Re: When is a Journal Open Access?

From: Stevan Harnad <>
Date: Mon, 4 Dec 2006 12:25:05 +0000

Matthew Cockerill is raising again the issues that have already been much
discussed on this list. The gist of our continuing disagreement is that (1)
Matt thinks OA requires further rights to be licensed, over and above free
online access, whereas I suggest that the rest already comes with the free
online territory (and shat doesn't is neither necessary, nor part of OA).

    "Free Access vs. Open Access" (began Aug 2003)

Here is the latest exchange:

On Sun, 3 Dec 2006, Matthew Cockerill wrote:

> Does it make a difference whether the version of an article that is
> made available is the authors manuscript, or the official final
> published version? Yes, that matters.

Yes, but *how much* does it matter, and compared to what?

Being able to access the author's peer-reviewed postprint, versus *nothing*
matters a lot. And that is the status quo for 80% of the research article corpus

But being able to access the publisher's PDF, versus just the author's
postprint matters, matters very little, and it is certainly not the
primary problem today.

And if the depositing of the undeposited 80% of postprints is delayed or hesitated
about because of needless hankering after the publisher's PDF, we are all a lot
worse off.

Forget about anything but the author's postprint for now, insofar as green OA is
concerned. If the publisher's PDF is made OA too, all the better, but consider it
a bonus, not a necessity. The postprint is a necessity, the target of OA
self-archiving mandates, reachable, and already long overdue. The publisher's
PDF is not.

> Does it matter what rights are attached to an article? i.e. whether
> you can reuse it, quote from it, download it for analysis? Yes, that
> matters too.

Matt, I think you have not read the exchange with Peter: All the essential uses
come with the territory if an article is self-archived by its author, free for
all, in his IR: accessing, downloading, printing-off locally, storing
locally, data-crunching. The other re-uses are simply *not part of OA.*

(OA is about *online* access; it was the *online* medium that made
OA possible and reachable. OA never promised the right to print and
distribute multiple hard copies, for example -- although who would bother,
when one can distribute the URL instead, in the online era?)

And journal articles are not like software or empirical data or Disney
film-clips. Their *content* can be used, as always (with attribution
and citation, and within limits, quotation), by all would-be users
(mostly researchers). but their *form* -- i.e., the code, the text --
is not for re-use, or modification, or re-publishing, or re-distribution;
nor does it need to be.

We've been through all this many times, Matt (and with your predecessor,
Jan Velterop, before that), on the "free" vs "open" thread.

> Are these incremental benefits, on top of the limited form of open
> access that Stevan advocates? Yes.

Yes, some of them are, but tiny ones. And holding out for them delays OA: the flea
on the tail, wagging the dog.

> Would Stevan's form of open access be some progress compared to what
> we have today? Absolutely.

It is not my form of OA. It is what OA has always meant: The new possibility of
making journal articles free for all by placing them on the Web for free. It
was the advent of the online medium that opened up this very obvious and
natural possibility, and self-archiving is the fastest and surest way to reach

> But that does not mean that these issues should be ignored or treated
> as irrelevant to open access.

They *are* relevant, but not in the way you are supposing: It would be
very *bad* for OA at this point to insist on a gratuitous and arbitrarily
enhanced "definition" of OA, at the cost of the ordinary vanilla OA that
is fully within reach (and now needs only to be mandated).

> We need to recognize that there is a spectrum of opinion as to what
> is covered by the term open access, and the degree of importance to
> attach to different aspects of it.

It is not opinion that matters now, but getting those articles free
online: 80% of them are not there yet. This is not the time to fuss
about the icing on the 80% missing cake, at the expense of further delaying
the cake.

> Stevan's definition is itself just an incremental step from the non-
> immediate "Free access" endorsed by the DC Principles group:

And free is just a slippery slope away from "lower-priced."

But let's be serious: immediacy is an obvious property of OA. Something
isn't freely accessible if/when it is not freely accessible. If this
causes confusion, attach "Now" to OA and speak of articles as being (or
not-being) OA-Now (rather than OA in year 2050).

And the same common sense applies to the need for OA articles to *remain*
OA. Only a light-headed lawyer (or advert huckster) would flirt with
the idea of "peek-a-boo OA" -- free for 30 seconds, then you pay the toll.

So, yes, making the obvious explicit -- that free accessibility is a
property of an article, and the article has to have the property *now*
(immediately) and continue to have it permanently in order to be OA in
any but a cynical, empty sense -- was an afterthought, something we left
out of the original BOAI version, but so what? It's not too late to fix
it, so it makes sense.

    "Proposed update of BOAI definition of OA: Immediate and Permanent"

> Is such delayed open access sufficient?

Delayed free access is better than no free access (just as temporary
free access is better than no free access, and lower-priced toll access
is better than higher-priced toll access), but it is not OA, and we need
to work for 100% OA.

The difference between embargoed deposits and embargoed access, however --
i.e., the difference between the optimal Immediate-Deposit/Optional-Access
(IDOA) Mandate and the suboptimal Delayed-Deposit Mandates that prevail
at the moment -- is that (although neither is OA), DD reinforces and
entrenches embargoes and delays for years to come, whereas IDOA (with
the help of the EMAIL EPRINT button) neutralises embargoes and will
usher in 100% OA as surely as day follows night.

The one thing to remember is the following fact (not conjecture; fact):

    The only thing that stands between us and 100% OA is keystrokes.

IDOA mandates the keystrokes, and (human) nature will take care of the rest.

    Generic Rationale and Model for University Open Access
    Self-Archiving Mandate: Immediate-Deposit/Optional Access (ID/OA)

> Stevan's goals for open access (which I understand as immediate
> access to some version of the article, somewhere) are stronger than
> those of the DC Principles group, but more limited than those of
> BioMed Central and PLoS, who aim to make the world's research
> literature into a resource that can be freely read, used,
> distributed, downloaded and analyzed.

That is absolutely correct.

But if we keep waiting for the rest of the 24,000 journals to convert to
the Gold BMC/PLoS model, we will be waiting till the heat death of the
universe; whereas Green OA is only a mandate (and some keystrokes) away.

(And, I might add -- though it's only a hypothesis -- 100% Green OA is
also the best hope for an eventual conversion to Gold OA.)

> The point of the Guardian campaign is not that you can't get a free
> weather forecast online - you can.
> But the underlying data for those services, although it has been
> collected at public expense, is not licensed so as to be freely/
> openly available for reuse and this limits the creativity that can be
> applied to making use of this data, and means that we have a much
> less interesting and useful set of tools for working with that data
> than we otherwise could.

But journal article texts are not data, or software, or Disney film clips.
All researchers need is online full-text access (OA): no more, no less.

     "On the Deep Disanalogy Between Text and Software and Between Text
     and Data Insofar as Free/Open Access is Concerned"

> Research articles are themselves data that can be mined and used (for
> example, by bibliographic tools such as Google Scholar, but also by
> more sophisticated text mining tools).

And that comes with the free online territory, so why are we formalizing or
legalising about it here? Google harvests green IR content just as it harvests
gold BMC content.

> Whether articles are free to be used in such a way is surely
> pertinent to their "open access" status.
> [I think even Stevan would agree that if a publisher website or
> repository blocked all robots including OAI harvesters and Google
> from indexing its content, then it would be difficult to consider it
> as providing effective open access.]

If a publisher does, yes, I agree. But we are talking about *author*
self-archiving in their own institutional IRs. And IRs don't block
harvesters, they invite them.

> > SH:
> > 2. Whatever contractual conditions are attached to an "OA article"
> > are meaningless or irrelevant, or overridden by the fact of the
> > article having been posted openly (which has achieved by being OA)..
> Far from being an irrelevance, Creative Commons licenses have been a
> great success in providing a mechanism to flag the fraction of the
> freely accessible web where the copyright owner explicitly permits
> readers to reuse and redistribute the content. (compared to most free
> access content on the web which is explicitly or implicitly flagged
> 'all rights reserved' ).
> [This is sufficiently important that both Google and Yahoo include
> the ability to filter by Creative Commons license on their advanced
> search pages.]

To repeat:

(1) Google can and does harvest all self-archived IR content.

(2) All the *usage* rights that researchers and harvesters need for full-text
journal-article content come with the free online territory (including
linking, downloading, viewing, storing locally, printing-off locally, and

(3) Other "re-use" rights are neither part of, nor needed for, OA (e.g.,
republishing or redistributing online or on paper).

(4) CC licenses are useful, desirable and welcome -- but *not necessary for
OA*, and a deterrent to OA if needlessly insisted upon as an extra

    "Making Ends Meet in the Creative Commons" (Jun 2004)

    "Open Access Data Archiving: A Complement to Article-Archiving"
    (Mar 2005)

> Similarly, the open source movement has depended for its success on
> licenses such as the GNU Public License, the BSD license and the
> Apache license.

True, but irrelevant to OA:

    "How OA is related to OS and FS"

> These type of legalities are not irrelevant and cannot be assumed.
> Just because someone lets me read their source code doesn't mean that
> I can assume I have the right to do what I want of it. The open
> source license tells me just what I can and can't do.

Correct. But peer-reviewed article texts are not source code, and
their authors do not want to give you the right to do whatever you
want with it! They want you to be able to find it, read it, cite it
with attribution, and use its *content* to build your research and
applications upon (just as they did in paper days), including data-mining
and data-crunching the text, if that helps. Unlike with software (which
has not content, only form, code), however, researchers do *not* want
you to modify or republish their "code".

This is a *profound* disanalogy between the nature and needs of OS and OA and
it is remarkable how widely and reflexively it is misunderstood.

> Rather than expecting everyone to agree on a precise definition of
> open access, I think that as with open source software, there will
> continue to be different approaches with different emphasis.
> Is that such a bad thing?

A very bad thing that the arrival, at long lost, of something so simple,
natural, reachable, optimal, inevitable and overdue as free online
access to research articles should be complicated or delayed by needless
formalism and legalism about superfluous definitional details or criteria.

Stevan Harnad

> On 2 Dec 2006, at 11:28, peter murray-rust wrote:
> > At 01:47 02/12/2006, Stevan Harnad wrote:
> >
> > [lengthy and useful explanation of OA snipped...]
> >
> > Thank you very much - this makes it clear and I am sorry for
> > misunderstandings. To summarise:
> >
> > "A journal article is OA if it is accessible free online permanently.
> > OA is about commonsense, not operational definitions or a formal,
> > legalistic business. OA is defined on the article, not on the
> > journal. OA refers only
> > to online access to the online version." [SH quotes]
> >
> > Corollaries:
> > 1. It is an irrelevance or nonsense to talk of OA journals.
> > 2. Whatever contractual conditions are attached to an "OA article"
> > are meaningless or irrelevant, or overridden by the fact of the
> > article having been posted openly (which has achieved by being OA)..
> > 3. The various declarations of OA are overly complex and should be
> > replaced by the first SH sentence above and commonsense.
> >
> > "distractions that have been slowing our progress in reaching the
> > optimal
> > and inevitable. (I have no theory of why we tarry so) "
> >
> > I apologize for the distraction - it came from an over-literal
> > interpretation of the BOAI.
> >
> > Peter Murray-Rust
> > Unilever Centre for Molecular Sciences Informatics
> > University of Cambridge,
> > Lensfield Road, Cambridge CB2 1EW, UK
> > +44-1223-763069
Received on Mon Dec 04 2006 - 12:38:16 GMT

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:48:38 GMT