and gratis in accordance with the academic golden rule -- for there is not enough money in the world to compensate them for their heroic services, so don't even think of that) have other things to do with their time besides instantly evaluating every one of your papers and mine ("they" are, after all, US).

> Unfortunately, as long as our own mainstream auditory journals oppose
> on-line dissemination of pre-publications, and enforce their opposition
> through automatic rejection of papers disseminated this way, and as long as
> our mainstream granting agencies insist on peer-reviewed publications as
> representing the major (if not the sole) proof of scientific productivity,
> Professor Harnad is putting the cart before the horses. Moreover, even a
> cursory visit at the web sites he suggests makes it clear that, should an
> unsuspecting colleague except his offer and post his/her paper on the
> preprint archive, he/she may shoot him/herself in both feet at once.

Don't be so fatalistic. Look instead at the empirical data. No feet were shot in Physics, where the game is now over:

http://xxx.lanl.gov/cgi-bin/show_monthly_submissions

> This kind of short-term, individualistic rationale distresses me.
>
> If we all shoot, then the bullets may not reach our feet but,
> hopefully, the heart (or rather the wallet) of the sharks of scientific
> publishing.

Don't demonize the publishers. You would do the same in their shoes.

They will only scale down to what is optimal and inevitable for research and researchers when they clearly feel that they have to, and for that, WE researchers have first to realize what is optimal and inevitable for us, and act accordingly. Class action, in the form of universal self-archiving, will accomplish both goals: to free our journal literature and to send our publishers the message that they must restructure themselves to accommodate it.

http://www.arl.org/scomm/subversive/toc.html

> True: If only a small proportion of us follows Harnad's lead, we might
> end up in troubles...
> It is a case of the well-know prisoners' game: if we cooperate, we all
> win (maybe less than if we play alone), but if we don't cooperate,
> some, maybe the majority, will lose a lot.

There was no prisoner's dilemma in Physics. See URL below. Moreover, we now have the advantage of the precedent of Physics already in place. They are, after all, kin of ours, hence part of the "class action."

http://xxx.lanl.gov/cgi-bin/show_monthly_submissions

> It doesn't take a complex demonstration to be convinced that now that
> the cost of publishing scientific results has dramatically dropped, the
> existence of publishers who charge huge prices and prevent widespread
> dissemination of the papers, is an anomaly.

It is not their existence that is an anomaly, but the continuing needless expenses. Subversion through self-archiving will bring this into line with reality. The demand for paid paper journals will not vanish at once (no significant cancellations have yet been detectable in Physics, though they will no doubt come, eventually); there will be time for rational restructuring; but meanwhile the free online literature will already be there for us all.

http://amsci-forum.amsci.org/archives/september-forum.html

> It seems obvious that a very small amount of the funds that are devoted
> to research could be invested in paying the few people needed to
> maintain scientific electronic journals, which content could be
> accessed freely by anyone. I am not against private enterprise and
> indeed, this job might very well be done by private publishers, if they
> can offer a better service than public agencies. (Why not have the
> source, that is the author(s) pay a reasonable amount to have the paper
> published. The price would pay for the few hours (or less) of work
> needed to format the paper for electronic publication, and maintain
> servers).

There is no need for new entities to take this service over from the established journals; they have the experience and expertise; they need only restructure for the new circumstances, which are indeed likely to entail scaling down to online-only, and selling, instead of the journal itself (which will be archived free for all), only the service of implementing peer review and certification:

http://www.cogsci.soton.ac.uk/~harnad/nature2.html

> Note that in this scheme, there is of course no reason why the
> reviewing process should be any different for these journals than for
> the ones we currently have.

Correct.

> I am unsure whether this will ever happen: current publishers don't
> want to lose the goose with the golden eggs, and are fighting hard to
> prevent this from happening. Rather, some of them try to install a kind
> a pay-per-view system. This makes me sick...

The cost-recovery model that publishers are attempting to retain is Subscription/Site-License/Pay-Per-View (S/L/P). I have dubbed S/L/P the "trade troika," because all three are predicated on access-barriers, because they are selling a product, the article/journal, rather than a service, the quality-control/certification. Up-front payment for this service makes most sense, because it frees the literature from toll-barriers. The author-institution, instead of subsidizing the literature by a huge S/L/P expenditure to buy it back, instead pays for it up-front, out of only a small portion of its own S/L/P savings!

So there is no need to look for outside subsidy (except initially, during the transition period). Because the cost of implementing quality control alone will be so much lower than the current costs of doing it all, the S/L/P savings themselves will be enough to cover the up-front costs with plenty left over to spend on essential things (such as books, which definitely do NOT fall under this nontrade model, because books-authors, like book-publishers, want fees or royalties from the toll-gate receipts, whereas with the journal literature this is not, and never has been, the case).

> At this point, we have the choice between two attitudes:
>
> 1) an egoistic attitude: putting our career before our scientific
> ideals, and not caring about this issue: just compete to publish in
> the "best" journals.
> The tax-payers will pay the costs, and what's the problem if our
> colleagues can't access the information?...
>
> 2) a responsible attitude:
> - refuse to submit or review papers in journals handled by publishers
> that refuse to allow free access to the papers (either on the authors'
> web site or on their own).
> - fight to convince journals editors to change publishers: why a
> journal couldn't move to cogprints? The journal may consists of a web
> page with links to the accepted papers.

There is another option, which is having your cake and eating it too: Continue to submit to your established journal of choice, but self-archive as well. This subversive path has been followed, with astounding success, in Physics:

http://xxx.lanl.gov/cgi-bin/show_weekly_graph

We need not keep debating it all, meanwhile falling prey to Xeno's Paradox. We need only stride ahead and self-archive. The infrastructure for it is in place (at all our home institutions) and more is on the way:

http://www.nih.gov/welcome/director/ebiomed/ebiomed.htm

http://library.caltech.edu/publications/ScholarsForum/

> If, as an author, this is impractical right now, the correct stance is
> to *disobey inane copyright laws* (and convince your peers to do so):
> publish accepted papers on servers like cogprints.
> It is only if the scientific community does this massively that we have
> a chance to prevent the pay-per-view system to win.

Correct. But don't be so sure you are disobeying laws either. There are massive untested and unreflective ambiguities and vaguenesses here: There is absolutely no law about self-archiving unrefereed preprints (only arbitrary and unenforceable policies on the part of some journal publishers), and the law about self-archiving of refereed reprints has a slippery slope with respect to the versions: How many changes in my unrefereed preprint constitutes stepping over the line and making it into a refereed reprint? Besides, authors need not and should not sign away their self-archiving rights; here too a class action is in order.

The critical factor (and everyone keeps forgetting this) is that copyright law is intended to JOINTLY protect the publisher and the author from theft of text. This is fine for royalty- fee-based books and magazines. But where the author wants to GIVE the text away rather than to sell it, it becomes a very different ball-game...

http://www.cogsci.soton.ac.uk/~harnad/science.html

> They are people who try to make a better world happen. Why not take the
> example of the Free Software Foundation (www.fsf.org) to create a Free
> Science Foundation?
>
> Christophe Pallier
>
> Response
>
> Cher Christophe,
>
> You have completely missed the point of my comment to Stevan Harnad's
> suggestion on electronic publication. The main reason is that you see
> the situation from your own environment which, thanks to a slew of
> built-in protection for academics and researchers that you benefit from
> in France as well as in the majority of Western European countries,
> makes the question of whether to publish or not by-and-large optional.

No such thing. Publishing in rigorous refereed journals is critical everywhere in the active scientific/scholarly world. France cares just as much for "impact factors" as the UK or US do.

> Let me tell you that in the U.S. it is not. In other words, you are
> quite pampered-spoiled by our standards. Maybe you should also be
> informed that the majority of American contributors to the auditory
> list is able to do research through the sole support of government
> agencies that adhere to the publication policies I outlined in my note.
> Thus, you should not try to admonish those of us for whom there is no
> alternative but to adhere to these policies. If you want the policies
> to change, address your criticism to the agencies.

Funding agencies mandate only that the research findings be published (in reputable refereed journals) -- not that those journals should black access to them in return for refereeing and certifying them. On the contrary, there is a strong move toward retention of self-archiving rights:

http://www.cogsci.soton.ac.uk/~harnad/science.html

> For your information, personally I would be quite in favor of having
> all publications available electronically. My private opinion, however,
> weighs very little in this matter which amounts to fighting windmills
> stronger than even Stevan Harnad's personal opinion: according to what
> I gathered from the information on the web pages he wanted us to see,
> his many years of effort, alas, have accomplished very little. The
> establishment is strong and you guys in France are unlikely to be able
> to export a second French Revolution to conquer it. >
> Pierre Divenyi

Courage, chers cocombatants! The battle has been won in Physics, and all that's needed to carry this on to the rest of the disciplines is to emulate what the Physicists did (thanks to Paul Ginsparg, to whom all power and glory!). Just self-archive, and the rest will take care of itself.

--------------------------------------------------------------------
Stevan Harnad
Professor of Cognitive Science
Department of Electronics and Computer Science
University of Southampton
Highfield, Southampton
UNITED KINGDOM
http://www.cogsci.soton.ac.uk/~harnad/
http://www.princeton.edu/~harnad/
ftp://ftp.princeton.edu/pub/harnad/

Stevan Harnad, University of Southampton, May 30, 1999

> Dear Stevan,
> as guest editor for Brain and Language, the manuscript for
> a special issue has been sent off to Academic Press and I
> guess is in production. I wondered whether I could advice
> the contributors of the special issue to post their articles
> in the cog archives. AP has a paragraph about "Personal
> Servers" in the Notice section at the end of B&L saying
> "When an Academic Press journal accepts the work for
> publication, the authors may post it, in its final accepted form,
> on their personal servers (but not on any organized preprint
> server) with a notice _Accepted for publication_ in ....etc.
> After publication, authors may post their Academic Press
> copyrighted material on their own servers without permission,
> provided ....."
> (Now they don't talk about "organized preprint servers"
> any more.
>

> I know that some of my contributors cannot post their
> articles on personal servers. Are the cog archives
> considered as a "organized preprint server"?
>

> You may understand that I do not want to encourage my
> authors to something that puts them into trouble. But
> I would like to support free electronic availability
> of articles.
> Have you had any experience with AP on these matters?
>

> I know you must be awfully busy with all your engagement
> in on- and off-line publishing, so this is not an urgent
> message.
>

> best wishes,
> Brigitte Stemmer

Dear Brigitte, sorry for the long delay: Overload. The answer is that the personal/public server distinction is completely incoherent and hence untenable: Every "personal" server is public -- reachable by anyone on the Web, indexed by all search engines, duplicated in countless public cache sites, mirrored, etc.; in addition, any "public" server in which the author archives himself, password-protected, and can add/delete as he pleases, is "personal."

So publishers who try to make and enforce this non-distinction are just playing word games in trying (desperately, and doomed to lose) to hold onto something that cannot be held onto, logically, morally, or practically. (The real distinction they are trying to re-create here, but cannot, is the distinction between privately distributing one's own offprints vs. publishing them with another publisher, but of course that does not fit the new situation at all!)

So you can decide for yourself whether you want to collaborate with AP, and reinforce this incoherent, illogical distinction, which will have the effect of deterring some authors for a few more years, through ignorance and timidity on their part, or you whether you prefer to take a step for good sense and progress and what is indisputably infinitely better for research and researchers and will prevail sooner or later, by telling them that they have the right to self-archive online as they please, their own papers, for which they never received or requested a penny from AP! The only thing that should/can be forbidden is (1) to sell them, or (2) to publish them with someone else who sells them.

For a model of the copyright agreement of the near future, see the self-archiving policy of the American Phsyical Society, publisher of the most prestigious and highest impact journals in Physics:

http://www.cogsci.soton.ac.uk/~harnad/Hypermail/Author.Eprint.Archives/0006.html

Cheers, Stevan

--------------------------------------------------------------------
Stevan Harnad
Professor of Cognitive Science
Department of Electronics and Computer Science
University of Southampton
Highfield, Southampton
UNITED KINGDOM
http://www.cogsci.soton.ac.uk/~harnad/
http://www.princeton.edu/~harnad/
ftp://ftp.princeton.edu/pub/harnad/

Stevan Harnad, University of Southampton, May 29, 1999

On Fri, 28 May 1999, Paul Smaglik wrote:

> ANALYSIS
> By Paul Smaglik
>
> It's going to be a preprint service. It's going to be a reprint
> repository. It's going to kill off society journals. It's going to save
> them. It's going to compete with commercial titles. It's going to
> complement them.
> There appears to be no consensus on the effect E-biomed, a
> potential government-backed electronic publishing service proposed by
> Harold Varmus, director of the National Institutes of Health, will have
> on other journals-both paper and electronic. Nor does there appear to be
> much agreement on what form that service will take. "There's a vagueness
> in Varmus' proposal," comments Stevan Harnad, professor of cognitive
> science at Princeton University and the University of Southampton in
> England.

But you MUST add: "but once this is resolved, there is a viable core that can have the revolutionary impact of freeing the biomedical journal literature for one and all forever."

> Varmus, who acknowledges that the proposal is young, calls that
> vagueness "evolvability" [see Varmus interview, page X]. While words
> like "vagueness" (and nonwords like "evolvability") are being applied
> specifically to E-biomed, they might well serve as accurate labels for
> electronic publishing as a whole. The field has splintered into a myriad
> of permutations. E-publishing now includes electronic reprint sites,
> such as a cognitive science one run by Harnad;

Paul, to put this in context, you should really say:

E-publishing now includes electronic preprint and reprint sites, such as the remarkably successful Physics Archive at Los Alamos, run by Paul Ginsparg, and its emulators in other disciplines, such as CogPrints, the one in cognitive science run by Harnad;

If you don't put it this way, the statement is neither representative nor informative. There is NO DISTINCTION between preprint and reprint servers: they are for self-archiving by authors, who can put either preprints or reprints in there.

URLs:

http://xxx.lanl.gov/cgi-bin/show_monthly_submissions
http://cogprints.soton.ac.uk/">

> E-only journals, such as
> MedGenMed, run by former JAMA:The Journal of the American Medical
> Association Editor George Lundberg;

To emphasize that there is NO INCOMPATIBILITY between running eprint archives and e-only journals, I suggest that you mention Psycoloquy, the e-only journal I have been running since 1990, the first peer-reviewed e-only scientific journal, and, paradoxically, sponsored by the American Psychological Association (APA), the biggest and most prestigious paper-journal publisher in Psychology. This just shows that Learned Societies can be extremely progressive (funding free journals) at the same time as being reflexively regressive (APA has one of the most restrictive copyright policies at the moment, along with Science and the New England Journal of Medicine; in contrast, Nature is more progressive, and the American Physical Society (APS), the APA's counterpart in Physics, has the most progressive copyright policy of all, one that will serve as a model for all the others):

Copyright:
http://www.cogsci.soton.ac.uk/~harnad/science.html
http://www.cogsci.soton.ac.uk/~harnad/Hypermail/Author.Eprint.Archives/0006.html

Psycoloquy:
http://www.princeton.edu/~harnad/psyc.html

> preprint archives, such as a popular > physics site sponsored by Los Alamos National Laboratory;

Drop this in favour of the prior mention above. It is artificial and counterproductive to try to distinguish preprint and reprint servers. The only separating issue is copyright, and as noted, this varies from journal to journal, rather than from preprint to reprint.

> and commercial
> sites by publishers including Elsevier Science and John Wiley & Sons.

Stress that sites like Los Alamos and CogPrints are for self-archiving by authors and they are FREE, whereas sites like Elsevier's and other journal proprietary archives are for FEE!

> Which of these disparate sites will thrive may depend on issues that
> affect them all. Who deservedly holds the copyright for research
> articles, what is the tolerance for "tolls" on the Information Highway,
> and how will the Web change the nature of peer review.
> Subversive Proposal
> Harnad suspects-and hopes-that Varmus' plan will eventually
> resemble the "subversive proposal" Harnad made years ago. That proposal

how about "converge on" rather than "resemble."

Also, I strongly suggest, for completeness, that you also mention the Scholar's Forum, an initiative parallel to the NIH one, and potentially even bigger, because it includes ALL disciplines, not just the biomedical ones, and it includes the top US Universities (and potentially all of them), not just NIH:

http://library.caltech.edu/publications/ScholarsForum/

Don't underestimate this CalTech initiative. I have heard through the grapevine that there is a GREAT deal of muscle behind it, as it is being promoted by the Provosts of the US Universities, the ones most conscious of the huge drain on University budgets represented by learned journals, as well as the huge limitation on the potential impact of University research represented by the access barriers of the Subscription/Site-License/Pay-Per-View (S/L/P) system that they too would like to subvert.

http://www.chronicle.com/free/v45/i04/04a02901.htm

> calls for authors to archive their published, peer-reviewed papers on
> their own Web sites and give them away for free. Harnad notes that since
> authors don't get paid for their efforts, "there's no reason they
> shouldn't be able to give their own work away."

Note here that they already DO give it away, and always have done, in the form of reprints that they themselves pay to produce and mail (for free) to those who want them; the Net will just become a big, universal reprint distributor for the author.

> To have their work
> mounted, authors-or institutions-could perhaps pay the journal that
> published it a fee that is less than the yearly subscription rate.

NO, NO! Please don't attribute this completely counterproductive view to me! Authors have already given their papers to their publishers for free, so that their publishers can sell them. It would be absolutely grotesque that authors should now, like libraries, PAY to buy back their own work so that they themselves can in turn give it away for free! Please think before saying such things!

> Such

> an arrangement would reduce the journal's role to peer review and a seal
> of approval.

THAT is the service that authors' institutions can and should continue to pay for, but not through access-blocking S/L/P but through direct, up-front payment for this quality-control and certification service, out of only a small portion of the institutional S/L/P savings.

PLEASE get the logic and pragmatics of this straight, otherwise you are simply advocating another variant of S/P/L access-blockage (in which the author's institution is now paying L -- a global site license -- for the "right" to make specific papers of their own accessible to everyone for free: there is no justification whatsoever for that; only the quality control and certification service needs to be paid for!).

> The drawback? Copyrights. Many journals do not let authors
> retain copyright. And commercial and society-based journals likely
> wouldn't voluntarily give up the subscription income that makes them
> viable. Still, Harnad thinks that may change. He notes that the American
> Physical Society recently gave copyright control back to authors who
> submit papers to its journals. "The game is over in physics," Harnad
> comments. If other societies followed suit, that could pressure
> commercial publishers to do the same. But that's a big "if," notes Helen
> Atkins, director of database development at the Institute for Scientific
> Information (ISI) in Philadelphia. "[Harnad has] been proposing this for
> a long time. The basic idea he has is very interesting. I don't see
> anybody doing it."

APS is somebody; and they are not the only ones: Perhaps you should do a survey of evolving copyright policy. My bet is that it's moving toward the APS model. Subversion from self-archiving in Los Alamos, CogPrints, E-Biomed and Scholar's Forum will help hasten the process.

ftp://aps.org/pub/jrnls/copy_trnsfr.asc

> The top biomedical journals are especially protective
> of copyrights, but if E-biomed becomes enough of a force, that could
> change. Harnad thinks E-biomed will be more successful as a reprint
> repository for existing journals than a new, competing one. "We don't
> need more journals," he concludes.

And continue: "We need infrastructures that will facilitate self-archiving by authors, as Los Alamos has done in Physics. The rest will follow suit, as it has with the APS."

http://www.cogsci.soton.ac.uk/~harnad/Hypermail/Author.Eprint.Archives/0006.html

> E-only Options
> Lundberg would disagree, although he declined to comment on the
> need for E-biomed. He is the Northwestern University-based editor of
> Medscape General Medicine (MedGenMed), a clinical medicine Web site
> designed for both patients and practitioners; it will publish its first
> peer-reviewed articles soon. Lundberg hopes the site will trade on the
> brand name of Medscape, the popular clinical site that launched
> MedGenMed April 9. The journal will be a curious combination of old and
> new approaches to publishing.
> "We don't plan for this journal to be an annual, a quarterly, a
> monthly, or a weekly. We will publish articles as they are found to be
> of value. The date of publication will be the day it goes up," Lundberg
> says.

This is not news. It is already the policy of hundreds of e-only journals, such as Psycoloquy (and many others).

You are here pitting something fairly humdrum and unenterprising against something revolutionary, as if they were somehow either on a par or alternatives. They are neither. E-only journals are one thing, free archives another.

The only common point is FREE e-only journals: Is Lundberg's free? If not, then it's just S/L/P barriers all over again, in a new medium. If it IS free, then it is probably premature -- like the new author-page-charge based Institute of Physics (IOP) free-only journal, which will, I fear, fail, because the culture is not yet ready for it: E-only journals can't be financed up-front until (1) the community has, and becomes addicted to, the journal literature for free, online, hence (2) S/L/P cancellations occur, freeing a portion of those savings to pay for (3) up-front charges for quality control.

In other words, till the fields are first softened up by subversion, authors will neither see the point, nor have the institutional support, for up-front expenses; besides, page-charges have a bad reputation today, having added insult to injury as an ADDITIONAL expense, over and above S/L/P tolls, in the paper era; and they still have the smell of vanity press. All this will change, but subversion and its ensuing changes in user culture must come first.

http://amsci-forum.amsci.org/archives/september-forum.html

But the point is that Lundberg vs. E-Biomed is an unbalanced and incoherent opposition.

> On the other hand, he will avoid the so-called "fluid" peer review
> with which some E-journals have experimented. "We intend to use a more
> traditional form of peer review-traditional in the sense of shielding
> the identity of the reviewers from the authors and readers and not doing
> open peer review, whereby you put up something ... and get everybody to
> shoot at it." Lundberg feels that such approaches dilute the authority
> of a journal.

These are all platitudes. Most of the new e-only journals (e.g. Psycoloquy, for over TEN years already) use classical peer review, with anonymity, etc. Why parade these platitudes in the same breath as truly new and potentially revolutionary stuff?

> Hosting such "fluid" documents can actually strengthen an
> electronic publication, argues Rick Luce, [**TITLE**], at Los Alamos
> National Laboratory. "One of the things that the medium clearly can do
> is turn static documents into living documents." That approach may
> explain why physicists have embraced the Internet as a research tool.

I assume you are referring here to open peer commentary (as opposed to classical peer review): This too is one of my specialities. The PAPER journal I founded in 1978 (Behavioral and Brain Sciences, BBS, published by Cambridge University Press), specializes in peer commentary. And, yes, the online medium is infinitely better suited to that, and will make a lot more forms of peer commentary -- formal and informal, refereed and unrefereed, on unrefereed preprints and on refereed reprints -- possible and permanent.

http://www.princeton.edu/~harnad/bbs/index.html

But this is a side-issue! Commentary is a supplement, not a substitute for the classical system. And what is at issue here is the classical system, that is, the 14,000 refereed journals presently constituting the journal literature (or the 6500 subset of them covered by ISI). THAT is the literature that subversion could make online and free. Save the frills for another story; let this be about delivering the quality controlled goods such as they are.

http://citd.scar.utoronto.ca/EPub/talks/Harnad_Snider.html

> The success of the Los Alamos preprint resource may have inspired Varmus
> to draft his own proposal. However, publishers in the two disciplines
> have different preprint philosophies. Physicists have used preprints
> long before the Internet to communicate rough ideas and then, with
> feedback from others, to shape those preprints into publishable papers.

True, but this isn't just about preprints any more. Nor does Los Alamos contain only preprints!

> In biomedicine, on the other hand, some of the most respected journals
> will not touch an article that has existed as a preprint in any
> form-sometimes even on a personal Web page. "There are plenty of
> biomedical publishers who won't accept a paper if it's been mounted
> anywhere," notes Atkins. But that, too, may change, especially if
> E-biomed shapes up as Varmus envisions.

Indeed it will change, as this policy, quite simply, has no justification whatsoever; it is purely self-serving -- with the very minor exception of papers whose unrefereed dissemination might endanger public health: these are a minuscule subset of the biomed literature and can be treated as a special case, as E-Biomed is in a position to do; they are certainly no justification for holding all the rest of the literature hostage to such restrictions; that is done merely out of publisher self-interest: to protect a revenue stream by not allowing themselves to be "scooped." But once the service provided by journals scales down to quality control and certification -- with no question of S/L/P sales to worry about protecting -- this will become the non-issue that it always should have been: Except where public health might be put at risk by premature publicizing, it is no business whatsoever of a journal's whether or not an author has disseminated a preprint of the unrefereed draft.

http://trauma-pages.com/harnad96.htm

I might add that journals also will "not touch" a paper that has already been published in another refereed journal, and in this they are fully justified: For referees referee for free, and it is an abuse of their services to ask them to referee a paper that has already been refereed and published. But in an online world free of S/L/P barriers there will be no incentive to "re-publish" work that has already been quality controlled and certified by some journal. It is already in the public eye, as accessible as anything else. HERE is where comments and citations from peers can draw attention to a paper that might have appeared in a journal that was lower in the prestige/impact hierarchy than it might have deserved to be. Peer commentary can help correct the oversights of classical peer review, but multiple submission, being an abuse of a scarce resource (referee time) will be as unacceptable with free e-only journals as it is now with S/L/P paper journals. Nor will page charges make it any more acceptable; for referees are, and will remain unpaid: there is not money enough in the world to compensate them for their heroic services, donated gratis to a prestigious journal or granting agency by reason of an academic golden rule.

> However, E-biomed might have the opposite effect. Large
> publishers will continue to deny publication of material that appeared
> as a preprint, will resist giving up copyrights, and will do whatever
> they can to charge for full-text articles pulled from the Internet. "I
> think the focus has been 'Protect the revenue stream,'" Luce notes. "If
> you and I were the journals, we wouldn't want to go along with this,"
> Harnad agrees.

I agree. But now we come to what I have called the "Faustian Bargain": There is a profound conflict of interest in this, one that is unique to the refereed-journal literature (it is NOT true of books or magazines), and that places research and researchers on one side, and publishers on the other.

http://www.arl.org/scomm/subversive/toc.html

The Faustian Bargain in the past was that all authors assigned copyright to their publishers because that was the only way they could gain the immortality of PUBLICation. This was fine for the trade literature (all books and magazines), because those authors contributed their texts for fee or royalty, and shared in the take from the toll-gate receipts. But this was never true of the refereed journal authors, who wanted only to reach the eyes and minds of their fellow-researchers with the reports of their research findings, so their work could have its full potential impact, and be built upon as broadly as possible.

Yet they too had to assign copyright, because there was no other way to cover the real expenses of paper dissemination. The access-restrictions imposed by the toll-barriers were against their interests, but the only alternative was even worse, namely no access at all.

ftp://ftp.princeton.edu/pub/harnad/Harnad/HTML/harnad90.skywriting.html
http://www.cogsci.soton.ac.uk/~harnad/Papers/Harnad/harnad91.postgutenberg.html

In the online era (what I've called the "PostGutenberg Galaxy of Scholarly Skywriting") this is no longer true. There IS a way of covering the much tinier expenses (of quality control and certification) without the need for any access barriers.

http://www.princeton.edu/~harnad/nature.html

And if, in this newly unveiled conflict of interest -- which could never be resolved in any other way in the Gutenberg era, but now can be -- the publishers insist on continuing to impose the trade model, with its S/L/P-barriers, when it is no longer necessary, then (and only then), the research community is in an excellent position to bolt -- for of course we are not only the authors and the readers, but also the referees and the editors.

I do not believe it will come to this, however, which is why I advocate subversion rather than confrontation or defection. For the copyright laws in this Faustian domain -- where authors don't WANT to be protected from the theft of their own texts! -- is not only unjustifiable, but also unenforceable. Everyone can post a preprint: Will journals be sending virtual agents around trawling for preprints 24 hours a day all over the Net, to compare with all incoming submissions? How alike must 2 texts be to count as the same draft? And once the preprint has been refereed and accepted, will the publishers then trawl for lookalikes on the Web again? How different does the self-archived version have to be from the accepted final version in order to count as just another preprint, rather than a "reprint."

This is all nonsense, of course, because there is not only no logical or practical basis for making such distinctions when the AUTHOR wants to give it away (there are plenty of bases for it when it is another author or publisher who is trying to steal the text-AUTHORSHIP rather than the text, but that's plagiarism, and not what's at issue here), but there is also absolutely no justification for it: It is against the interests of both research and researchers to try to enforce such arbitrary strictures in the PostGutenberg Galaxy -- for THIS special, give-away literature (often publicly funded already, with publication mandated by the funder).

> Links to Success
> Large commercial publishers see E-biomed as a threat and a
> challenge. "As written, the Varmus proposal almost encourages a
> reduction in the number of journals available to authors," comments
> Brian D. Crawford, vice president and general manager, life and medical
> sciences, at John Wiley & Sons of New York. Karen Hunter, senior vice
> president of development at Elsevier Science, agrees. "If it's intended
> to replace journals, I think that's a concern."

That's just because of the vagueness of the initial draft of the proposal. As soon as it is brought more into focus, and the inessentials and incoherencies are dropped, it will become clear that SELF-ARCHIVING the entire literature is what is at issue, and this entails no reduction whatsoever in the number of journals: It is intended to reconstitute every single one of them (via author self-archiving) online, and for free, thereby ushering in the optimal and inevitable outcome of this process, and encouraging the publishers to restructure themselves so as to continue providing a useful service to this new, smaller niche.

It is the false impression that E-Biomed is trying to spawn a new breed of rival journals that has gotten publishers' hackles up, but this will be remedied in the next draft. The subversion, on the other hand, will continue to be inherent in the project, as it should be, but that is not something against which either a logical or an ethical or even a practical case can be made: It must be tolerated by the publishers, and adapted to. There is neither a means nor a justification for trying to stop it.

> Both Crawford and Hunter
> agree with Harnad and Varmus that the publisher's strength is the name
> and reputation of its journal. Both publishers are trying to boost both,
> by taking advantage of interactive communication. They're building links
> to other references, adding sound and animation to Web publications when
> appropriate, and hosting online discussions-enhancements that, in many
> cases, first appeared in electronic preprint, reprint, and
> E-journal-only formats. Crawford and Hunter use the term "value-added"
> to describe those features.

Ah me, the "value-added" argument. Here is the quick rebuttal:

ADD-ON enhancements for a fee are just fine. Add them and then try to sell them. But do not try to hold the refereed article hostage to those add-ons: Let a generic, quality-controlled, certified draft be self-archived for free, and then continue to try to persuade the user community that they are better off with an enhanced version, with add-ons, for a fee (S/L/P).

My prediction is that the user community will prefer the free, no-frills version. Then, and only then, will publishers realize that there is no hope of sustaining S/L/P barriers, and will scale down to up-front payment for peer review and certification.

Note, however, that subversion is an end in itself either way: The goal is to free the refereed literature. Self-archiving does that. WHETHER or NOT a parallel S/L/P version proves to be sustainable, the goal will already be attained by providing the free version.

> That potential added value could well spring from the
> competition of the other electronic information sources. And
> noncommercial sites, such as Stanford University's Highwire Press and
> Los Alamos' experimental "Library Without Walls" have been adding
> similar features. Highwire Press, founded in 1995, mounts 127
> high-impact science and technology journals. Highwire has for years been
> adding many of the hyperlinked features that commercial publishers are
> now exploring. The "Library Without Walls" project lays one search
> engine over a multitide of databases, including PubMed, ISI's Web of
> Science and other massive journal repositories.

This is all highly non-revolutionary stuff: It is merely about driving S/L/P prices down. Subversion is about eliminating them altogether, to produce a completely access-barrier-free literature for one and all. Cheaper S/L/P will solve some researchers' access problems, but freeing the literature will solve everyone's, everywhere. That's a difference between night and day.

> The competition between the noncommercial sites and the
> commercial ones will likely increase as a result of the E-biomed
> controversy. "Some of that conflict, frankly, is healthy," Luce opines.
> "What I see are two spheres of where papers go or where you might access
> literature. One I'll call an informal sphere-which would include things
> like preprints and more informal communication. The other I'll call a
> more formalized sphere. And that would be where there's very careful
> peer review." That formal sphere touches both the public and private
> sectors, because online journal publishing is not now one single thing.
> And perhaps, E-biomed notwithstanding, it never will be.

This is all exceedingly murky, and based on the vagueness of the current draft of the E-Biomed proposal. The real categories are these: Free self-archived preprints AND reprints vs S/L/P-toll-based reprints. That's all! It's not preprints vs reprints, informal vs formal, peer review vs peer commentary. That's all just fog and confusion.

> Medscape General Medicine (MedGenMed)
>www.medscape.com/Medscape/GeneralMedicine/journal/public/mgm.journal.html
>

> Elsevier Science
> www.elsevier.com
>
> Highwire Press
> www.highwire.org
>
> Los Alamos "Library Without Walls" project
>lib-www.lanl.gov/lww/welcome.html
>
> Los Alamos Preprint site
> xxx.lanl.gov
>
> Stevan Harnad homepage
> (including links to "Cogprint" cognitive science reprint site; the
> "subversive proposal"; and discussions and essays about electronic
> publishing)
> www.princeton.edu/~harnad
>
> Wiley Interscience
> www.interscience.wiley.com
>
> [Q&A HED]
> Varmus Seeks Societies' Support for Electronic Journal
>
> On May 5 Harold Varmus, director of the National Institutes of Health,
> unveiled a draft proposal for E-biomed, an electronic publishing system.
> The plan sketches out several routes for a U.S. government-backed
> system. It includes provisions for electronic preprints, perhaps
> resembling the Los Alamos National Laboratory-hosted physics site, as
> well as original publications, perhaps matching newer E-only journals,
> such as MedGenMed, or existing print journals' electronic versions. In a
> recent conversation with News Editor Paul Smaglik, Varmus hinted that he
> seeks to cooperate with society journals and perhaps compete with
> commercial ones. The following interview has been edited for length and
> clarity:

The commercial vs. learned-society, bad-guy vs good-guy dichotomy, is at best simplistic, at worst simply erroneous. The big, successful Nonprofits, whether Learned-Society or University, are virtually indistinguishable from the Commercials in their means, ends, policies and prices. Examples are the American Chemical Society and the American Psychological Association and, for that matter, the American Association for the Advancement of Science (which has one of the most regressive copyright policies).

Yes, the Learneds and Universities are more likely to come round once they smell the subversive coffee, than are the Commercials; but at the moment, it is as little in their immediate interests to do so as it is for the Commercials. So there is no point expecting a priori cooperation. The APS (Physics) is the most progressive, but note that that is only AFTER 8 years of resoundingly successful subversion by Los Alamos!

No: Self-archiving is the way, and not some attempt to separate the good guys from the bad guys and make a deal before there is any de facto pressure to make a deal.

> Q: How will E-biomed affect existing publishers?
> A: I think it's too early to say what the impact will be. We have to
> distinguish among the various kinds of publishers. We proposed this to
> stimulate discussion and to move toward developing a system of
> publishing that does at least three things: take advantage of the
> flexibility in electronic publishing; move toward providing full,
> unfettered, and seamless access to the entire biomedical literature; and
> [create] a system that has the ingredients for evolvability in various
> directions. Things are changing very rapidly, and it's important that we
> have a system in which peer review and copyright holding and different
> means of raising money can evolve along with the mechanics of the
> system. We're obviously concerned about how much this methodology will
> cost and how it will be paid for.
>
> We have to distinguish between two kinds of publishers: publishers that
> are private, profitmaking organizations and those run by societies that
> represent the interests of thousands of scientists. We've been working
> largely with some of the societies to discuss some of these issues, and
> we are certainly aware of and concerned about the possibility that
> revenues could decline for scientific societies. I think we have to have
> a culture change here that may take a while to develop; that is, people
> who are potential members in scientific societies-scientists like
> me-have to be reminded or taught that societies do a lot more for their
> members than simply give them cheaper access to a few journals.
> Societies lobby effectively for the concerns of scientists and many
> political and cultural venues. They run important meetings, they worry
> about the future of minority, female, or young scientists, and they
> advocate usefully on behalf of all those folks. They develop a sense of
> community within a scientific discipline. If any individual society's
> journals were no longer purchasable, but instead just available on
> E-biomed or some similar site, it would not be an issue of deprivation.
> Indeed, the societies should be showing their members that they are
> participating in a new wave of more useful dissemination and
> presentation of information. No one is being deprived of anything by the
> absence of a cut rate on a weighty journal that can now be accessed much
> more easily through the Web.

I don't think it's all about lowering S/L/P prices. It's about no longer holding the report of research hostage to S/L/P at all (not even in the interests of supporting learned societies' other "good works": their best work would be to free the research literature of needless and obsolete access barriers at last, to the eternal benefit of research and researchers).

> Q: How important is the issue of who holds the copyright?
> A: Some people, I think, overemphasize that. I think it's significant,
> and I would prefer to see authors hold copyrights-they've written their
> proposals that way. But it would not bother me if we had editorial
> boards that participated in E-biomed who wanted to try to hold
> copyrights while others didn't. I think that's where evolvability has a
> major role. People who feel strongly-there are probably a lot of
> them-will say, "I'll submit my electronic publication to an editorial
> board that has high visibility and is highly respected and also allows
> me to retain my copyright." And others might say "I don't really care"
> and may go with another board that wants to hold their copyright. We'll
> have to see how important that really is.

Alas, this is still part of the vagueness of the first draft. People will continue doing EXACTLY what they have done till now, which is submitting to the highest quality/impact journals in their subject area. Copyright is critical ONLY inasmuch as it attempts to block free self-archiving. THOSE are the substantive issues.

E-Biomed is not and will not be a journal or journals. It will be a free archive for self-archived preprints and reprints, with the possibility of official "overlays," in which the paper is authenticated by the journal itself. That is merely a matter of tagging and encryption. There is still equivocation to be carefully resolved here about just what E-Biomed itself is meant to be and to deliver. I am suggesting that the coherent core is a reliable, permanent, useable infrastructure for self-archiving by authors, with the option of authenticated overlays by journals, if and when they are ready for it (as the APS is already with Los Alamos).

> Q: How closely will the final structure resemble the Los Alamos model?
> A: That's been the preprint system. Our proposal ... is built in a way
> that would allow our community to either have that system or to have a
> very high proportion of postings be reports that have been reviewed,
> edited, and stratified by traditional hierarchies. I think the way our
> culture works now-given its size and the number of publications-we're
> going to remain in the camp of reviewing and identifying journals with
> different status.

Still far too vague on the critical essentials: Los Alamos is NOT just a preprint archive. It is a SELF-archive, hence authors can put in whatever they like, preprint or reprint, and they do.

So THAT barrier has already been crossed. Exactly the same should and will be true of E-Biomed. Authors can, as a first approximation, self-tag their refereed journal reprints as such. That's good enough for subversion. Once the user community is addicted to E-Biomed as the locus classicus for the journal literature, instead of the S/L/P corpus, then "official" authentication overlays will come onto the horizon, but not before, or a priori, for (apart from newborn journals, which are irrelevant), for established journals such an a priori arrangement would be to shoot themselves in the foot before they had even had a chance to make alternate arrangements to restructure for the free Skywriting era.

> Q: What about cultural differences between the physics community and the
> biomedical community?
> A: [Physics] has a hundred authors writing one paper and they don't
> publish as much; the need to stratify the literature by hierarchy and
> status is less of a problem. But there are some things that my lab does
> and I'm sure other labs do that might not ordinarily constitute a
> publication. But it's useful information if I can deposit it somewhere.
> It could be a conversation, a posting, say, that my other colleagues who
> work on wnt genes might want to see; I could never put it into a
> reviewed manuscript with space constraints. Nevertheless, it might be
> useful. And rather than put this into my own Web page, which everybody
> would have to consult one by one, it would go into a central repository,
> which a search engine could pick up.

Yes, yes, the preprint sector will hold many treasures. But the essence of it all is the reprint sector, which will free the entire biomedical journal literature.

> Q: What do you make of journal publishers' arguments that their major
> asset is their status, their seal of approval?
> A: I agree with that. I think people have gotten the idea that because
> we're proposing alternative routes, we don't value editing-I spent a lot
> of time as a scientist editing and reviewing. And I believe that's
> useful. I do think that, in the last few years, because there have been
> so many manuscripts submitted and because there's been such a tight view
> of the hierarchy, that people spend an awful lot of time revising
> papers, sending them to different journals. This is a very inefficient
> process, which we ought to be able to make more effective. I think all
> of us have a lot of quite significant papers that spend a year bouncing
> around and undergoing fairly minor corrections before our colleagues can
> see them. I don't like that. I'd like to see my colleagues' work earlier
> and I'd like them to see mine sooner. I think we all can envision ways
> in which the process could be speeded up in electronic format.

Again, too vague. Self-archiving unrefereed preprints solves part of this, and is highly desirable and commendable. But there is no need to speak about any of the problems of peer review here -- either its quality, its efficiency or its timing. There ARE such problems, to be sure, but they are not to be conflated with the project at hand, which is to free the peer-reviewed literature (such as it is!) for one and all.

There's room for projects to improve peer review, speed it up, and what have you. But let us not LINK the fate of the clearcut and eminently desirable goal of freeing the literature with the more hypothetical and conditional one of trying to improve peer review. Such issues should be disentangled completely from the plans for E-Biomed or they will simply raise needless opposition from the defenders of classical peer review, or worse, will make the prospects of a free literature -- already highly desirable a priori -- depend on the prospects of various peer review reform schemes, schemes which may or may not prove successful (and certainly require a good deal of prior empirical testing before being implemented at all, let alone implemented en masse).

Cheers, Stevan

--------------------------------------------------------------------
Stevan Harnad
Professor of Cognitive Science
Department of Electronics and Computer Science
University of Southampton
Highfield, Southampton
UNITED KINGDOM
http://www.cogsci.soton.ac.uk/~harnad/
http://www.princeton.edu/~harnad/
ftp://ftp.princeton.edu/pub/harnad/

Paul Ginsparg, May 29, 1999

> Date: Sat, 29 May 1999 15:18:40 +0100 (BST)
> From: Stevan Harnad
> To: Paul Smaglik
> cc: "Harold Varmus (NIH Director)" ,
> Gene Garfield ISI

> > electronic publishing as a whole. The field has splintered into a myriad

note that researchers are not really interesting in "electronic publishing" per se, that notion builds in the idea of carrying over all the inadequacies and artifacts of the paper methodology into the electronic realm -- what we're really building towards is some sort of knowledge network for the sciences, very different from what the publishers have in mind by unnecessarily perpetuating the current partitioned research database

> Drop this in favour of the prior mention above. It is artificial and
> counterproductive to try to distinguish preprint and reprint servers.

yes, the official name here is the "Los Alamos e-print archive", where the current definition of "e-print" is something self-archived by the author (in accord with stevan's comments later on, e.g.: > Los Alamos is NOT just a preprint archive. It is a SELF-archive, hence
> authors can put in whatever they like, preprint or reprint, and they do.
)

> > The drawback? Copyrights. Many journals do not let authors
> > retain copyright. And commercial and society-based journals likely
> > wouldn't voluntarily give up the subscription income that makes them

in this context you could also mention our policy forum from Science 281 (1998) 1459, http://www.sciencemag.org/cgi/content/full/281/5382/1459, where we argued that agencies funding research require that author retain copyright as a matter of public policy interest for publicly funded research.

> > Hosting such "fluid" documents can actually strengthen an
> > electronic publication, argues Rick Luce, [**TITLE**], at Los Alamos

Rick is the director of the Los Alamos research library, and as such head of the library without walls project that you mention later on. on the other hand, he currently has nothing to do with the e-print archive which is not operating out of the library here, but instead from the theoretical division (where i started it in 1991).

> > The success of the Los Alamos preprint resource may have inspired Varmus
> > to draft his own proposal. However, publishers in the two disciplines
> > have different preprint philosophies. Physicists have used preprints
> > long before the Internet to communicate rough ideas and then, with
> > feedback from others, to shape those preprints into publishable papers.
>
> True, but this isn't just about preprints any more. Nor does Los Alamos
> contain only preprints!

actually even the premise is not entirely true. the preprints that physicists circulated prior to the internet were not intended to communicate "rough ideas" , they were invariably manuscripts simultaneously submitted to journals, and a significant percentage appeared later in journals (i.e. after the referee process) in a form identical to the original "preprint" (perhaps with a few typos added by the publication process). "rough ideas" were communicated in seminars or talks at conferences, not in preprints.

> Los Alamos Preprint site
> xxx.lanl.gov

Los Alamos e-print archives

> > Q: What about cultural differences between the physics community and the
> > biomedical community?
> > A: [Physics] has a hundred authors writing one paper and they don't

> > publish as much; the need to stratify the literature by hierarchy and

i'm sorry, but this is ABJECT NONSENSE, and no sensible person would wish to be quoted saying that in this context. there is only one small area of physics in which this is true, namely high energy experimental physics, but this involves well under 1000 papers per year, and only a small percentage of those have "hundreds of authors", hence under 1% of physics papers have "hundreds of authors" and characterizing the entire field in this fashion is simplistic ignorance.
note that the physics archives started not with high energy experimental papers, but with theoretical papers, that in general have *fewer* authors than biomedical papers, and written by people who publish just as frequently as biomed people. the real distinction here is between experimentalists and theorists (with the theorists typically more computer literate and forward-looking), and in biomed there are fewer of what we would regard as pure theorists, more the sort of cultural difference at issue here.

pg

Stevan Harnad, University of Southampton, May 28, 1999

On Tue, 18 May 1999, Gene Garfield wrote:

> 2. Peer-reviewed depository - published papers.

> Research reports submitted for publication in a
> Current Science Group journal (irrespective of
> whether they have been placed previously in the
> pre-print depository of that or any other journal)
> will be fully peer-reviewed in the traditional way.
> The accepted research reports will be placed in the
> peer-reviewed depository and will be available free
> of charge to any individual through the web.

Hi Gene:

Here is the fatal flaw (I think) in this approach: You are setting it up EITHER to compete with the classical journals ("submit there or here"), a competition you are certainly going to lose, OR to "collaborate" with conventional journals ("submit there AND here"), a collaboration you are also going to lose, because classical journals are not interested in publishing what has been published elsewhere, and they are insistent on retaining exclusive copyright over what they themselves have published.

(The latter will change, under pressure from author self-archiving, because the author has all the moral rights here, but it will certainly NOT change under pressure from a competing peer-reviewed journal!)

And, last, you are (wrongly) assuming, in a world where referees are already hard-pressed and doing what they do only because of the golden rule, that there is a large enough set of willing/able peers out there to referee a paper not only ONCE, but TWICE (one set for the classical journal, another set for yours).

> Vitek Tracz, Chairman of the Current Science Group,
> said:
> "We believe science publishing is on the brink of
> some of its most exciting and far reaching changes.

It is indeed, but it is important to realistically sort out just exactly what those changes will be, and why they will happen. The core insight is that journal authors, because they want to maximize access to their work, will self-archive it online, freee for all, and the user community will of course prefer that free version. This in turn will change the market for journals, and will force them to resturcture themselves. But this all involves a reform of the classical journals; there is no niche for rivals here unless the classics fail to reform in the face of self-archiving and give up the game to other players. But that's a bit of the way down the road and I doubt that that will be the way it goes.

> #3
> Dear Steve: I have finally had a chance to take a brief look at the Cal Tech
> proposal and I have been trying to follow your various arguments reported in
> Sci. Amer.,Nature,etc.
> The Scientist has an issue coming up shortly concerning publishing and I
> would like to see a commentary by you, if possible, that could state in 1700
> words or less your basic ideas and make very few assumptions about what
> readers know, even though they are mainly research scientists.
> You can cut and paste or whatever to make a coherent piece and if you don't
> have the time to make the deadline we can schedule it in another issue. This
> topic will not go away and we will have to continue to educate the
> scientific community.

I'll be happy to do it: When's the deadline?

> As you know anything you publish with us will go up on the web free of
> charge and we would encourage you to include links to appropriate
> archives,etc. for those who want the full background.

Fine.

> I think there will be significant objection to having all this run by Los
> Alamos and even for that matter NLM were they to adopt the Ginsparg
> programs. Having it in government hands is problematic.

That all depends on what you mean by "all this." I am now coming to realize that the idea expressed in my own Subversive Proposal of 93-94, which I then thought was just one hypothesis among many, is turning out to be EXACTLY what will and needs to happen. Self-archiving is the name of the game, not rival journals, or rivals to the journals. So all Los Alamos and NIH and others are providing is the infrastructure for universal self-archiving. In parallel, the universities (Scholar's Forum) are ALREADY offering it, in the form of authors' local home-institution servers. The Los-Alamos/E-Biomed approach is a global one. But in reality, the result will be a virtual library that seamlessly integrates these two approaches, with one serving as a reliable backup for the other.

On the contrary, governments and big research institutions are PRECISELY the ones that authors know and trust, and will hence entrust their self-archived papers to. They are JUST the ones to provide the infrastructure, especially if it is suitably distributed, so as to assuage any fears of monopolism or imperialism.

> Have you ever talked to Vitek Tracz? He is in London and I have urged him to
> contact you. I will try to send you his web site announcing his plans.His
> compnay is called CURRENT SCIENCE.

Not yet, but I'd be happy to: What's his email?

> Your comments on the role of citations are relevant and I wonder if you have
> seen the search engine developed by NEC of Princeton?? I think it is
> www.citeseer.com but I will also check that later
> and make sure you get the right address. They have developed what they call
> "Autonomous Citation Indexing" which means they take anything that is on the
> web and create all the citation links and display the citing context. The
> fellow is named Steve Lawrence, an Englishman I believe.

I know the project, and know Lee Giles, his collaborator, from my years at Princeton; indeed, we cite their project in our own proposal (which will be formally announced by NSF/JISC next week). We will be in contact with them.

> I think you somehow have avoided the issue of redundancy. I think you may
> argue too strongly against the idea of "virtual journals" assuming that
> profiling will do the trick.

Not sure what you mean. I am a STRONG advocate of redundancy: backups, mirrors, distributed archiving, local and glabal archiving, interoperability, the lot! I'm also for "virtual journals," in the sense that the final, refereed, accepted paper is redundantly archived publicly, and the role of the virtual journal will have been to peer-review it and, if accepted, to certify it as such with the journal brand name.

What I don't agree with is needlessly trying to compete with the classical brand-name journals now. That will just fail, and so it should. What authors should do now is NOT to switch from the high-impact, high-quality brands they know and trust; they should stick with them, but subvert too, by self-archiving. THEN let the chips fall where they may.

> However, I think that you as an editor should
> know that journals are not just aggrtegations of closely related material.

Of course not. And the items in an issue rarely have much to do with one another. But a known, trusted, quality-assurance brand continues to be what authors (and readers) want, and there is no reason why it should not continue to be.

> If you are talking about invisible colleges then you are probably right.

College-schmollege: The peers are the college, and each journal (and granting agency) has its own subset of them; that is medium independent.

> Derek Price use to say that 200 scientists could support a journal, but even
> that number is too large for the kind of journal you are talking about. You
> are right in the sense that one can develop a profile as I have with the
> ASCA system for 35 years(now called Research Alert) but my profile includes
> dozens of Source journals as a kind of mini-Current Contents. If everything
> gets poured into one barrel I myself would worry about browsing such a huge
> collection. And knowing how scientists work I do not think they will
> maintain their profiles properly without the kind of rubrics that journal
> titles represent.

I think you have misinterpreted me on this point: I'm for MAINTAINING the diversity and hierarchy of journals; that's one of the things I was criticizing E-Biomed for, rather than advocating it. But I'm also against trying to COMPETE with the existing journal hierarchy right now, when what we need is subversion (by self-archiving), which we can win, as Los Alamos has proven, rather than competition, which we can only lose.

> Of course this will have to be tested, but I would urge you not to push that
> point too strongly, especially in the transition. Perhaps if the system
> includes a scheme for co-citation and word clustering we can create
> "virtual" journal contents for browsing.

I'm not sure what point you're talking about. In the citation proposal, we will interlink everything in Los Alamos. That does not make it into one big journal, it makes it into a better interlinked ARCHIVE which contains all the brand-name journals, suitably tagged (and free).

> What has happened to the Links Project at Southampton?
> How often do you come back to Princeton?
> Best wishes. Gene

The ISI links project was taken off-line, at ISI's insistence, after only a few months. A pity, because I would have liked to leave it on-line much longer, to give more people a chance to taste it.

It is understandable that ISI is ambivalent about this, though, for it takes only a little imagination to deduce that if there were a free, full-text archive of the journal literature, the citation linking possibilities and benefits would be much greater than the mere linking of ISI abstracts plus reflists (even if paired with click-through, pay-as-you-go links to journals' full-text proprietary data bases).

Do you see the fundamental rift between for-fee and for-free in this new world?

Cheers, Stevan
--------------------------------------------------------------------
Stevan Harnad
Professor of Cognitive Science
Department of Electronics and Computer Science
University of Southampton
Highfield, Southampton

Dale Purves, May 28, 1999

Harold,

E-BIOMED is a great idea. A transition to electronic publishing of this sort is inevitable, and having the process occur in a deliberate manner under the auspices of the NIH is certainly the right way to do it. As Editor-in-Chief of the JOURNAL OF NEUROSCIENCE during the late 80s and early 90s, I know full well that the list of advantages for E-BIOMED -- and the cost savings -- could have been put much more strongly than in your prudent and politic proposal. Let me know if there is anything I can do to help things along; I would be glad to do so.

Best regards,
Dale Purves

David Goodman, Ph.D, Princeton University Library , May 27, 1999

The Biology e-print server--is there a consensus?

There is general agreement about the need for a replacement or major change in the existing system of biomedical journals, and also general agreement that a system based on a e-print server such as LANL is an appropriate alternative. However, there remain some fundamental differences about the optimization of many of the specifics.

As a relative outsider to the details of these discussions, I suggest that it may not be appropriate to be concerned initially about defining or achieving the ideal system of scholarly publication, and that it would be more appropriate with our limited current experience to have the more modest goal of a generally acceptable initial step.

The criteria for what this should include should emphasize the areas on which there is agreement, leaving the necessary flexibility for the as yet unresolved areas, and providing a simple and inexpensive system that could be rapidly initiated with a minimum of special funding or administration. Otherwise, while we are trying to find the best way, they system will continue to deteriorate to the point of true collapse. Though this will certainly force us to adopt an alternative--any alternative--it is not likely to lead to an optimum one.

As I see this discussion, the area of general agreement is that the fundamental distribution of material should be via an e-print mechanism very similar or identical to LANL, operated on a cooperative basis. Fortunately, this is know to be achievable, is known to be inexpensive, seems to be fully adequate to future extensions and adaptations, and does seem to be acceptable to all the various parties--even the commercial publishers seem to be able to live with it. This is the basic portion, and perhaps all that should be initially included.

The other portions, which involve areas for disagreement, include the need for linked mechanisms for peer review or other alternatives for quality control,
the need for rudimentary quality control upon entry,
the need for formal archiving. ` the possible role of print publications,
the roles of commercial publishers, professional societies, and universities,
the mechanism of funding via subscriptions, page charges, or hidden overhead,
The widely divergent but equally assured views on these matters are perhaps evidence that the best way to deal with these issues is to postpone them, while ensuring that the developing system will be open to any of the suggested solutions.

The peer review mechanism proposed by Varmus would seem a very good and practical one, but it is certainly not the only possible mechanism; since it does not seem to command universal assent--essentially everyone has an individual choice of a favorite mechanism--it might be advisable to implement this as a later optional addition, along with other alternatives.

The proposal for a initial quality filter does not also seem to have universal assent. I rather doubt that the lack of need for one in physics will prove the case in biology. But if it proves necessary, it can be added, at a level which will exclude only truly irrelevant material--I suggest the mechanism in the Varmus proposal may prove unnecessarily elaborate.

Almost all librarians, am distrustful of the long term usability of a system without formal archiving arrangements. But since many of the scientists regard this as excessive caution, perhaps the existing stability provided by the mirroring of the LANL system is enough to start with. Print is not being immediately abolished, and we will have opportunity to devise more permanent long-term solutions.

It is apparently totally uncertain what the effects of this system will be on print journals, and on the commercial and society publishers. All possible views are represented: those who foresee and perhaps intend that the conventional publishers will have no future role,
those who see them becoming at most merely editing and reviewing bodies, those who accept a limited role for conventional publication for certain classes of material,
those who see a dual system, with the e-prints simply a preprint mechanism and the conventional paper and electronic journals continuing as before,
those who see e-prints being incorporated into the conventional system and operated by the established publishers.
The existing experience in physics is hardly decisive: the print journals and their formal electronic equivalents continue, though they are less used. This confirms the desirability of being able to adapt to any of the technological and cultural developments.

Some publishers may refuse to permit papers to be published in their journals if they are previously included in the system. This has not proved to be the case in physics, but there the most prestigious journals are published by a professional society, not a commercial publisher, whereas in molecular biology, many of the most prestigious journals are now part of a major commercial publisher. Similarly, some publishers already insist that only the preprint version, not the version as published, be included in the server. I suspect that either approach by a publisher will prove counterproductive, and authors will prefer those publishers who permit papers to be available in final form on the server.

Some scientists under the existing competitive conditions are uncomfortable with a system without the built-in time delays provided by publication, and it is possible that this may cause some to submit only their less important papers. If this proves to be the case, appropriate optional protection can be added--as a minimum, a paper can be added to the system only at formal publication.

Before proposing elaborate financial or administrative arrangements, it might be better to see if they prove to be necessary. The basic portion of the system seems to be operable with very low direct costs, and it is at least possible that the total cost of the eventual system may be less than at present. The e-print proposal is not likely to fail altogether, but its low utilization would prove very detrimental to the chances of achieving meaning reform. It is important for its success to be rapid, and unequivocal. This can most reliably be achieved by starting with the basic distribution mechanism only.

David Goodman, Ph.D.
Biology Librarian & Co-Chair, Electronic Journals Task Force
Princeton University Library
http://www.princeton.edu/~biolib/

Jack W. Bonney, Vancouver Canada, May 27, 1999

i am interested in pursuing more information regarding your proposal for centralized electronic publishing of bio-medical information.... my key interest lies in what tools you will be using...

traditional wordprocessing tools are nor suitable in many cases because they were never designed for several key issues:

1. huge volumes,
2. automated hyperlink generation,
3. multi-media file types; i.e. text, graphics, audio/video clips, and
4. multi-media production; i.e. paper and electronic formats with the same look and feel.

you may be interested to see that the kind of publishing upon which you are embarking is far more complex than most people realize.... until their project has been started and by then it is too late....

visit the following site for more details on what has been developed in this area (note the link to the book on the subject):

http://www.radiotower.com/ssi-can/

there are two components: the first one, collaborative knowledge development can work with existing "shelf-ware" tools, such as lotus notes, microsoft packages and others.... the second one, deals with the far more complex issues of publishing, namely the databank creation, integration, management, and dispersal followed by revision and updates.... if you have not made a final decision on the publishing "management" tools that you will be using, and would like to find out more on a top-of-the-line solution built specifically for a task such as yours, simply reply with your questions and concerns and i will try my best to answer them for you.

thank you in advance for your consideration and courtesy.

sincerely,

jack w. bonney,
vancouver canada

FTP INTERNET DATA ARCHIVING: A Cousin for PSYCOLOQUY,92.3.29.data-archive.1.skoyles http://www.users.globalnet.co.uk/~skoyles/pdata.html

Dr. John R. Skoyles
London, UK

Check out my Golden House-Sparrow award winning homepage http://www.bigfoot.com/~skoyles

Eugene M. Renkin, University of California, Davis, May 26, 1999

Dear Dr. Varmus:

This is in response to your E-BIOMED draft and proposal of May 5th. I want to express my appreciation of your addressing the future of biomedical publications in this network communications age, and to submit my comments on your proposals.

I believe your plan to extend electronic coverage of bioscience publications back from the dates currently available to electronic search is practical and most commendable. Electronic accessibility of the full text and illustrations of these older publications, which are often overlooked, would be a strong support of modern research. It could be achieved gradually, over several years, at relatively small cost. NIH would probably have to administer and fund this project, because it is unlikely that commercially published journals would be willing to undertake it. I should consider it a suitable use of NIH funds.

I fear your proposal for centralized electronic publication of biomedical research reports would have more disadvantages than advantages. Centralization of editorial boards would surely have a narrowing effect on the scope of reports accepted for publication. And conventional journals of long standing and high reputation would disappear. Current journals are trying to adapt themselves to electronic publication in ways that might suit them individually. NIH would do better (with less expense) to help them in their efforts, and coordinate retrieval of papers via a common search engine.

And as for your suggestion of a repository for unedited material, that would not only be a horrendous mess, but a total waste of time effort and money. We all have our own web pages (or could have if we wanted to), and can print anything we like in them. Besides, we can communicate our exciting ideas and findings to colleagues in our fields via multiply-addressed e-mail.

Again, thanks for the opportunity to comment.

Yours sincerely
Eugene M. Renkin
Professor Emeritus
Dept Human Physiology
University of California,Davis

Note: These are my personal opinions, not necessarily those of my Department or University.

Jerrold M. Ward, DVM, Ph.D., NCI-FCRDC, May 25, 1999

Great idea! At least try it out and see what happens. (woops, i didnt get this statement approved by my division director)

Jerrold M. Ward, DVM, Ph.D.
Chief, NCI Veterinary Pathology NCI-FCRDC

http://www.ncifcrf.gov/vetpath

Sincerely,
Thomas E. DeCoursey, Ph.D.
Professor, Department of Molecular Biophysics and Physiology
Rush Presbyterian St. Luke's Medical Center
url: http://www.rushu.rush.edu/molbio/physiodec.html

Chemistry: http://www.virginia.edu/~chem/faculty/etzkorn.html
Biophysics: http://www.med.virginia.edu/ed-programs/biophysics/fae8m.html

Carolyn Wilson, May 25, 1999

I have reviewed the proposal posted about developing E-biomed for electronic publication of papers and data. There are some parts of the proposal which are potentially of great value to the scientific community, while other portions of the proposal I view with some skepticism.
The first tier of review, using the Editorial Board mechanism and peer review, would provide certain advantages over print media, such as the immediate posting of papers upon acceptance of publication (obviating the 2-3 month, or more, lag time often found with print journals). An additional advantage to this process is the suggestion that additional data or subsequent modifications (in themselves not deeming separate publication), could be appended to the initial publication.

The second tier of publication is less appealing. First, the cursory review of two individuals who would validate the report is subject to a higher degree of cronyism and conflict of interest than a real review with written comments for evaluation by the editor of the journal. One approach that may help to alleviate this problem somewhat is to use the criteria frequently used by many journal and evaluation committees for avoiding conflict of interest: exclude current or former collaborators and mentors, and individuals with financial ties. Even with these additional precautions, this second tier is still subject to becoming a repository for data that is suspect, and could potentially become a junk yard of otherwise unpublished/unpublishable data. It seems that it may not be worth the resources to maintain a huge system to provide access to a large amount of information of a lower standard and of potentially limited use.

An alternative to the approach proposed in E-biomed of having non-peer reviewed data published, may be to have an electronic chat room for informal postings of questions, data, new techniques, etc. These type of postings should not be viewed as publications at all.

A final general comment: if this is less expensive than print journals and scientists can get publications to "press" more rapidly than traditional journals, and eventually with its own citation index, publication in e-biomed will be viewed as prestigious, could this potentially put print journals out of business? If so, is this a desirable outcome?

Ajit Varki, Professor of Medicine, University of California, San Diego, May 24, 1999

The E-Biomed concept has many worthwhile features, but there is a real danger that the baby will be thrown out with the bathwater.

As R. Paul Robertson, wrote on May 18, 1999: "it will inevitably undermine the academic processes of tenure evaluation and grant application evaluation......Would NIH be happy funding a grant loaded with E-Biomed references as a rationale or a progress report? I doubt it. "

An important value of the peer-review system is that it acts as a filtering mechanism to weed through the tremendous amount of biomedical research that is submitted for publication, helping readers to decide which articles to look at first with the limited time available. A secondary value is that it greatly expedites the processes of academic evaluation and grant review. Knowing where the papers were published allows one to divide applicants of any sort into three groups: those that are clearly outstanding, those that are clearly poor, and those in the gray zone that deserve very careful evaluation and a more detailed reading of the actual papers. Eliminating this filtering function will greatly increase the amount of time that we all will have to spend on (a) general reading; (b) Academic reviews; (c) Evaluating candidates for jobs; and (d) grant application reviews.

There is no doubt that the peer review system is far from perfect, and needs a lot of fixing. However, it is not broken enough to justify replacing it a new system that will generate a huge amount of additional work for everyone. I would like to see the best aspects of E-Biomed be integrated with the existing peer-review system, in a manner that also improves the latter.

Ajit Varki
Professor of Medicine
University of California, San Diego

Cass O'Malley, TechBooks, May 24, 1999

Dear Mr. Varmus;

I have just read your essay on the proposal of E-Biomed and as an account executive involved in providing electronic publishing solutions this is very exciting to me! My one and only comment is that you may want to offer a RFP for the composition and tagging (SGML or HTML) for this new service and this may help with costs for all involved because of volume. Please do not hesitate to call if TechBooks can be of any assistance or service. Good luck!

Cass O'Malley
TechBooks
www.techbooks.com