[This is an exchange with someone at University-X who has *not* agreed to
having his words posted, so I have abridged, paraphrased and completely
camouflaged his points and his institution.]

> PARAPHRASE: Open Access business models should be tried, and may
> one day prevail.

I agree, but I believe open access through self-archiving can and will
precede open access publishing and its accompanying change in business
model.

> PARAPHRASE: PLoS seems to have thought it through.

I'm not sure PLoS  thinking (which is in terms of governmental subsidies
and/or institutional licenses) will scale up to all or even most of
refereed research (20,000 journals, 2,000,000 articles annually). I
believe open access through self-archiving must come first, and only
then, and only *if and when* journal toll-revenues should ever shrink
(and corresponding institutional toll-savings grow) to the point where
peer review needs to be paid for in another way, only *then* will there
be a transition from toll-access to open-access publishing. But by then
the important part (universal open access) will already be behind us.

> PARAPHRASE: BMC may be underpricing to gain more sponsors.

I actually think BMC's $500 is closer to the (asymptotic) mark than PLoS's
$1500. Because PLoS is explicitly targetting the very highest quality
papers http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/2530.html
there's still a lot of excess fat on the PLoS bone. Hence here too,
I believe it will be self-archiving -- offloading all access-provision
and archiving onto the distributed institutional archives -- that will
eventually cut down to the only essential cost that still needs to be
covered in online-era peer-reviewed publishing (namely, the cost of
providing the peer review service itself).

> PARAPHRASE: Perhaps in a few decades...

Not if I can help it -- no more decades of needless delay, I mean)! I
am sure that open access is already universally reachable immediately,
and is in fact well overdue! It may be decades before the transition to
open access publishing, but I hope it will only be a few more years at
most before we have universal open access to the entire refereed corpus
(through institutional self-archiving).

> PARAPHRASE: Your view that self-archiving needs to come first sounds
> plausible, but $1500 seems closer to the pricing target than $500. And
> the revenue per article is much lower than $500 when access is
> unlimited.

The relevant figure is of course not the *revenue* per article (which
is the old, reader-end way of thinking, based on the toll-access model)
but the (model-neutral) *cost* per article. And to determine the size of
that, we need to specify the price *for-what* per article? Every product
and service being provided now (peer review and copy-editing, markup,
paper version and its distribution and marketing, online version and
its storage, distribution and marketing, online enhancements, etc.)?
or just an essential subset of it?.

I am betting that of all of these paper-era products and services,
the only *essential* online-era service will turn out to be peer review
itself (and possibly some editing) -- the rest being either jettisoned
or offloaded onto the distributed network of institutional archives,
self-archiving their own research output, both pre- and post-peer
review. And the price of peer review alone is far closer to $500 per
article.

But you have touched on some very important points. Let me try to put
them together into what I believe is a coherent picture of what is
actually going on today:

There is a straightforward *incoherence* in reckoning per-outgoing-article
peer-review service charges on a fixed annual institutional-rate basis,
on the same model as institutional access-tolls (licenses).

Fixed annual institutional rates are appropriate for access-tolls on annual
*incoming* articles, as they are now, but not for a stable open-access
model, which must be based strictly on each individual *outgoing* article
submitted to a particular journal for peer review. The reasons for this
are simple, and several:

    (a) Journals are independent, individual entities, selected by
    (and competing for) submitting authors and quality. Institutions
    cannot make a priori collective contracts with individual journals,
    committing themselves (on behalf of their authors?) to any annual
    quota of submissions (let alone acceptances), the way they can with
    journals that they subscribe to or license (reader-end).

    (b) The BMC-like institutional "membership" deals are hence an
    artifact of the fact that there exist virtually no open-access
    publishers apart from BMC right now, so BMC's cost-recovery model can
    be put forward in what looks like a universal way -- but it would
    immediately stop making sense if many other publishers (including
    competing biomedical publishers) were to approach universities with
    the same kinds of membership offer! The BMC solution does not scale
    up, either to the rest of the 20,000 peer-reviewed journals, or to
    the rest of each member-institution's researchers (if it were ever
    to become a real constraint on where researchers were allowed to
    submit their papers).

    (c) The BMC-like institutional membership only looks viable and
    attractive now, when it is taken (by default, being the only
    candidate) to be the *paradigm* for open-access itself (whereas
    in reality it is only one of many, many possibilities) and when
    it seems to be the only game in town; and (perhaps most important)
    when BMC has nothing to lose -- and everything to gain -- from using
    its memberships to try to capture a larger share of the biomedical
    authorship through this sort of indirect institutional "lock-in"
    of future institutional submissions. (Right now, the BMC share of
    biomedical research as a whole is still minuscule.)

In short, the BMC open-access publication model has not been thought
through by the research and library community *at all*, whereas BMC itself
has only thought it through (understandably) from its own bottom-line
standpoint (and improvising as they went along, helped along by the
rising tide of pro-open-access sentiment in the research community).

For what is undeniable today is that there is a growing demand by
the research community for open access. Yes, immediate open-access
publishing would be *one* possible way to fill that demand, but
there is no way to get universal open-access publishing immediately
(toll-access publishers are understandably disinclined to convert,
especially since the research community's demand has so far only been
expressed in the form of petitions and polemics -- plus the small number
of new open-access journals that have so far been created; and authors,
too, are wary of new journals, open-access or otherwise).

There is, however, a way to have immediate, universal open access,
without the need to first wait for publisher and author conversion from
toll-access publishing to open-access publishing, and that is through
self-archiving. Instead of petitioning and polemicising and making
altogether too much of the regrettably few open-access journals that
exist so far for researchers to submit to or publish in, researchers
can make their insistence on immediate open access -- for their *own*
work at least -- immediately felt and realized by self-archiving it.
http://www.ecs.soton.ac.uk/~harnad/Temp/unto-others.html

There is absolutely nothing (including copyright:
http://www.eprints.org/self-faq/#copyright1 ) to stop all researchers
from doing this immediately -- except their own confusion and uncertainty
about it. That confusion and uncertainty is what I, for one, am dedicated
to working to dispel. And the outcome will be not only open access
itself, but a potential basis for converting to open-access publishing
too -- but *only if and when it is ever needed* -- on a rational
institutional basis: paying for the institutional peer-review costs,
on a per-outgoing-paper basis, out of the annual institutional windfall
toll-access savings (if/when they ever get big enough to force the
transition, i.e., *after* the fact), rather than on the basis of fixed
annual licenses or other notional sources (i.e., *before* the fact).
http://www.nature.com/nature/debates/e-access/Articles/harnad.html#B1

> PARAPHRASE: BMC charges my University -- "University X" -- about
> $500 x 10 = $5000 [actual figures altered so as not to identify any
> institution, but ballpark is the same] for its yearly membership. (Our
> faculty were, predictably, not in favor of the deal.) So far, about 20
> University-X researchers are already publishing in BMC journals annually
> [actual figures altered, but ballpark is the same].

The faculty disinclination toward the BMC deal is quite understandable
(though in itself it is certainly no evidence that it's not a good
idea)! From your own figures, this amounts to a publication subsidy to
about 20 University-X (biomedical) researchers per year right now, while
everything else stays the same: All of University-X's other incoming
journal-tolls still have to be paid. Universal open-access by this route
is nowhere in sight. University-X did this in the hope it would be an
investment in more such open-access journals, but it will only work
as long as BMC alone grows at the expense of the existing toll-access
journals. If other open-access journals start to vie for "membership,"
University-X will have to start paying attention to how many articles are
actually being submitted to each journal (maybe you should already start
doing a per-journal reckoning now, as a sample!) to make sure that the
membership-fee for each balances with the publication rate. And all this
at a time when these membership-fees must all be paid *in addition to*
toll-access costs (with no sense of when and whether those toll-costs
will diminish).

Wouldn't it make more sense for University-X -- *in addition,* not
instead -- to also throw its full weight behind a concerted policy
of institutional self-archiving for all of its own refereed
research output too (rather than worrying only about how to
help open-access publishers find ways to make ends meet?):
http://www.ecs.soton.ac.uk/~harnad/Temp/archpolnew.html
http://www.ariadne.ac.uk/issue35/harnad/

Those University-X researchers who are already clamoring for open access
would then have it; University-X research visibility, uptake and impact
would rise for *all* University-X researchers (rather than for just 20 of
them) and at a one-off cost, for now, of closer to $5 per article (for
self-archiving alone) rather than $500 (which would be needed only
if/when the toll-access cardhouse ever falls). (And the $500-per-article
charge would only come if and when the competition from open-access usage
worldwide ever generates the institutional windfall toll-access savings,
worldwide, out of which to pay for it -- on a per-article basis.)

> PARAPHRASE: With 20 articles instead of 10, that already makes
> it $250 per article instead of $500. Any more and BMC will have to
> raise its rates, causing financial hardships for member universities.

Don't worry for BMC! If they manage, they manage. If they later raise rates
and institutions balk, cross that bridge when you come to it. Worry now
about the *rest* of University-X's research output, over and above the 20
articles in question!

(And remember, on the self-archiving model, peer-reviewed journal
publication downsizes to peer-review service provision [and certification]
alone, whereas both BMC's expenses [and your own reckoning of what it
must all cost] are based on co-bundling with a lot of other products,
services, and enhancements that may well prove unnecessary by the time the
need for a transition makes itself felt [if/when it ever does]; by then
the journal publishers will have done a lot of unbundling, cost-cutting
and downsizing to essentials. BMC is not yet ready to see an open-access
journal publisher as merely a peer-review service provider/certifier! But
usage pressure from the vanilla self-archived open-access versions may
well force a re-evaluation of what is a necessity and what is not, once
access-provision is offloaded onto the institutional archives.)

And if BMC should ever go belly-up (as many other journals, on-paper
and on-line, have done before), those University-X (or other) authors
who have published therein will nevertheless continue to have the
peer-reviewed articles they published therein accessible to and usable
by the world in perpetuum -- by simply self-archiving them, along with
all their toll-access articles, in the University-X Eprint Archive.

> PARAPHRASE: I hear that each article in the 95 BMC journals averages
> one download per month.

I think that's a considerable underestimate. I'm sure that BMC open-access
articles do not get, on average, more or less downloads and citations than
other comparable-quality open-access articles (whether self-archived or
published in open-access journals) -- which is, on average, a lot more
downloads and citations than comparable-quality toll-access
articles get (4.5 times as many, according to Laurence 2001
http://www.nature.com/nature/debates/e-access/Articles/lawrence.html )
In other words, the impact-enhancing benefits of open access are not
in dispute (whereas the instrinsic quality-level of BMC articles is a
separate matter, on which I have no views, or information).

The relevant question is this: Does University-X want, right now, for
*all* of its researchers, the enhanced impact that open access is currently
providing for only its 20 BMC authors -- or does it prefer to wait
decades for it?

http://www.ecs.soton.ac.uk/~harnad/Temp/self-archiving.ppt

Stevan Harnad

NOTE: A complete archive of the ongoing discussion of providing open
access to the peer-reviewed research literature online is available at
the American Scientist September Forum (98 & 99 & 00 & 01 & 02 & 03):

    http://amsci-forum.amsci.org/archives/september98-forum.html
                            or
    http://www.cogsci.soton.ac.uk/~harnad/Hypermail/Amsci/index.html

Discussion can be posted to: [log in to unmask]