[First, apologies for cross-posting. Second, a note of explanation
    about this posting, which has just appeared in the American Scientist
    Open Access Forum
        http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/5957.html
    and is not by me, Stevan Harnad, but by my brother, John Harnad,
    a physicist. Although I will follow the posting with a critique
    by me on some points of detail, I want to stress that I am in basic
    agreement with the gist of John Harnad's point that a hasty CERN-led
    forced conversion to OA (Gold) Publishing in (part or all of) physics
    at this time is likely to cost more, is likely to divert funds from
    research, and should be deferred in favor of expanding the practice
    of OA (Green) Self-Archiving and parallel co-existence. -- SH]

-------------------------------------------------------------------------
    Is OA (Gold) really a desirable goal for scientific journal publishing?

            John Harnad, Director
            Mathematical Physics Laboratory
            Centre de recherches mathématiques
            Université de Montréal

Seven Reasons why Open Access (Gold), as a policy objective, should be
more carefully scrutinized within the scientific research community

An article appeared in the January edition of 'Physics World', under the
title 'The Open Access debate', combining the differing views of two
physicists who have played an active role on the issue of publishing
policy in physics: Rudiger Voss of CERN, who was author of the lab's
recent report 'Open-access publishing in particle physics', and John
Enderby, until recently the President of the Institute of Physics,
vice-president of the Royal Society of London, and lead officer for its
publishing activities. Their views differ considerably, with Voss taking
an active position in favor of the 'author pays' Open Access policy model,
and Enderby a more cautious one about its possible implications.
http://physicsweb.org/articles/world/20/1/4/1?rss=1.0

Since CERN, in part through the advocacy of Rudiger Voss, has apparently
declared itself in favour of the Open Access (Gold) Publishing model,
this will likely have a strong influence on subsequent developments,
at least in high energy physics publishing. The arguments opposed
to this position need to be given adequate consideration before any
decisions are implemented that may possibly impact more broadly on the
research community. Some of these arguments are discussed in detail in
the following.

First, the definitions:

Open Access (Gold) means: a journal charging nothing to readers for access
to the electronic versions of articles published in it. This is to be
distinguished from Open Access (Green) which means a journal allowing,
or encouraging the simultaneous deposit of peer reviewed published papers
in publicly accessible, linked institutional repositories, or central
repositories like ArXiv.

It is a pity that the same expression 'Open Access' is used by many to
refer to both these policies - without making clear the fact that two
logically and practically distinct concepts are being confounded. This
may cause misunderstandings which, in some instances, can even camouflage
a hijacking of objectives. The 'Gold' version of Open Access involves
several questionable implications for the scientific community, and those
advocating it within the community should seriously rethink its logic.

The main point to recognize is that the only mechanisms by which a
journal can operate in 'Gold' Open Access mode are: 1) direct support
through public or other institutional grants, 2) advertising revenues,
3) subscription costs for paper versions that exist in parallel with the
free, electronic versions, or 4) transferring the costs for a major share
of its overhead, and profits, to the authors. (Of course, in most cases,
this does not mean the personal bank account of the authors, but their
research grants, if they have them, and can afford it.) Charging only
for the paper version of subscriptions, when the electronic version is
accessible to all for free is, in most cases, not likely to be a viable
way to cover costs or make a profit. Therefore, the only candidates for
this are, either: the authors, or paying advertisers, or direct grants
from public or private funding agencies.

Of the roughly 2500 journals currently listed by the Lund University
Directory of Open Access Journals, a large number function by charging
their authors very hefty publications fees (e.g., those published by
the OA (Gold) publishers Biomed Central charge their authors 'article
processing charges' that are typically of the order of $1500 US
(1120 euros) per article). Some (e.g., in biomedical research) have
advertising revenues that are adequate to sustain them, possibly when
combined with professional association fees and subscriptions. But this
is not a feasible model in a majority of areas of scientific research.
In some cases, direct government support is adequate to sustain OA (Gold)
journals, but, for the most part, these are journals having limited
geographical scope, and little or no international standing. There are
also those that are effectively 'in-house' publications supported by a
specific university department, largely through volunteer work, without
a paid full-time professional staff. These generally are also of rather
limited scope and distribution, serving a somewhat narrow segment of
the research community.

The following outlines seven reasons why OA (Gold) is currently a bad
general policy objective for the scientific community.

1) In most areas of research, no alternative to 'author pays' or
'subscriber pays' models currently exists that is compatible with
maintenance of quality. There do not exist sufficient direct grants from
government or other research funding agencies to publishers, nor revenues
from private advertisers, to allow a majority of journals to become
either publicly funded or self-supporting through advertising revenues.

2) There is a large variation across domains of research in the percentage
that would have to be attributed from available research grants to
cover publications costs if they were to be transferred mainly to the
'author pays' mode. In some domains, where the scale of research grants
is very high (e.g. experimental high energy physics and some domains
of biomedical research), this may be only of the order of 1-2%. But
in others (e.g. theoretical and mathematical physics), where research
grants are available only on a more modest scale, this could easily rise
to 10-15% if applied to all journal publications. Thus, these areas would
be relatively penalized by an order of magnitude regarding monies that
must be subtracted from other, 'direct' research purposes.

3) Those researchers who do not have substantial research grants -
which includes those from countries that cannot afford high levels
of research support, and individuals from countries in which a highly
selective process of grant attribution excludes a large percentage of
potentially active researchers from the benefits of grant support -
would be particularly penalized by such a mode of charging.

4) It is highly unlikely that public (or private) funding agencies will be
willing to increase their budgets to cover such extra publication charges
for authors, even if they express themselves in favour of 'Open Access'
and continue to allow this (as most do now) as a legitimate item within
the budget of a supported researcher. The implication is that the extra
costs for publication charges will have to be subtracted from other,
current research expenditures. For those, e.g., in the 10-15% category,
this means, effectively, a 10-15% cut in their 'actual' research budgets.

5) The notion that 'Open access' will miraculously cut the costs to
publishers, making it possible either to charge lower subscription
rates for paper versions, or more modest page charges than have been
applied in the past, is a fallacy. The erroneous logic behind this
is based on the expectation that, since electronic versions are much
cheaper to produce, reducing the volume of paper printed versions (or
eliminating these entirely) will greatly diminish the overhead of the
publishers, making OA 'Gold' much more cost-effective. This is simply
confused thinking. Although there is certainly a diminution of costs to
be expected due to the increasing emphasis on electronic publishing,
this will be the case because of ongoing developments in technology,
and habits, not due to 'Open Access' or any other mode of cost-revenue
balancing. If the journal is reliant entirely on its electronic version,
which is free, it has to generate revenue somehow. The loss of income
from subscriptions for OA 'Gold' publishers will be the overwhelming
factor, pressuring them to transfer costs to the authors. However, if
this becomes the main source of revenue for publishers, the rate for
page charges can only become even higher than in the earlier days of
mainly paper printed versions of journals.

6) The notion that money that would be saved by libraries will be made
available to the researchers who will henceforth have to cover the costs
for producing journals from their research budgets is also, in most
academic settings, erroneous. There is no mechanism for such a transfer.
In most academic settings, the sources and methods of distribution of
funds for these two purposes are completely distinct, and it is nothing
but wishful thinking to imagine that there will be an automatic
adjustment that balances a major transfer of the financial burden from
one to the other. The high costs, if they remain high, will simply be
transferred from library budgets to researchers budgets, without any
adequate compensatory mechanism to offset the change.

7) The scientific quality of journals would be negatively affected by
transferring the burden of costs from subscribers to researchers. This
mechanism is not likely to ever be applied universally to all journals in
a given field, and those journals which do not rely on hefty page charges
for their operation will, as in the past, tend to be the more prestigious
ones, where an author must provide an article that is of sufficiently high
calibre to justify its publication, whereas the page-charge journals,
being reliant on this income for their sustenance, will tend to accept
lower calibre contributions, provided the author is willing to pay.

For the immediate future, taking into account the variations in sources
and levels of support available for funding scientific publishing across
different domains, a 'hybrid' model with adequate choices and flexibility
would best serve the community. In some areas, either because of the
availability of advertising revenues, or very high levels of research
grant support, perhaps an OA 'Gold' policy can be sustained. But in a
large part of the scientific research community, such a model would entail
an unjustified and unwise transfer of the burden of support for scientific
publication costs to the researchers and their existing resources.

Given the currently available resources, a large-scale switch to 'Gold'
Open Access is neither beneficial to the quality of scientific publishing,
nor in the interests of most researchers. This does not imply that
the 'ideal' of Open Access (i.e. cost free access to the scientific
community) is not desirable or achievable. A large part of it can,
however, be achieved without transferring the cost burden to authors'
research grants, simply by relying mainly on freely accessible
institutional repositories, or central repositories such as ArXiv,
to provide universal access. Naturally, such repositories bases do not
provide the 'value-added' or guarantee of quality that the peer-review
refereeing system does, and hence cannot substitute for it, nor do
they, under present conditions, provide adequate guarantees of long
term preservation for posterity. But the parallel existence of the two
does both, provided the peer reviewed, referee-based publishing journals
continue to accept the co-existence of such no-cost access to essentially
the same body of published papers, divested perhaps only of the luxury
of standardized formatting. It is up to the publishing authors - and in
their interests - to see to it that they do so.

John Harnad, Director
Mathematical Physics Laboratory
Centre de recherches mathématiques
Université de Montréal