I am posting this exchange with my colleague Steve Hitchcock in the hope that it will elucidate the logic of up-front author page-charges and their role in freeing the journal literature for everyone.
Stevan Harnad shi> From: Steve Hitchcock <[email protected]> shi> Open Journal Project shi> Multimedia Research Group, Department of Electronics and Computer Science shi> University of Southampton SO17 1BJ, UK shi> Tel: +44 (01)703 593256 Fax: +44 (01)703 592865 shi> Open Journal Project Web page http://journals.ecs.soton.ac.uk/ > shi> Somewhat daunted by the length I nevertheless read shi> your reply to the... proposals and found it very effective. I like shi> the emphasis on changing one variable at a time. A couple of points I want shi> to clarify. How are the physics 'overlay' journals recovering costs, i.e. shi> charging users for the 'refereed reprints'? NO, definitely not. The American Physical Society (APS) journals continue to charge for the paper edition as well as for the APS On-Line edition (with many added enhancements) that APS sells. But the Los Alamos (LANL) public Archive version is free. This is a critical point. The "overlays" are a service to the authors, in that authors can submit papers directly to APS journals by depositing the preprint in LANL. They can also deposit their final accepted drafts (but not the proprietary APS version, with the enhancements, which only comes from the paid APS Archive) in LANL. As you can no doubt guess, I am betting that Users will be mainly using the free version. If there is still a market for the paid version, that's fine, but irrelevant to the goal of providing the entire corpus online for free, which is achieved as soon as ALL publishers fall into step with the benign and responsible stance of APS. shi> Of course you advocate author page charges, but not exclusively I presume. shi> S/L/P are also possible. Can you resolve possible confusion on this in the shi> following two paragraphs? S/L/P are possible, but I am opposed to them because, by definition, they mean toll-barriers instead of a free refereed journal literature. But let us make some critical distinctions: As long as authors can and do publicly self-archive their unrefereed preprints AND their refereed reprints, the goal is achieved. S/L/P are only objectionable if they in any way prevent that. Of course if there is a free public Archive, then a parallel fee-based one with added enhancements is perfectly fine too, if there is a market for it. So I certainly don't oppose S/L/P for that parallel source. But there is one critical interdependency: In order to appear in refereed form in the free archive, a paper needs to have been refereed and accepted by a journal. How is that quality control to be paid for? For if it is not paid in any way, the quality control and journal system collapses. <http://www.ecs.soton.ac.uk/~harnad/nature2.html> It is conceivable to pay for the quality control via S/L/P in PARALLEL (i.e., there is a free version in the Public Archive, plus a fee-based version in paper, and in the Publisher's Online Proprietary Archive), and there is no reason anyone should object to that, except that it makes no sense: Why would anyone want to continue paying for what they can get for free? On the other hand, to attach an S/L/P price tag to the FREE version would make it no longer free, so that is out of the question. So page-charges are my candidate for how everyone could have their cake and eat it too (assuming that people will not long want to keep paying for what they can get for free): (1) Quality control continues to be paid for, hence it continues to be provided, and publishers continue to have their niche (albeit a much scaled-down one). (2) Libraries save their entire 100% serials budget, in exchange for redistributing <30% of it to their institutional authors' publication Funds. (3) Authors draw on the publication funds to pay their page charges up-front. Now do you see the logic of why it does not make sense to recover that 30% directly via S/L/P barriers? > sha> The ARL initiative is largely backing new forms of licensing. Inasmuch > sha> as these retain the author's right to self-archive for free, they are > sha> commendable; inasmuch as they help to preserve S/L/P barriers -- in the > sha> form of L alone -- they are counterproductive. shi> shi> ...later... shi> > sha> I have tried to sketch out a way > sha> <http://www.ecs.soton.ac.uk/~harnad/nature.html> based on switching > sha> to cost-recovery from up-front page-charges, but the promotion of > sha> universal self-archiving by authors does not require a commitment to a > sha> specific transition scenario. > shi> So S/L/P and page charges are all non-exclusive options? I hope that the above has clarified this: It is a logical fact that you cannot have something for free that you pay for by S/L/P. "L" stands for "License," and a License, whether Individual, Institutional, Consortial, or National is an access toll to be paid -- or not paid. Those Individuals/Institutions/Consortia/Nations that cannot pay it do not have access to the literature in question. Therefore it is not free. The logic of L is precisely the same as the logic of S and P. And the publishers know it; which is why Arnoud de Kemp at Springer regularly delights in informing audiences at which we both talk [at loggerheads!] that we are in fact BOTH arguing for the same outcome: A seamless, interconnected, global, online literature, free for all, paid for "up-front": All that's needed is a Global Site License! So just pay us the 100% now, and if there are indeed economies to be made that will scale it all down to 30% or lower, we will make those economies, and pass them on to all of you... But what this is envisioning is something that would I think make Adam Smith convulse in his grave: A Global Consortial Monopoly on the part of all publishers, where they name the 100% figure, and their sole, inelastic customer shells it out obligingly up-front, and entrusts the rest to the market-free beneficence of the Monopoly. Nonsense. Better to have the "market forces" distributed, but not via S/L/P, with its access barriers, which is how they are distributed now, but via author page charges, which has the virtue of calling a spade a spade. For once there is a free public archive for accessing the papers themselves, it is clear that the only indispensable product/service that online-only journal publishers are providing is quality control. The customer for that service is the author and, more specifically, his institution, for quality-controlled publications are the measures of research productivity on which their funding is in turn largely based. Referees actually do the quality control for free, but implementing it costs a little money, and that's what the page charges are to pay for. And just a small portion of the 100% annual savings from the termination of all S/L/P is the natural source for these funds. shi> Do we need to make another distinction between refereed reprint (author's shi> version) and the edited version of the same paper? One version can be made shi> available from the archive for free, the other for-pay. If there is no shi> distinction, then author page charges are the only way forward. But then shi> prescribing page charges as the only way goes against the 'one variable at shi> a time' principle. You said this a bit confusingly, but if you meant more or less what I said above -- free unrefereed preprints + free refereed final drafts + paid final drafts with publisher add-ons -- then the answer is yes, and the contingency is that, assuming that the paid version under these conditions will not find a market, the only way to pay for the quality control (the "Invisible Hand") behind the no-frills, vanilla version (the refereed, final draft) is up-front author page-charges. shi> Which brings us to the role of publishers and existing journals. Your broad shi> advice to the biologists is base the (LANL-like) archive on simple shi> principles and let market forces develop enhancements. The market decides whether it is worth paying for publishers' proprietary add-on enhancements. Peer-review itself needs to be financed in any case. But I also suggested that the Online community itself will develop powerful search/link enhancements for its Archive, for free. So I don't expect much of a market for proprietary enhancements by primary publishers (and that is even truer for the secondary publishers -- abstract, index, citation services -- who will have much more trouble finding something of value to add once the entire corpus is online for free [which includes its abstracts, keywords, reference lists, and full texts!]). shi> You also suggest shi> that it is desirable if many existing journals, esp. prestigious journals, shi> can adapt and attach to the archive, and to this end that the E-Biomed shi> proposal aims to bring publishers on board, if possible: > sha> Work out agreements with a sufficient proportion of established > sha> journals, as in the case of APS/LANL, and this will be a highly > sha> attractive feature, and will hasten the success of the ... Archive. > shi> and > > sha> Rather than announce it as a > sha> fait accompli, a priori, that established journals will allow their > sha> authors to self-archive online preprints and to submit to the journal > sha> via the Archive, I suggest you confirm this with a sufficient number of > sha> them so that you have a viable and attractive package to offer > sha> prospective archivers. > shi> This is fine until we get to perhaps your most contentious statement: > > sha> Journals just become quality control > sha> tags; otherwise, they are an outmoded concept. > shi> I hope it is not the case that journals just become tags. I would not argue shi> that journals should remain the same, but what incentive is there for shi> publishers to cooperate in allowing journals to be reduced to this? I think shi> there has to be scope for journals to add value, and they have to have some shi> way of exploiting that value. Are the physics 'overlay' journals just shi> 'tags'? Perhaps I should have said: "Implementers of quality control and its certification via a tag, the journal name." Now if you don't like that, then you don't like the online incarnation of the refereed journal corpus! For, absent paper, and absent also the need to provide the archival version online, that is indeed the only ESSENTIAL service that publishers still provide. (Add-ons are anyone's game, but I'm betting the community will do it better, and for free, than paper publishers, re-tooled as online-service-enhancers will be able to do. Remember one critically important thing: We are only talking about this highly anomalous subset of the entire world literature, namely, refereed learned journals. In this peculiar, non-trade domain, authors GIVE AWAY their work for free. This is not true of books, popular magazines, or just about any other product, online or otherwise, literature or otherwise. That non-trade, give-away reward structure -- where authors want to make maximal impact on the eyes and minds of their peers, not to extract dollars from their pockets in exchange for their products -- is what leads inexorably to the conclusion that quality-control certification is the only thing that needs to be provided or paid for in this very nonstandard "market.") shi> I think all this is possible - free archives with value-adding (for pay) shi> services overlaid - providing the one variable at a time rule is followed shi> and we don't try to prescribe too many features at once. Amen. -------------------------------------------------------------------- shi> Date: Thu, 29 Apr 1999 10:08:10 +0100 shi> From: Steve Hitchcock <[email protected]> > > sha> NO, definitely not. The American Physical Society (APS) journals > sha> continue to charge for the paper edition as well as for the APS On-Line > sha> edition (with many added enhancements) that APS sells. But the Los > sha> Alamos (LANL) public Archive version is free. This is a critical point. > sha> The "overlays" are a service to the authors, in that authors can submit > sha> papers directly to APS journals by depositing the preprint in LANL. > sha> They can also deposit their final accepted drafts (but not the > sha> proprietary APS version, with the enhancements, which only comes from > sha> the paid APS Archive) in LANL. > shi> This appears to be along the lines I was suggesting, i.e. free copy of shi> author's final version but for-pay enhanced proprietary version. But I'm shi> still unclear what 'overlays' are in the context of the archive, or how the shi> concept works in practice. If I go to xxx, how can I ask to see only papers shi> in the archive that have been accepted by APS journals, or by a specified shi> APS journal? Or do I have to go through APS' proprietary service to shi> discover that? Or, in the archive, do overlays exist not so much at the shi> journal level but at the level of the single paper, i.e. replacing the shi> original e-print with the refereed version? If you could point me to an shi> example that would be useful. Thanks -- Steve Hitchcock The answer is trivial: An Archive like LANL is a database and we are just discussing parameters on which searches can be set: You could do a search on, say, "Strange Attractors" on the whole archive, or you could restrict your search to the refereed journal sector, "R," or you can restrict to a particular journal, "JX," or even to a particular journal/year/month, etc. What you retrieve when you retrieve the refereed journal article is not necessarily the page images of the journal: Those may only be available from the publisher's proprietary, fee-based archive (if one still exists), with all kinds of add-on enhancements. The public archive will definitely contain only the "vanilla" version of the refereed final draft: the text and figures will be the final ones, but the formatting may not be the same as the printed page image (so what? paragraphs are the relevant entities on-line, not pages); the free version will, however, almost certainly be citation-linked, commentary-linked, update-linked, etc. So it will have plenty of useful add-ons of its own. So the "overlay" could take several forms: For a journal that has scaled down to providing quality-control certification only, fully funded from author page charges, the overlay will indeed be the official, authorized, final accepted draft of all papers. For a journal that still sells a paper version and/or maintains a proprietary online archive funded by S/L/P, with enhanced versions, the version in the public overlay will be the vanilla one, not the enhanced one; but it will still be identical in content, if not precise form, with the enhanced version. > sha> Perhaps I should have said: "Implementers of quality control and its > sha> certification via a tag, the journal name." > shi> This is curious. You appear to have moderated the contentious statement. In shi> fact I think there is a vision where a journal can become a powerful set of shi> tags, link tags, more powerful than is suggested by the added comment. But shi> there must be more, and my contention is with the word 'just'. > shi> Over the last four years in the Open Journal project we promoted the idea shi> of journals as 'open', mediated by links, so what are such journals but shi> collections of (link) tags: shi> 'The project also expressed a longer-term vision of link-based journals: shi> "For publishers the "journal" now becomes a set of links: substitute links shi> for the glue of the paper journal" (Hitchcock 1996).' shi> http://www.ariadne.ac.uk/issue5/open/ > shi> So I hope that e-journals will become largely collections of tags - the shi> knowledge process that is editing and publishing can effectively be boiled shi> down to link data - but the Open Journal project showed that there has to shi> be more, and that is where there is scope for value-adding. shi> http://www.dlib.org/dlib/december98/12hitchcock.html#Hitchcock96 Tags and links are technological matters. The main "tag" I'm concerned with is the price-tag, in particular, its absence! If the price of quality-controlling (refereeing/editing) is paid up-front, there is no need or justification for attaching any further price-tag to the quality-controlled final product, as quality-tagged by the journal brand-name. -------------------------------------------------------------------- Stevan Harnad [email protected] Professor of Cognitive Science [email protected] Department of Electronics and phone: +44 1703 592-582 Computer Science fax: +44 1703 592-865 University of Southampton http://www.ecs.soton.ac.uk/~harnad/ Highfield, Southampton http://www.princeton.edu/~harnad/ SO17 1BJ UNITED KINGDOM ftp://ftp.princeton.edu/pub/harnad/
