Re: May 12 CERN meeting on implementing the Berlin Declaration

From: Stevan Harnad <harnad_at_ecs.soton.ac.uk>
Date: Mon, 24 May 2004 22:26:32 +0100

          CRITIQUE OF SCHLOEGL/VELDEN ROADMAP
          FOR IMPLEMENTING THE BERLIN DECLARATION

               Stevan Harnad

This is a critique (I hope a constructive one) of Robert Schloegl's and
Theresa Velden's (S/V's) Proposed Roadmap for implementing the Berlin
Declaration.

http://www.zim.mpg.de/openaccess-cern/berlindeclaration.html
http://www.zim.mpg.de/openaccess-cern/program_prelim.html#roadmap
http://www.zim.mpg.de/openaccess-cern/presentation-oa2berlin-roadmap-proposal.pdf

The problems with the S/V Roadmap are, I think, substantial, but readily
remediable:

(1) The (library's) problem of journal-affordability is, as so often
happens, treated as if it were the same thing as the (researcher's)
problem of accessibility and impact. It is not, and OA is meant to
be the direct remedy to the researcher's access/impact problem and
only indirectly and eventually a possible solution to the library's
journal-affordability problem. Trying to treat OA as a direct solution
to the affordability problem only produces a confusing mixed-agenda with
little likelihood of any solution and hence no clear rationale for
implementation.

(2) Open Access itself is, for much the same reason, treated
as if were the same thing as Open Access Journal Publishing. It is not.
Open Access Journal Publishing is merely one of the two roads to Open
Access -- and not the faster, broader, or surer of the two.

(3) As a consequence of (1) and (2), the concrete implementation proposals
themselves are both vague and unfocussed.

The solution is simple:

    (A) A clear and explicit definition of Open Access (OA) (already
    provided by the BOAI as: "immediate, permanent, toll-free, full-text,
    online access to peer-reviewed journal articles")

    plus

    (B) a clear and explicit commitment to providing OA, by the two means
    available: (B1) publishing articles in OA Journals (wherever possible)
    and (B2) otherwise publishing articles in existing journals but also
    self-archiving them in OA Archives.

That is what needs to be implemented, no more, no less. Instead, the
Schloegl/Velden (S/V) Roadmap is missing one of the two roads to Open
Access, B2 (the far wider, faster and surer of the two) altogether,
and it fails to give clear or concrete directions for using either of
the roads, B1 or B2!

> Roadmap Proposal Robert Schloegl Theresa Velden

> Open access is the replacement for the conventional scholarly
> communication paradigm and not its 2nd class counterpart

There is no paradigm and no replacement! OA is something that an author
can *provide* for his peer-reviewed journal articles. There are many
"classes" of peer-reviewed journals: higher quality ones, with high
standards of peer review, lower quality ones, with lower peer-review
standards, etc. None of this has anything whatsoever to do with whether
or not the articles are OA, or whether or not the journals are OA.

The reason S/V describe OA as a "paradigm change" is because they are,
as noted, equating OA with OA publication (the "golden road" to OA). This
is both incorrect and counterproductive. Publishing in an OA journal is
only one of the two ways to provide OA, and it is profoundly limited by
the fact that there exist only about 1000 OA journals (5%) today, whereas
there exist 23,000 non-OA journals (95%)! The road to implementing OA
that is completely missing from this "Roadmap" is the one for that 95%,
the "green road" of self-archiving.

Yet instead of being treated as the *access* road that it really is,
"archiving" is only discussed in the road map in connection with
"preservation." It is odd to be worrying about the preservation of
as-yet-nonexistent OA traffic! Odder still for the green road, which is
merely a supplementary means of access, for those who cannot afford the
toll-roads, and not a substitute for it. (Why focus on the preservation
of a near-non-existent supplement, rather than on first *providing* it?
Particularly as the primary, non-OA version is the one that requires the
preservation, not the supplement.)

A paradigm change may (or may not) eventually occur, in the peer-reviewed
journal-publishing business in the online era, but the goal of the OA
movement is not to provide an eventual paradigm change for the publishing
community, but to provide immediate OA for the research community.

http://www.nature.com/nature/debates/e-access/Articles/harnad.html#B1
http://www.publications.parliament.uk/pa/cm200304/cmselect/cmsctech/399/399we152.htm

> Roadmap Pre-conditions for Open Access

> ascertained quality assessment due to immediate access to
> primary source information interconnected with interpretation
> and secondary information

I could not follow all this terminology, but assuming that we are talking
about the yearly 2.5 million articles that appear in the existing
24,000 journals, I assume peer review continues to be the quality
assessment, OA provides universal access, and the users do the rest
(read, use, interpret, discuss, cite) in their own subsequent research
and articles. This usage can in turn be "assessed" through measures of the
download and citation impact of the articles in question. It is meanwhile
being demonstrated that the only way for usage and impact to go, given OA,
is up: OA articles are used and cited far more than non-OA articles --
and that's what OA is all about, is it not?

http://www.nature.com/nature/focus/accessdebate/21.html

So this first rather complicated clause of S/V's roadmap does not seem
to say anything concrete or relevant at all.

> superior capabilities of digital medium ( e. g. enhanced
> multimedia representation) for scholarly communication

With virtually all of the major journals now being hybrid -- i.e.,
publishing both a paper and an online edition -- it seems rather late
in the day to be preaching the virtues of the digital medium: They are
already known, and implemented. What needs to be implemented is OA:
i.e., immediate, permanent, toll-free, full-text online *access* to
those digital articles!

> Increased productivity of discovery process due to unlimited
> and unrestricted access to all relevant information

Yes, OA means maximized access, which means maximized impact (usage,
productivity, progress). That is what OA is all about.

> journal crisis as driver

Yes, the journal affordability crisis was what first alerted us all
to the access problem, but now it seems to be half-blinding us to the
self-archiving solution: OA is a solution to the access problem, not the
affordability crisis.

> Internet/ WWW as matured enabling technology

Yes, the online medium was what made it possible to implement open online
access (OA). But we have already implemented the medium. We now need to
implement the access.

To put it another way: The *real* road is the Internet, which is already
there. The gold and green roads to OA are "virtual" roads, one implemented
by creating/converting OA journals and publishing our articles in them,
the other implemented by self-archiving our non-OA journal articles to
make them OA.

S/V's terminology is again complicated here, and, once decoded, it
turns out to be just stating the obvious, but without any concrete
implementational content.

> request for ubiquitous connectivity

Isn't the time to worry about imperfections of Internet connectivity
*after* we have provided OA on what there is of it, rather than
before? (It is not the Net we are trying to implement, but OA!)

> Evolution of Berlin Process
>
> Group of drivers: CERN, CNRS, MPG, + x.

So far, these distinguished institutions have signed their names to the
*principle* of OA, which is extremely welcome. But what is needed now is
an implementation of those principles.

To put it another way, one can declare that Autobahn use is a good
thing, but if one wants to be a *driver* (rather than a pedestrian
cheering the principle), one needs to build and use a car!

> representative organization for each country.

Representative of what? And what can and should *countries* do? It is
researchers who write the articles for which researchers seek OA. It is
their institutions that employ the researchers. And it is national (and
non-national) funding agencies that fund their research.

Those are the players. Now what exactly is a *country* to do to
implement OA?

> Implementation of Berlin Declaration

But what does it *mean* to implement the Berlin Declaration? What is one
to *do*, and *how*? Was that not to be the outcome of the CERN meeting? At
least one very concrete and specific implementation proposal -- indeed,
a dual-roadmap -- was submitted:

    http://www.eprints.org/signup/sign.php

but one sees no sign of it in the vague generalities that actually
emerged from the CERN implementation meeting. Nor are they in the S/V
Roadmap.

> building of OA alliance (technical, political)

An alliance between whom and whom, to do what? Until concrete
implementational specifics are stated, this is still just a rally behind
a vague principle!

> identify milestones and deliverables

This is the language of grant applications, before we have even
stated what needs to be done, and how! How about this for a target:

    100% OA provision for all institutional peer-reviewed
    article output by 2007?

And this for milestones:

    25% OA by 2005, 50% by 2006 and 100% by 2007?

Deliverables? Well, apart from the above (which sound like both
milestones and deliverables), it might help to be more
specific about the road-map:

For all institutional researchers publishing in peer-reviewed
journals, the milestones above can be reached in the following way:

    (1) For this particular article, does there exist a suitable
    OA journal (i.e., it is in the right area of specialization, has
    sufficiently high quality standards, etc.)? If so, submit the article
    there, and if accepted there, count that as one step toward the 2005
    institutional milestone of 25%.

This strategy should be successful for about 5% of articles, although
this will vary with the specialty as the 5% OA journals are not uniformly
distributed across disciplines. For covering OA journal publication fees,
5% contingency funds could also be provided or found.

    (2) For the 95% of articles for which there is no suitable OA journal,
    or where the article is not accepted by the OA journal to which it is
    submitted, or where there are no funds available to pay the OA journal
    publication fee, submit the article to one of the 83% of journals
    that is "green" -- i.e., has officially given its green light to
    author self-archiving. Count each accepted, published, self-archived
    article as one step toward the 2005 institutional milestone of 25% OA.
    http://www.ecs.soton.ac.uk/~harnad/Temp/Romeo/romeosum.html

    (3) For articles submitted to the 17% of "gray" journals that have
    not yet given their official green light to self-archiving, either
    ask to self-archive each article on an individual basis (and if
    you get the green light, go ahead), or self-archive the preprint
    prior to submission and link a corrections file after peer-review.
    The 25% and 50% milestones will easily be met via the existing gold
    and green journals, and meanwhile the percentage of green journals
    will continue to increase. (Gold too, but far more slowly.)

The above are concrete means/ends, milestones and deliverables. They
require the implementation of a specific, concrete institutional
OA-provision policy such as:

http://software.eprints.org/handbook/departments.php

> standards for OA platform interoperability and long-term
> archiving.

These interoperability standards, like the Internet and Web itself, are
already in place and growing healthily. Why does an OA implementation
proposal have to break down open doors? What is urgently needed is not
more connectivity and more interoperability but more OA content
provision!

And what is the point about long-term archiving? The self-archived OA
articles from 1991 are still online and OA today, in 2004. What is not
there is all those articles for which OA has not been provided. Nor is
their problem that they have not been preserved: Their non-OA incarnations
were preserved (on paper and online). Digital preservation is not the
problem of OA. Access provision is.

> Include sciences as well as humanities.

All disciplines with a peer reviewed journal article literature should
provide OA to it.

> Education and Awareness

> Communication internally Common OA advertisement material
> by Berlin signatories addressed to coworkers and leaders of
> organisation.

Yes, but what should this communication and advertisement say? Until the
Berlin Declaration signatories commit themselves to concrete
implementation measures such as the above, they are merely propagating a
vague principle (which, by default, amounts to: Wait for more OA
journals to be created or converted, but until then just communicate and
advertise!)

> Communication to political players

Communication of *what*? Communication is fine, once we have settled on
what we want implemented, but otherwise it is empty.

> Signatories lobby politicians to set favorable boundary conditions
> (financial, legal and carrier issues).

What does this mean? Here is something it *could* mean:

    Already in place -- on the part of both research
    institutions/employers and research funders -- is the requirement
    that the research one is funded to do must be published (publish
    or perish). The only additional "boundary condition" needed is to
    extend the boundary of this publishing mandate to: "Publish your
    articles *and* also provide OA to your articles." This does *not*
    mean only, or even primarily: "Publish them in an OA journal," for
    that requirement could only be met for about 5% of articles today. It
    also means publish them in any suitable journal, but also self-archive
    them.

Now OA journal publishing can and should be encouraged and facilitated
by providing the funds to cover any OA journal publication costs, as
well as by encouraging the founding and conversion of OA journals. But
that is a slow and partial boundary condition. The other concrete
steps to be encouraged and facilitated are the creation and filling of
institutional OA Eprint Archives for self-archiving institutional article
output. The milestones and deliverables for the creation and filling of
institutional OA Archives can be digitally monitored, institutionally as
well as globally, as follows:

http://archives.eprints.org/eprints.php

Institutional OA provision can also be encouraged and facilitated
by implementing institutional usage/citation monitors that
automatically compare OA and non-OA research impact (to reward OA
providers and entice non-providers): http://citebase.eprints.org/
http://citebase.eprints.org/analysis/correlation.php

> Signatories encourage learned societies to support open access
> peer reviewing converting proprietary journals into open access

It is always good to encourage non-OA journals to convert to OA, but
that is a hopelessly and needlessly slow road! Far more important is to
also encourage institutional authors to provide OA!

> Discredit impact factor as appropriate measure in career
> evaluation and tenure promotion

This is an arbitrary and *extremely counterproductive* recommendation
and it is hard to see how it made its way into an OA implementation
proposal! The following are facts (references cited at end of this
essay):

    (1) Citation counts are correlated with both journal
    quality and journal rejection rates, hence they are valid
    performance indicators for career evaluation and assessment.
    http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/1111.html

    (2) Citation frequency is also highly correlated with research
    assessment outcomes, and not just when it is consciously counted:
    even when other correlates of quality are used.
    http://www.ecs.soton.ac.uk/~harnad/Temp/self-archiving_files/Slide0007.gif

   (3) Most important, the single most important rationale for OA itself
   is that it maximizes the usage and impact of research by maximizing
   access to it.
   http://www.ecs.soton.ac.uk/~harnad/Temp/impact.html

It is a logical fact that access is a necessary condition for usage
and impact. It would be a *great* shame if the single most important
reason for providing OA were to be obscured or undermined by a completely
orthogonal and speculative crusade against the measurement and usage of
scientometric performance indicators!

I strongly urge dropping this subjective and untested animus against
objective impact measures in the Open Access context, where it can
only cause needless harm, and where it does not belong. This is about
implementation of OA, not about evaluation ideology.

My guess is that the only reason this erroneous non-sequitur was obtruded
into the S/V Roadmap at all was because of the conflation of OA
with OA journals, mentioned earlier, and hence the perceived need to
enhance the value of OA journals by downplaying their citation impact,
for fear that it would be lower than non-OA journals.

Well there is no need to magnify OA journals by diminishing the predictive
power of impact measurement, as it has now been reported by ISI that OA
journal impact factors are no lower (or higher) than those of comparable
non-OA journals. Hence there is no longer any need for special pleading
on behalf of OA journals.

(In any case, the right comparison to make is not between OA and non-OA
journals, which can never be equated, and are like comparing apples and
oranges: the right comparison is between OA and non-OA articles in the
*same* journal and same year and issue. And there the impact of the OA
articles is substantially higher than paired non-OA articles.)
http://www.ecs.soton.ac.uk/~harnad/Temp/isi-impact.html

> Legal Issues

> adoption of creative commons license model by signatory
> organizations.

Why?? There is no need for the CC License model in order to provide OA
to an author's articles published in non-OA journals! 83% of them have
already given their authors the green light to self-archive. And many of
the remaining 17% will do so if asked. To insist instead on a CC license
even if the journal is unwilling is to place a constraint on authors,
and on the prospect of OA provision, that is neither necessary nor
helpful. It puts authors into conflict with their journals where there
need be no conflict. (And, as usual, it does so because of a blinkered
focus on the golden road of OA journal publishing -- where the CC license
is fine, and unproblematic, but only a solution for that 5% of articles
-- ignoring the green road which is able to provide immediate OA for
the remaining 95% without having to enter into needless conflict with
the author's journal.)

> recommendation to co-workers.

Recommend what? To maximize access and impact by self-archiving, or to
enter into needless copyright conflict with their journals?

> seek agreement with commercial publishers to accept use of
> creative commons license.

Why? Why not just take publishers up on the green light they have already
given and self-archive, at least for the 83% of journals that are
already green?

> allow for temporary exclusive license to a publisher as transitory
> solution.

Why again? This is not OA: This is embargoed access, precisely at the
fertile growth region of research when access is needed most! OA is
"immediate, permanent, toll-free, full-text online access."

> Sustainable, technical infrastructure

This is already well on the way, hence irrelevant. What is needed is OA!

> set up technical infrastructure at organizational level.

What is this? The only "infrastructure" needed is the web, and
institutional archives. Both are there already. Archives only need an
(implemented) policy of filling!

> ensure compliance to standards for interoperability (like OAI-
> MPH).

This is breaking down long-open doors, again. The free OAI-compliant
software is available, the archives are up, they merely need an
implemented policy to *fill* them.

> take care for persistence and long-term availability (e. g.
> introduction of globally unique, persistent ID for documents and
> data sets, cooperation with national archiving institutions).
> enforce efforts of communities to establish (selection) standards
> (formats, description schemes) that facilitate long-term
> archiving.

Why all this focus on the preservation of the content of near-empty
OA archives, when the real problem is *filling* those archives?

> Facilitating retrieval of open access material

Not a problem. *Provision* of OA material is the problem.

> unique digital object identifier.

Not the problem!

> mark-up language.

Not the problem!

> standard (community specific) metadata.

Not the problem!

> develop new, appropriate genre types.

???

> develop new, appropriate search and retrieval tools/concepts.

Not the problem!

> Address business models for open access

Business models? OA journals may need business models, but institutional
OA archives do not -- or no more than institutional websites require
a "business model."

What is needed is the concrete implementation of institutional OA
provision policy.

> agree on definition of open access content.

The definition is there already: immediate, permanent, toll-free, online,
full-text access to all peer-reviewed journal articles, by anyone who
has access to the web. And the access must be immediate, otherwise it is
not OA. (OA also needs to be permanent, but the continuing accessibility
of articles in OAI archives that were made OA in 1991 confirms that this
too is a non-problem: Non-existent -- because non-archived -- articles
are the problem!)

> acknowledge that creation and dissemination of research results
> are equally indispensable elements of scholarly communication.

Acknowledged. But these are merely principles, and what is needed for
implementation is concrete implementation details, not more principles.

> acknowledge on an individual level that open access costs as
> research costs (research grants).

This is again focussed on the golden road of OA journal publishing, the 5%
road. What is also needed is a policy for the other 95%, and there it is
not about providing for article publishing costs, but about providing
OA for published articles.

> signatories share on an institutional level a mutually distributed
> load of community-driven communication services.

This is hopelessly vague: Share what load? To do what? The web is
there. All researchers already have access. Establishing OA archives
is trivial. Implementing a policy of filling the archives is not.

> signatories engage in bridging the digital divide.

Banging down open doors again. All countries, rich and poor, benefit
from OA provision, both as users and as providers.

> Institutional Immediate Measures

> enforce open-access publishing policy on all levels of
> organization.

What does this mean? Publishing in OA journals? That's the 5% solution.
What about the other 95%?

> install steering committee at top executive level.

Committee to steer what, when no concrete implementation proposal has
been made, other than publishing in OA journals where available?

> create organizational competence center.

For what? Here is an idea: Proxy self-archiving:
"Let us Archive it for you!"
http://eprints.st-andrews.ac.uk/proxy_archive.html

And there are many other concrete ways to facilitate self-archiving:
http://software.eprints.org/handbook/

> assign open access policy coordinator.

Good idea (once there is indeed an OA policy!):
http://software.eprints.org/handbook/universities.php

> ensure long-term funding and guarantee long-term operation.

Of what?

> Future Actions

> Berlin signatories report on implementation of roadmap in a
> series of meetings.

So far, the S/V Roadmap contains one (golden) road: OA journal
publishing. And one recommendation: Fund it.

That covers 5% of the traffic. Now I suggest adding the (green) road
for the other 95%.

> The drivers group organises meetings every 6 months. Alliances
> for targeting specific tasks of roadmap are recommended.

First organize a concrete implementation plan, then work on getting it
adopted. The right plan is simple, but it needs to be explicitly stated,
as above, not left as vague inferences one might possibly make from
equally vague principles.

Stevan Harnad

    Lee KP, Schotland M, Bacchetti P, Bero LA (2002) Association of
    journal quality indicators with methodological quality of clinical
    research articles. AMA-JOURNAL OF THE AMERICAN MEDICAL ASSOCIATION
    287 (21): 2805-2808

        "High citation rates... and low manuscript acceptance rates...
        appear to be predictive of higher methodological quality scores
        for journal articles"

    Ray J, Berkwits M, Davidoff F (2000) The fate of manuscripts rejected
    by a general medical journal. AMERICAN JOURNAL OF MEDICINE 109
    (2): 131-135.

        "The majority of the manuscripts that were rejected... were
        eventually published... in specialty journals with lower impact
        factor..."

    Donohue JM, Fox JB (2000) A multi-method evaluation of journals in the
    decision and management sciences by US academics. OMEGA-INTERNATIONAL
    JOURNAL OF MANAGEMENT SCIENCE 28 (1): 17-36

        "perceived quality ratings of the journals are positively
        correlated with citation impact factors... and negatively
        correlated with acceptance rate."

    Yamazaki S (1995) Refereeing System of 29 Life-Science Journals
    Preferred by Japanese Scientists SCIENTOMETRICS 33 (1): 123-129

     "There was a high correlation between the rejection rate and
     the impact factor"

Smith, Andrew, & Eysenck, Michael (2002) "The correlation
between RAE ratings and citation counts in psychology," June 2002
http://psyserver.pc.rhbnc.ac.uk/citations.pdf

Oppenheim, Charles (1998) The correlation between
citation counts and the 1992 research assessment exercise
ratings for British research in genetics, anatomy
and archaeology, Journal of Documentation, 53:477-87.
http://dois.mimas.ac.uk/DoIS/data/Articles/julkokltny:1998:v:54:i:5:p:477-487.html

Holmes, Alison & Oppenheim, Charles (2001) Use of citation analysis
to predict the outcome of the 2001 Research Assessment Exercise for
Unit of Assessment (UoA) 61: Library and Information Management.
http://www.shef.ac.uk/~is/publications/infres/paper103.html
Received on Mon May 24 2004 - 22:26:32 BST

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:47:28 GMT