Free Access vs. Open Access

Stevan Harnad harnad at
Sun Jan 4 20:40:47 EST 2004

The quintessence of the disagreement between (on one side) Mike Eisen
(PLoS) and Jan Velterop (BMC) and (on the other side) myself (and Peter
Suber, Barbara Kirsop, and Sally Morris) is contained at the end of the
very last sentence of Mike's latest posting:

On Sat, 3 Jan 2004, Michael Eisen wrote:

> things like conversion to XML and the right to make articles available
> in different forms at different places - are practical prerequisites
> for the success of open access.

I hope it is evident to any reader who has not lost sight of the true current
state of the peer-reviewed journal literature today that the only prerequisite 
worth talking about -- the big one, the one we are nowhere near having met -- is
toll-free access to the 2.5 million yearly articles in the 24,000 peer-reviewed

This fact is fundamental to understanding practical reality here! To keep
focussing instead exclusively, or even primarily, on the (undeniable,
welcome and admirable) success of the new fleet of BMC and PLoS journals
is to lose track of the actual numbers involved, as well as of time:

The open access being provided today by publishing in those
1000 open-access journals amounts to <5% of the 2.5 million
articles published every year. The articles in those 1000
open-access journals are not the only articles that are openly
accessible, however. About three times as many articles annually
are open-access because their authors have self-archived them:

But even with the help of that second road, annual open access today is
still under 20%. And although open-access provision through self-archiving
is growing faster than through open-access publishing, and *could* reach
100% virtually overnight, even that is still growing far, far too slowly:

There are two (complementary) ways to provide more open access: (1)
by creating or converting 23,000 more open-access journals, finding
funds to cover the author-costs of publishing in them, and persuading
the authors of the remaining articles (80%) to publish them in those
open-access journals instead of the toll-access journals they publish
them in now; or (2) by persuading the authors of the remaining articles
(80%) to self-archive them in their institutional open-access archive
(creating the archives is trivial, and can be done overnight):

*That's* the prerequisite to be met. Without it, all we have is a small and
arbitrary subset of the journal literature. Meeting the prerequisite
is a matter of persuading authors and their institutions and funders
that providing open access is in their own best interests and within
their reach.

Fifty-five percent of journals already support author self-archiving, and
many of the rest will agree if asked.

So the delay in meeting the prerequisite resides not with journal
publishers but with authors, their institutions and funders.

What have XML or republication rights to do with any of this? Absolutely
nothing! They are just ways to while away our time while *not* meeting
the prerequisites.

> I think an important point has been lost in the various threads on this
> topic.
> 1) universal free access to the peer-reviewed literature, in any form, would
> be a wonderful thing - both in its own right, and because it would almost
> certainly lead to universal open access

Then why is PLoS with its considerable resources promoting only
open-access publishing (BOAI-2), instead of also promoting, *at
least* as vigorously, the other road that "would almost certainly lead
to universal open access"? (BOAI-1: self-archiving.) For the same
money, and in the same breath, the dual unified open-access strategy
could be promoted, instead of just open-access publishing:

Unified Dual Open-Access-Provision Policy:
    BOAI-2 ("gold"): Publish your article in a suitable open-access
            journal whenever one exists.
    BOAI-1 ("green"): Otherwise, publish your article in a suitable
            toll-access journal and also self-archive it.

There's a lot more room for immediate growth via the green road than
the golden road right now. And both lead to the same destination. Why
are time (and advertisement space and revenue) being spent on arguing
that "Free access is not Open Access" in BMC's "Open Access Now"
(or in your own postings on the topic, Mike)?

We should instead be doing everything -- together -- to maximize open access,

> 2) the greatest challenge facing open access advocates is convincing authors
> to make their works available freely and/or openly by either publishing in
> open access journals or by self-archiving
> One of the reasons that I and many others are so ardent in our defense of
> the stronger form of open access - one that explicitly permits
> redistribution and reuse - is because we believe that the uses these
> freedoms will enable are a critical part of making open access more
> attractive to authors.

But what is the sense of going for the "stronger" form of access when you don't
even have the "weaker" form yet (and the weaker form ""would almost certainly
lead" to the stronger form)?

How on earth do you imagine that it will help persuade authors to
self-archive if you tell them that self-archiving does *not* provide
open access: that they must also negotiate republication rights and
format in XML! (Especially as you agree that it "would almost certainly
lead to universal open access"?) Adding that further needless handicap
on self-archiving would immediately reduce it to next to 0%!

> A simple example is PubMed. Virtually all scientists who work in fields
> whose journals are included in PubMed use this database as their primary
> tool for searching the literature, and this is unlikely to change. 

Of course they use it. But PubMed is mostly just abstracts, not
full-texts! It's just a search engine. Physics has Inspec. Chemistry has
Chem Abstracts. The rest of the disciplines have ISI's Web of Science.
Of course PubMed, unlike Inspec, Chem Abstracts, or ISI is free, which is
wonderful. But it is still just abstracts. What has that got to do with what we
are discussing, which is toll-free access to the full-texts?

> PubMed is free, simple, efficient and fairly comprehensive, and, with
> links to journals on publisher websites it provides a gateway to the
> online scientific literature. Of course, most of the articles are behind
> toll barriers, and these barriers are not transparent even to scientists
> at the wealthiest institutions.

Mike, you are comparing apples and oranges (abstracts with full texts)
and we are going in circles: It's toll-free access to the full-texts --
the fundamental prerequisite for all of this -- that we are supposed to
be talking about!

Elsevier's Scirus is likewise free, covers
the literature across all disciplines, and likewise dead-ends on publishers'
toll-gates. What is your point?

> An exception are the articles in PubMed Central - these are freely
> available to anyone, with a prominent link provided in PubMed.

PubMed Central contains the full texts of about 100 of the 1000 open-access
journals we were talking about earlier. Again: What is your point? What about the
annual 2,000,000+ articles that neither appear in open-access journals nor are
self-archived? Don't we want to get the message to their authors that they can and
should provide immediate open access to those articles (and how, and why)?

> Because scientists in their role as readers experience the utility of
> PMC on a daily basis, they recognize the advantages of journals that
> deposit their content in PMC when they choose the journals in which they
> publish, and PLoS has received considerable feedback from authors who
> cite immediate availability in PMC as a major reason for their choosing
> to publish with us (I'm sure BMC has had a similar experience).

This is commendable and very welcome, but it only concerns the
contents of the <100 open-access biomedical journals archived in PubMed
Central. There are thousands of toll-access biomedical journals. 

Moreover, scientists and scholars in other disciplines are
having exactly the same experience as PubMed Central users
with, for example, the cross-archive search engine, OAIster which harvests all OAI-compliant
institutional archives. This includes not only all the OAI-compliant
open-access journal archives, but all the OAI-compliant institutional
open-access archives of their own authors' toll-access journal article
output! That means, as noted earlier, open access to at least three times
more journal articles than just those published in open-access journals!
(Citeseer managed it even without OAI-compliance: ).

So researchers are indeed getting a taste of what it would be like if
they had open access to the entire refereed literature -- but a taste
is all it is. It is nowhere near being something they can rely on
exclusively or predominantly yet, for the simple reason that it is still
at least 80% incomplete.

What we should be talking about here is how we can get it 100% completed,
as soon as possible. Not about how it needs to be in XML, or coupled
with republication rights!

> Recently, the NCBI has begun linking their sequence, structure, taxonomy
> and other databases to the full-text articles in PMC, thereby increasing
> their utility and their impact. As people start to use these tools,
> the attractiveness of journals in PMC - especially those that make their
> content available immediately - will grow.

I agree that one of the incentives for open-access provision will come
from linking data to article full-texts. But alas linking can also be done
to toll-access articles! So I very much doubt that data-linkability
alone it will do the trick! (And data-linking can certainly also be done
with self-archived full-texts too.)

> The benefits of inclusion in PMC - and in other services that will begin
> taking advantage of the content published by BMC, PLoS and others open
> access publishers or made truly open access through other means - are denied
> to articles that are self-archived in a way that precludes their reuse and
> redistribution.

Nothing could be further from the truth! There are many powerful search
engines and metadata harvesters (like OAIster and Scirus) that harvest
open-access institutional archives, exactly as if they were all in one big
virtual archive. Once PMC becomes OAI-compliant -- it might already be --
OAIster will harvest it too. That way users can search the open-access
literature in all disciplines, just as they can in ISI's Web of Science.

    "PubMed Central OAI-compliance"

And only metadata need to be harvested. The full-texts can sit in their
own institutional open-access archives -- and be linked to, and accessed,
and downloaded, etc. -- exactly as if the harvester were itself a central
archive like PMC! That was the reason the OAI protocol for interoperability
was created in the first place -- and that is why, for example, we
created the free GNU software so institutions could create
OAI-compliant archives that were all interoperable.

But OAI-compliance is not a necessary feature of open access either! It
is merely an incentive to *provide* open access, because it enhances
functionality, harvestability and discoverability. (Unlike requiring
XML formatting for the text, however, or requiring that republication
rights or the creative-commons license be negotiated  with the journal,
OAI-compliance places no extra burden on the self-archiving author;
it just confers benefits.)

> While it is may be theoretically possible to do some of these things by
> crawling self-archived content, it is a practical reality that relying
> solely on such methods will diminish the attractiveness of open access 

I cannot imagine why and on what basis you are saying this, Mike! It is
not only "theoretically" possible to do *all* of this with self-archived
content, (1) it is already *being* done with self-archived content (example
given above), (2) it was self-archived content that was the inspiration for
creating the OAI protocol, and (3) OAI interoperability is one of the
incentives for self-archiving today.

But even OAI interoperability is not incentive enough. Which brings us
back to the prerequisite underlying all this, which is to find a way of
persuading authors to provide open access for all their articles.

> and is a major reason why I believe that the things that Stevan dismisses as
> "frills" or as organic food for the starving - things like conversion to XML
> and the right to make articles available in different forms at different
> places - are practical prerequisites for the success of open access.

Which brings us back to the point made at the very beginning of
this posting. 

Mike: Your reasoning is circular. We agree that we need open access; we
even agree that self-archiving "would almost certainly lead to universal
open access." But you seem to be forgetting how little open access we
actually have at the moment, and that most of that is not coming from
open-access publishing!

Your efforts are devoted to open-access publishing, which is fine,
and a valid road to open access. But instead of trying to exclude
self-archived articles from the definition of open access, you should
be including self-archiving in PLoS's efforts to promote open access
-- the unified dual strategy -- rather than exclusively promoting
open-access publishing. 

PLoS (and BMC too) should to it (especially in their negotiations with
institutions) not only because it would cost no more to systematically
promote the unified dual strategy rather than just the unilateral one,
but because it "would almost certainly lead to universal open access."

Stevan Harnad

NOTE: A complete archive of the ongoing discussion of providing open
access to the peer-reviewed research literature online (1998-2004)
is available at the American Scientist Open Access Forum:
        To join the Forum:
        Post discussion to: 
    american-scientist-open-access-forum at
        Hypermail Archive:

Unified Dual Open-Access-Provision Policy:
    BOAI-2 ("gold"): Publish your article in a suitable open-access
            journal whenever one exists.
    BOAI-1 ("green"): Otherwise, publish your article in a suitable
            toll-access journal and also self-archive it.

More information about the Jrnlnote mailing list