P. Ginsparg and S. Harnad, 5/2 & 5/3/99

SF = Scholar's Forum sh = Stevan Harnad

A major criticism that Stevan makes is that the proposal should really have as its the natural target the authors (or perhaps their sponsoring institutions, whether physical or funding), not publishers; and currently that structure is not so clear.

Another major concern in the proposal itself is the substantial infrastructure required:

SF > Before this is accomplished, research universities must assemble a Consortium to
SF > support the development in implementation of this model.

This will require far too high a degree of cooperation and coordination of interested parties in advance of getting anything off the ground (especially considering that the research enterprise is global, not just a national enterprise), so there must be some way to nucleate instead on a smaller scale and grow from there (or in multiple places and coalesce) rather than awaiting some solution to the chicken-and-egg question.

(And for this to come to fruition I'm afraid one would have to wait roughly as long as for the proverbial group of chimpanzees in front of typewriters to generate the complete works of Shakespeare)

Sadly, authors are not always ready to act in manifest self-interest, and we don't yet know precisely what forms of encouragement will be required in other disciplines. (but we do know something more will be needed, otherwise none of this discussion would be necessary -- all fields would already be well on their way...)

5/3 - SH>> My comrade-in-arms, Paul Ginsparg, is quite right. One can lead a herd of
5/3 - SH>> learned thoroughbreds to the most beneficent body of water, but it is
5/3 - SH>> anyone's guess what will make them drink.

5/3 - SH>> That is why as much of the time and energy of zealots like ourselves must be
5/3 - SH>> spent in ostension and polemics as in concrete steps to spruce up the
5/3 - SH>> communal watering hole still further.

sh > What is undeniable is that the current refereed journal literature….has always been
sh > an invaluable resource to researcher.

"Always" is a long time, and the current implementation of peer review is actually a relatively recent construct. It's not at all obvious that a one-size-fits-all approach is optimal for all disciplines --

5/3 - SH>> Indeed. Peer review is relatively recent (so is present-scale science and
5/3 - SH>> scholarship) and may well be improved upon some day, some way (so might
5/3 - SH>> democracy, and water-quality). But there is no reason whatsoever for -- and
5/3 - SH>> many reasons against -- LINKING the incontestable goal of making the
5/3 - SH>> water (such as it is, quality-controlled according to today's prevailing
5/3 - SH>> standards) available to everyone for free, with the much more indeterminate
5/3 - SH>> goal of devising, testing and implementing new systems of water quality
5/3 - SH>> control. Let us drink our fill for free of what we have so far been able to
5/3 - SH>> drink only for a fee; do not make the fulfillment of that basic epistemic need
5/3 - SH>> contingent on much more speculative schemes for making the water taste
5/3 - SH>> even better!

5/3 - SH>> And peer-review isn't and never has been "one size fits all": Disciplines
5/3 - SH>> differ in how they practise it, and even within disciplines, there is variation
5/3 - SH>> in the quality of journals, and in the rigor of their refereeing systems. That
5/3 - SH>> variation is all part of today's prevailing standard. Our goal should be to free
5/3 - SH>> all of this, online. Reform is possible, but let there be no linkage, and on no
5/3 - SH>> account should free availability become a hostage to the fortunes of any
5/3 - SH>> particular improvement scheme.

certainly many practitioners of my own discipline would dispute the efficacy or necessity of "classical peer review" to the advancement of science. I certainly wouldn't be willing to presume to say for other disciplines either that it simply must be maintained as is or that it must simply be discarded -- that's their business.

5/3 - SH>> Splendid! So don't! Leave it alone, and just give all disciplines the invaluable
5/3 - SH>> gift you have given to Physics: a way to access it all for free, along with the
5/3 - SH>> freedom to choose for ourselves whether to read unrefereed preprints,
5/3 - SH>> refereed reprints, or something in between or beyond.

But it is true that a raw distribution system should have a much better chance to be based on universal principles (though even this can be controversial, with some prominent researchers proclaiming absolutely that their discipline would "never" be attracted to distribution of unrefereed material---never clear if these are representative or just the most vocal, but they certainly exist so can't be dismissed out of hand).

5/3 - SH>> I don't know about universal principles, only about the prevailing water-
5/3 - SH>> quality, and the importance of all scholars' being able to drink their fill.

5/3 - SH>> It's entirely up to individuals whether they want to publicly self-archive their 5/3 - SH>> (1) unrefereed preprints or only (2) their refereed reprints (does any serious 5/3 - SH>> scholar prefer to keep (2) behind S/L/P toll booths?): Vive la liberte! Nothing 5/3 - SH>> hinges on individual decisions about (1); so no one need be detained by it.

It is also true that the details of the raw distribution scheme can be considered both before and at a logically distinct level from whether or how exactly the peer review and other information is layered on.

SF> We are adamant in our belief that "self-archiving" to a single pre-print server by
SF> authors when they submit a work is not building an archive.

Submissions to the LANL archives may turn out to be of greater longevity than that belief (i.e., I wouldn't bet against it just yet).

I am very skeptical of those who commit to see *beyond* the foreseeable future.

sh> I don’t know what "duplicative journals" means, and no one wants to put publishers sh> out of business.

(I believe SPARC does?) (pg)

Looks like simple misunderstanding -- at least some of the standards and protocols mentioned would simply be technical, involving storage and transmission formats and object identification schemes. Even (or especially) a self-archive would need some uniform standards.

Perhaps, but there is a practical matter of making encouraging authors recognize their self-interest and participate. As a matter of historical fact, the LANL archives owe *all* of their initial growth starting in '91 precisely to the daily e-mail alerts of new material received (according to subject matter). While one might have expected the later web access to replace this in some way, it is a matter of empirical fact that users of the archive insist on retaining the e-mail alert as an essential component. It is not supererogatory to mention some of the little details that can magically facilitate the more revolutionary possibilities in the long-term.

5/3 - SH>> Good point. So perhaps I should have said "search/alert" tools and not just 5/3 - SH>> search-tools (although I still consider all of this trivial).

5/3 - SH>> You are right, though, that sometimes an incidental feature, even a non-
5/3 - SH>> optimal one, can turn out to be the critical one that set the cavalry a-drinking. 5/3 - SH>> (But the reason email was the "hook" with LANL in '91 might have been that 5/3 - SH>> the Web was still young and not as widespread then. Who knows? But, fine, 5/3 - SH>> my buzzword will be search/alert henceforth...)


SF> 6. The Forum does not propose to change the peer review

sh> But then why mention peer review at all?

Maybe so they won't be taken to task for appearing to want to change peer review? (oops, can't win, they'll get taken to task anyway...)

5/3 - SH>> The truth is, that the Forum said it was wasn't out to change peer review and
5/3 - SH>> then described speculative structures in which it was changed after all!

5/3 - SH>> (Yes, the only sense in which peer review need be mentioned in connection
5/3 - SH>> with free archiving initiatives is (1) the archive does not provide peer review,
5/3 - SH>> (2) authors can self-archive both peer-reviewed and non-peer-reviewed
5/3 - SH>> papers, and (3) the possibility of official journal overlays is there too, along
5/3 - SH>> with overlays for new peer review schemes and entities, should anyone care
5/3 - SH>> to implement them.)

No one is suggesting megajournal, not sure why that strawperson frequently re-emerges.

Perhaps to make it clear precisely which elements of the current system one expects to maintain in the short term? Not everyone takes it for granted that the current peer review system should or can be maintained without modification, so this explains where they stand on this issue (and if coming down in favor of the concept of classical peer review is necessarily content-free, then others are equally guilty in this regard).

5/3 - SH>> Please re-read the Scholars Forum Proposal. There is a profound ambiguity
5/3 - SH>> about just who/what those "Editorial Boards" are to be. If they are the current
5/3 - SH>> journals (and their successors), there is no problem, and this should be stated
5/3 - SH>> explicitly. If they are meant to be new entities to compete with these current
5/3 - SH>> journals, spawned somehow in conjunction with the Scholars Forum
5/3 - SH>> Initiative, then again we are in the speculative reform business instead of the
5/3 - SH>> business of freeing the water.


It is certainly premature to worry about copy editing in any detail, or of how to create additional infrastructure to certify copy editors...

There are many conference proceedings in the LANL archives. These have always been among the most natural for the archives, with authors supplying "camera-ready" copy to the archives and editors providing the overlay of links that constitutes the proceedings.

SF> 10. The Forum proposes to develop a dynamic alternative to deadening email
SF> discourse surrounding works in servers.

It is true that such platforms already exist, and they're not obviously optimal for scholarly discourse, where one might actually prefer an edited e-mail commentary, or one in which only the threads that actually lead somewhere auspicious are highlighted.

The question is whether the "scholar's forum" is intended to be an implementation plan or a possible end-state model. It is unlikely as the former, since it would have to spring essentially fully formed from the head of Zeus. Most of Stevan's comments criticize this, and he emphasizes instead to focus on the most practical first steps. As an end-state model, ignoring how exactly we could tunnel there from here, it is not clear what exactly will be the incentives for publisher participation. (Stevan's model for supporting the organizers of peer review entirely through page charges at least constitutes a financial incentive structure, though some of the details of how exactly existing publishers [whom he states he wished to see remain in the game, maintaining the same journals as currently] could smoothly get from here to there remain problematic [or else once again, maybe they'd already be moving in that direction...])