Re: Nature's vs. Science's Embargo Policy

From: Stevan Harnad <>
Date: Sat, 11 Jan 2003 23:16:10 +0000

Replies to:
Jan Velterop (3), Fytton Rowland, Alan Story, Michael Eisen & Mark Doyle

> Date: Fri, 10 Jan 2003 10:05:41 -0000
> From: Jan Velterop <jan_at_BIOMEDCENTRAL.COM>
> I agree with Mike. Nature's new 'licence' is a
> 'pull-the-wool-over-your-eyes' version of what Elsevier calls the
> 'give-backs' and is nothing new at all, just a new PR exercise.
> Clever PR...

I'm not sure why you're saying that, Jan, but I hope you are prepared to eat
your words (as I am, if I'm wrong)!

> BioMed Central have a copyright and licensing policy that can truly be
> seen as leading: No
> restrictions on self-archiving and further dissemination whatsoever. That
> should be the 'gold-standard', not Nature's feeble attempt to look good
> without delivering any substance worth mentioning. Congratulating Nature
> for putting a new gloss on basically an old stance seems unnecessary
> sycophancy to me.

This sounds a bit shrill! (I'm not sure why...?) My solely interest is
in hastening universal open access (what I have called, never often
enough, the optimal and inevitable outcome for all refereed research
and researchers). Open access can be had instantly if researchers will
simply self-archive their refereed research in their institutional
Eprint Archives. Many don't do it because they believe it is somehow
illegal. Nature now explicitly reassures them that it's ok -- yet you are
unhappy: Why?

> Describing the new Nature licence as a 'gift horse' that shouldn't be looked
> in the mouth (in one of Stevan's earlier messages on this topic) is giving
> the wrong impression that the scholarly community should really sit back, be
> patient, shut up, swim on, wait what's being given to them and then be
> grateful for beads and mirrors.

Jan, this is beginning to border on the absurd! You know very well that
I've not been preaching patience and shutting up, but immediate open
access through self-archiving, right now! The Nature policy removes one
of the perceived obstacles to doing that. Why should that not be welcomed?

> They should simply expect more from publishing, and demand the right to
> self-determination of what can be done with their articles. Besides,
> there would be no point in looking a dead horse in the mouth anyway,
> apart from performing an autopsy.

What more self-determination should an author who is seeking open access
for his work expect than the right to provide open access to it? The
right to *pay* to have the self-archiving done for him? (And Nature,
and the other 20,000 toll-access journals don't look at all like dead
horses to me. What I would like to breathe more life into is researchers'
self-archiving efforts!)

> Of course, authors could always re-format their papers and flip into
> 'subversive mode' again. They could always do that anyway, and Nature's new
> formulation of their restrictions doesn't make that any different.

I have many times said that I now regret ever having called self-archiving
I (like many others) was earlier under the (completely wrong) impression
that the obstacle to open access to refereed research was publishers. I
now realize that it is not, and never was. The only obstacle is
researcher unawareness and inertia.

Dramatic empirical demonstrations of the causal connection between access
and impact will make researchers (and their institutions) aware, and
giving them and their institutions every means to go ahead and do this
effortlessly will combat the inertia. Removing the perceived obstacles
(like copyright worries) can only help too. Hence my gratitude (sic)
to Nature for their explicit support of self-archiving.

> There is a lot to be cheerful and optimistic about with regard to open
> access, but Nature's copyright licence ain't amongst it. The question
> remains, if Nature really permits self-archiving (which is what Stevan seems
> to believe), why don't they make their research papers available in open
> access or at least freely available after a short time (say a month or two)?

Supporting self-archiving does not mean *doing* it for the author! That
would be for Nature to become an open-access publisher like BMC (at a
time when it is far from clear that ends can be made to meet yet that
way): Is *that* what you wanted Nature to do?

> There's nothing to be lost for them that cannot be compensated by the gains
> they could make from such a policy, in my view. Open access advocates should
> keep up the pressure instead of relenting when offered a cigar from their
> own box.

Pressure for on whom to do what, how? It is not journals that need to
be pressured but researchers. Unlike journals, they can provide open
access without any risk of revenue loss, only the prospect of
lost-impact gained!

> Date: Fri, 10 Jan 2003 15:32:39 -0000
> From: Fytton Rowland <J.F.Rowland_at_LBORO.AC.UK>
> Every information system has to be paid for somehow, since it has
> unavoidable costs. The only way that you can (appear to) charge nobody is
> to have some kind of operating subsidy from somewhere. In discussion and
> analysis, it is helpful, in my view, to make a clear distinction between
> systems that charge the creator of the information and systems that charge
> the reader (or their respective institutions, of course).

Agreed. And it's useful also to remind ourselves just what it is that
each is charging *for*: In the one case, they are charging for *access to
a product* (along with all the attendant expenses of producing that
product). In the other case they are charging only for a *service*
(peer-review), with the product provided by the author, archived by the
author's institution, and openly accessible to everyone. That is why
the cost is so much higher in the first case than the second.

> Date: Fri, 10 Jan 2003 16:17:38 -0000
> From: Alan Story <>
> What I object to is one system that charges for the ability to "speak"
> dissing a system that charges for the right to "listen" (or perhaps more
> accurately, the right to produce and read accessible information) on the
> grounds that the former is in favour of open access and that the latter is
> not. Both are against it, albeit in different ways.

I think this is just plain incorrect (see above): One charges
user-institutions for access to a product (the peer-reviewed paper); the
other charges provider-institutions for a (much cheaper but essential)
service (peer review itself), with the product (the paper) being openly
accessible to all. (How is that "against open access"? We're trying to
free the peer-reviewed literature from access tolls, not from peer

> Many researchers, academic and otherwise, in many parts of the world have
> lots worth saying, but simply cannot "stump up" $US 500 each time that
> they --- and peer reviewers --- agree they have something
> worthwhile/important to say as the BioMed model requires. Sitting here in
> our well-funded universities in the US and Europe -- or well-founded
> compared to many parts of the world -- we tend to forget that many such
> researchers work at institutions where "of course" they don't have the funds
> to pay their staff for access to BioMed as a contributor...and not merely as
> a reader.

First of all, $500 per paper is the high-end estimate of what
peer-review costs will be. For many journals they may be considerably

Second, the lion's share of the current cost for the toll-access product
(for those who can afford it at all) is paid by the research universities
who also supply the lion's share of the contents of those journals.

If and when demand ever shrinks (because of open access provided by
self-archiving) to where toll-revenues can no longer pay the costs
of peer review, those research universities will by then by the same token
have the corresponding windfall toll-savings out of which to pay them
(and plenty left over).

My own guess would be that university serials budgets the world over are
roughly proportional to their own research output; that would mean that
accounts will in most cases balance. Does it take a lot of imagination
to note that to take care of that minority of cases where they do not
balance (very poor institutions, unaffiliated researchers) all that is
needed is a slight overall markup of the peer-review costs, to cover a
"slush-fund" for researchers whose institutions cannot afford the peer
review (or who have no institution)? Surely this makes more sense than
assuming that the minority of disenfranchised exceptions means that the
peer-review service fee model won't work!

> Date: Fri, 10 Jan 2003 17:29:45 -0000
> From: Jan Velterop <jan_at_BIOMEDCENTRAL.COM>
> We publish and believe in open access to research articles, but not in a
> free lunch. The $500 article processing fee has to be seen in the context of
> an amount as much as ten times that, which is currently being forked out per
> article by the scientific community.

True. But at BMC this is not yet merely a peer-review service charge,
with archiving offloaded onto the authors' institutions. When that is
the case, the cost may prove to be lower... (But as long as a product
is provided, instead of just the peer-review service, costs will not
be minimized to the essentials.)

> Date: Fri, 10 Jan 2003 10:41:02 -0800
> From: Michael Eisen <mbeisen_at_LBL.GOV>
> Why do you think the real cost is closer to $500?
> We spent a considerable amount of time and research trying to project
> exactly what it was going to cost to run our journals, and, for the near
> term, the number is very close to $1,500 per published article. I should
> note that this is very much in line (and generally lower) than the numbers
> one gets directly from established publishers or by extrapolating from their
> financial statements.

See above, about providing a product rather than just the peer-review
service; and about not offloading archiving onto the author's

> PLoS is committed to having our charges directly reflect our actual costs.

Even if the 20,000 toll-access journals of the world eschewed all profits
and merely charged tolls at-cost, that would not provide open access. And
redirecting the entirety of those costs to the author-institution as a
publication-charge is not viable either, because the figures are simply
too high! Cost-cutting and downsizing to the essentials is necessary in
order to make author/institution-end payment feasible and acceptable to
authors and their institutions.

And the payment will be for a service to the author-institution, not for
a product. The product is already provided by the author-institution. The
service is peer review. The archiving will be provided by the
author-institution too, but that entails not a further payment to the
publisher, but rather a further cost-cut by the publisher.

> It is, of course, impossible to project what these will really be until we
> actually start publishing. We are confident that these numbers will go down
> over time, as technology reduces the amount of time spent in managing peer
> review and processing the manuscript. It will, however, go to zero so long
> as professional editors are involved in the process. In the long term, the
> costs (and our charges) will scale with the amount of professional editorial
> input that goes into each journal and article.

Fine. But the higher the publication-charge levied initially, the lower
the probability of success, I would think.

> For some publications, costs of close to $1,000 will likely remain. For
> others publications, where no professional editors are used and where all of
> the labor comes from community volunteers, the costs will be near $0.

Then wouldn't it be better to use the subsidy to keep the charges much
lower initially, until open-access journals catch on (and institutional
toll-access windfall savings begin to build up) rather than risk
deterring authors and their institutions a priori with
unprecedentedly high charges?

> I am very confused by the reluctance of many on this list to embrace the
> idea that academic institutions and funding agencies should view the costs
> of publication as fundamental costs of doing research and pay them upfront,
> enabling the product to be made freely available from the moment of
> publication to the entire world.

Nobody could agree with this more than I do. But surely it makes a
difference whether the funds have as yet been freed to pay the (lower)
costs of open access? At the moment, institutions are double-paying,
when they support their current toll-access burden *and* pay for hitherto
unprecedented publication-charges. It's only natural that they should
balk *now*, even if they agree with the principle.

> It is pointless to try to frame this as a
> broad philosophical/moral argument about whether producers or readers should
> pay the costs. The fact is that the upfront payment model is practically
> superior to the current subscription model in virtually every way. It is
> more economically efficient, and thereby will be considerably cheaper on the
> whole. It is far more equitable, as there are no restrictions on who can
> access or use the material. And it will make the published material far more
> valuable by enabling the development and deployment of better tools to
> access and use published material.

I agree. But we have to find a way to get *there* (lower, up-front
payment for a service to the research-provider) from *here* (higher,
toll-based payment for access to a product). While so much money is
currently tied up in tolls *here* it will be hard to get institutions to
pay even more now, in order to get us *there*.

Let us hope that parallel progress in self-archiving (providing instant
open-access for free) will eventually loosen up the funds needed for the
transition (if and when it ever becomes necessary: and if it never becomes
necessary, we will still have open access!)

> So long as all open access, pay up front
> publishers ensure that authors without access to funds are still able to
> publish their works (which, I note, all such publishers including BMC and
> PLoS do), then I simply can not understand the objections being raised here.

Because even those institutions who can afford to pay still more than
they do now would rather not do so: What they want to do is to pay less,
not more.

> Date: Fri, 10 Jan 2003 14:01:42 -0500
> From: Mark Doyle <doyle_at_APS.ORG>
> >dg> For one thing, probably $500 is nearer the real cost than $1500;
> Well, you have to cover rejections as well as accepted articles. $1500
> is about what it costs APS per published article ( this covers
> ALL costs associated with our journal publications, including rejections
> which cost more on average to peer review than accepted articles and
> it assumes all income would come from submission side charges for
> accepted articles only).

Doesn't that bundle in a lot more than just the service costs for the
rejections? The rejection costs can, like a shop-lifting fee, be added
to the peer review costs for the accepted articles. So far that's fine;
but isn't about 2/3 of that $1500 for the *product* (paper version and
distribution, mark-up, archiving, etc.) rather than just the peer-review
for the accepted and rejected papers?

> Anyway, for a large US research institution, for APS publications, I
> find the money is about the same in both the new and old systems,
> but that it is no longer split between the researcher and the library.
> APS revenue per published article is equal to our cost per published
> article since we are non-profit. Applying this model to a for-profit
> publisher would (it seems to me) lead to a net savings for the institution.

APS journals are no doubt less expensive than commercial ones, but that
is not relevant. As I said, even if all tolls went down to cost only, as
long as the cost of open access still included all the costs of producing
the (former) toll-access product, it would still be needlessly (and
probably prohibitively) high. The cost needs to be pared down to
peer-review service-provision alone, or something close to it.

> Date: Fri, 10 Jan 2003 21:42:04 -0000
> From: Jan Velterop <jan_at_BIOMEDCENTRAL.COM>
> The actual price per article is very much dependent on the scale and
> efficiency of the operation, given that so much of the cost is fixed
> (the system) or quasi-fixed (staff). The $500 per article of BMC is
> based on assumptions of scale which we believe are achievable within
> a reasonable amount of time. These assumptions are ambitious, but not
> overly so. There is a lower limit to those economies of scale, though,
> because of basic technical and production costs and the like, and that
> lower limit now seems to be under $500, but only just. This may change
> over time (in either direction, I might add), but is unlikely to result
> in costs that are massively lower or higher.

Basically, this is still based on reckoning the costs for a product
rather than for just the peer-review/certification service, with the
rest offloaded onto the author's institution. As such, it is still
needlessly high.

> For now, the fee per article is only charged per published article (at
> least at BMC). That still is a (pragmatic) inconsistency in the input-paid
> model that eventually needs to be addressed, either by charging a higher
> fee proportional in some way with the rejection rate of a journal, or
> a non-refundable portion of the per article charge even if the article
> is rejected.

Once it's down to just the peer-review costs alone, the rejection costs
can be factored in like a shop-lifting fee (or possibly -- *this is
only a speculation!* -- a much lower submission fee could be charged
for the processing of all submissions, creditable toward the higher
acceptance fee for those that are accepted, but not refundable if the
paper is found to be unrevisable and rejected; this might serve as a
deterrent to the nuisance submissions that waste so much of referees'
scarce and freely-given time, for papers that prove to have been aimed
unrealistically high in the journal quality hierarchy, or simply hopeless:
perhaps journals could have reciprocity agreements, waiving the submission
fee if the author provides the prior journal's referee reports and
indicates what has been done to improve the paper in response to them).

Stevan Harnad
Received on Sat Jan 11 2003 - 23:16:10 GMT

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:46:49 GMT