Re: US Federal Research Public Access Act of 2006

From: Stevan Harnad <>
Date: Sat, 27 May 2006 20:32:09 +0100

> My primary aim is to develop a model OA legislation... I am consulting
> the FRPAA, 2006 bill and other texts. In formal legislation, we generally
> avoid any loopholes or pitfalls that may have a counter-effect on the very
> objectives of the legislation. My point here is not to implicate journal
> publishers in the OA legislation unnecessarily, but to develop a robust
> system that is entirely supportive and conducive to building national
> scholarly OA.

That is the optimal strategy. And it is precisely for that reason that
instead of recommending that you emulate exactly the proposals of the
FRPAA, the RCUK, or the European Commission, which all still have some
needless weaknesses as well as loopholes, I recommend that you change
a few of their parameters to make your own policy recommendation the
optimal one:

The changes are small but absolutely crucial:

(1) Mandate *immediate* deposit for *all* accepted papers (*not* deposit after
an interval determined by the publishers, or after a fixed 6-month embargo).

(2) Recommend, rather than mandate, immediately setting the access
to that immediate-deposit to Open Access: allow also the option
of setting access to Closed Access where necessary (i.e., if it would
contravene the author's copyright agreement with the publisher).

(3) This separate dual deposit/release policy requires that the full
text and metadata be deposited immediately, not after a delay; and it
requires that the *metadata* (only) be immediately set to Open Access, so
they are immediately visible and accessible worldwide; but it allows the
option of not setting the full-text itself to OA immediately. In effect, it
thereby transfers the force of any embargo/delay onto OA *access-setting* instead
of onto the deposit itself, which (to repeat) must be immediate.

(4) The US is probably the only country in the world that has enough collective
weight to go even further, because it represents such a large proportion of the
authorship of so many journals, and of the funding of so much published
research: The US can, I think, with impunity put a cap of 6 months on the maximal
allowable embargo period.

The Wellcome Trust has already done this, telling their authors:
"If your publisher does not agree to a cap of 6 months on the embargo,
go to another publisher!"
Wellcome, however, has the advantage, for this admirable and bold move,
that they are a private funder. So all an author can do, if he does
not like the Wellcome terms, is to seek funding elsewhere.

But the US Federal funding agencies are governmental, so they can
be lobbied, not only by publishers but also by researchers and their
institutions, if they object to the terms: And researchers may well
object to being told that they cannot publish in the journal of their
choice if it does not agree to a 6-month embargo cap!

But the dual deposit/release policy removes even this possible obstacle
to the adoption of the policy; for it does not require either switching
publishers or contravening the terms of their copyright agreement. It
allows embargoed access-setting but mandates immediate deposit (and
then semi-automatic email-eprint requests in the repository software
will take care of any gap-period in which the metadata are visible and
accessible but the full-text is Closed Access).

I think this slight parametric change answers your remaining concerns.

> In the case of FRPAA, 2006, one has to be dependent on the mercy of the
> journal publishers, presuming that 93% of them will immediately allow the
> researcher to deposit his research in OA. Please let us know how you have
> reached that figure.

That's 93% of journals, not publishers (as some publishers publish more journals
than others). I have reached the figure on the basis of each journal's own
official author self-archiving policy, as indexed, by journal as well as by
policy, in the Romeo directory, for over 9000 journals, including all the
principal ones:

But note that the dual deposit/release policy would be immune to publisher
objections and policies even if 100% of journals had an embargo,
because it does not mandate Open Access! It simply mandates immediate
repository deposit, which is an internal record-keeping matter, and that is
in no way the business of publishers or anyone else.

What gives the dual deposit/release policy its power is that (a) it also
mandates the immediate deposit of the bibliographic metadata (author,
title, journal name, date, etc.), which no one can embargo, and (b) the
webwide, uniform visibility and accessibility of the metadata immediately
makes the existence of the article visible as well, and allows each
individual would-be user who finds that the article is Closed Access to
use the semi-automatic eprint-request button we have designed for the
GNU Eprints Institutional Repositories (now also implemented for the
Dspace repository software as well, and easily implemented by all other
repository softwares) to request that the author email him the full-text:

So whereas the dual mandatory-immediate-deposit plus
optional-delayed-release policy does not mandate OA per se, it instead
-- in exchange for immunity from all possible objections on the grounds
of either copyright or author choice of journal -- requires immediate
deposit in all cases, no exceptions or delays, and relies on (trivial)
technology to fill any access-needs during any embargo-gap. (It
is unlikely, by the way, that once the dual deposit/release mandate
is widely adopted, embargoes will survive for very long -- and in the
meanwhile, all usage needs will be filled by the software technology,
using the unique capabilities of the new medium).

If the adoption of the proposed mandates instead keeps being delayed
and delayed, as it has been now for the past 3 years, for the very
reasons you are here raising, everywhere (except the Wellcome Trust plus the
 six institutions that have already mandated self-archiving)
-- and even if the mandates are adopted as *delayed deposit* mandates
(as the Wellcome policy is, with a cap of 6 months on the permissible
delay) -- the result will be that embargoes will only become *more
deeply entrenched* instead of being defused, as they are by the
mandatory-immediate-deposit plus optional-delayed-release policy, which
fixes both the date of deposit and the practise of depositing for all
authors and papers uniformly and optimally, allowing no exceptions.

It may be useful for you to bear in mind -- and also to inform your policy
advisees -- that the real thing that has held back 100% OA for a dozen years
now has not been copyright policy but *keystrokes*: Authors have not been
depositing their articles in their institutional (or central) repositories.
Two international, cross-disciplinary surveys for JISC by Key Perspectives
have shown that most authors don't and won't self-archive spontaneously,
but 95% of them will do see if/when mandated to do so by their funders
and/or their institutions
And the actual experience of the (few) institutions that have already
gone ahead and mandated self-archiving confirms the predictions of these
JISC surveys:

The self-archiving mandate should be thought of as a mandate to
perform those all-important keystrokes -- for record-keeping as well as
*metadata-exposing* purposes -- immediately upon acceptance for publication,
which is the natural date, the date when usage of the research can and
should begin, and the date that provides a fixed, objective landmark
for all papers.

(Swan et al. have written an excellent strategic
analysis of the question of institutional versus central
There is also excellent further policy guidance in: )

Once a "Keystroke Mandate" becomes widely implemented, 100% OA will not be far
behind; but keep delaying the keystrokes because of other worries, or
over-reaching, and you only keep delaying 100% OA.

> Nevertheless, a public law can not be dependent on the
> assumptions, if this 93% of publishers tomorrow decide not to allow
> researchers to deposit in OA, the law is then helpless.

Not the law I have just described! It is completely unaffected.
Please also see also the FAQ on this very question:

> In summary, there is a difficulty in constructing a law that guarantees
> 100% of OA even if the researcher/research-institute is willing to do so.

What is needed right now is not a law that guarantees 100% OA; what is needed
is a mandate that guarantees 100% immediate deposit of the full-texts and 100%
immediate visibility of the metadata. Nature will take care of the rest.

But delaying and delaying the adoption of the policy (or weakening it to
ineffectuality by allowing fixed or open-ended deposit-delays in order to
comply with every possible embargo) is a way of guaranteeing that 100% OA will
remain a long, long way off!

The obstacle is keystrokes, and what is needed is a *keystroke* mandate, i.e.,
an immediate-deposit mandate, for internal record-keeping and external
visibility purposes only. That is not an "OA law" but a law specifying
conditions on the fulfilment conditions for receiving public funding for
research: The resulting publications must be immediately deposited in an OAI
Repository and must immediately make their metadata visible.

Once that law is in place, 100% OA will quickly ensue of its own
accord. But keep delaying the law by agonising instead over what will
allow immediate OA for the full text in all cases -- or what interval
can be agreed upon for delayed deposit -- and 100% OA will be delayed
for yet another needless decade,

> Moreover, the semi-automatic email is not open access as the term
> is being used in the legal texts or definition of OA. It depends
> on the will of the author/publishers et al and access is platform/
> technology depended. A legislation has to be platform/ technology neutral/
> independent.

You are quite right that immediate deposit, immediate access to the
metadata, plus the email-eprint feature is not OA! But (1) it is almost
OA, (2) it will rapidly usher in OA, and (3) it is infinitely more
useful to research and researchers, now -- and for the very reasons that
OA is so important and needed -- than continuing to delay and agonise over
legislation that will somehow manage to formally accommodate publisher
copyright agreements, and lobbies, and author choice, all at once, and in

However, it is not true at all that the dual deposit/release "keystroke"
mandate depends on the will of the publishers, nor that access is
platform/technology dependent. All of OA, and indeed the definition
of OA, is dependent on some technology (the Internet and the Web)
and in particular, on the OAI protocol for metadata-harvesting
and interoperability. Hence all Open Access Repositories are
OAI-compliant. That's easy, free, and all that's needed.

Ans the mandate is platform-neutral too: It just requires immediate deposit of
the full-text and immediate webwide visibility of the metadata. The rest will
take care of itself.

As to dependence on the author's will: *All* authors of *all* 2.5 million
annual articles in the world's 24,000 peer-reviewed journals want to have as
many users and citers of their research as they possibly can. That is why, in
paper days, they would take the time, trouble and expense to mail paper
reprints to individual reprint requesters. With the d-d/r mandate, this
is all reduced to one author keystroke, upon receiving the semi-automated
eprint-request generated by the repository software. The only thing now
standing in the way of that option today is the other N-1 keystrokes
needed to deposit the full-text and the metadata into the repository. *And that
is what the d-d/r "keystroke" mandate is for!

Researchers never sought royalty revenue for their articles, and they never
sought to deny access. Mandate the requisite N-1 keystrokes and they will do
the Nth one happily for each individual eprint request (and to re-set access
from Closed Access to Open Access when the embargo expires or they tire of
doing the individual keystrokes -- whichever comes first!).

> Does it mean that there should be an exception to this law in cases,
> 1. where, researcher/ research institutions fail to obtain necessary
> permission from the publishers of the manuscript for reproducing in OA

No, as already stated, no permission is needed to deposit (Closed Access), nor
to expose the metadata (Open Access).

> 2. till the time when researcher/ institute obtain the necessary clearance
> from the publisher of the manuscript

No, the timing of the deposit is now independent of any publisher embargo
periods, which instead apply only to the timing of the Open Access-setting
for the (Closed Access) full text. The policy need make no mention of
embargo periods, except to recommend setting access OA immediately or as soon
as possible.

> Also, in some cases, especially in social sciences, researchers build their
> papers on their research experiences from various sources, not necessarily
> with that of current research funding. So, can government law bind them in
> a strict legal sense to make their publications OA?

No. A law based on conditions of funding (whether funding of research grants or
funding of research or academic institutions) can have no power over what it
does not fund. But it need not. The effects of a self-archiving mandate on
funded research will propagate to non-funded research rapidly as the beneficial
effects of Open Access on research accessibility, impact and progress are
increasingly felt across the entire world research community.

Right now, to repeat: the only thing holding us back from the
feeling of those beneficial effects, and from their propagation across
disciplines and around the world, is *keystrokes* -- not copyright:
keystrokes. Hence an immediate-keystroke mandate is all that is needed.

> The issue that I am raising here, and as reflected in FRPAA, is very simple
> and as such may not have implications for OA. But it [could engender]
> a legal controversy across the three parties. To be very precise, FRPAA or
> for that matter any OA legislation mandates a 'researcher' or a 'research
> institute' to submit the article in a copyright refereed journal.

The (proposed) law does not tell the fundee where to publish. (It
is *already* required that funded findings be published ["publish or
perish"]). The discipline and research community dictates where to publish
(in the highest quality peer-reviewed journal whose content and quality
standards it successfully meets). The d-d/r mandate would add only that the
fundee must deposit the publication (if it is an article rather than a
book) in an OAI-compliant repository (preferably his own institutional
IR) immediately upon acceptance for publication.

> The [FRPAA] says that it adheres to all the clauses of copyright
> law. Now if a researcher has to archive his paper in OA, as mandated
> by the OA Act, he has to comply with the copyright laws as well. If the
> journal publisher allows... the researcher to deposit the paper in
> [immediate] OA it is fine; otherwise the researcher is in danger
> of violating the copyright law of illegally distributing, reproducing,
> copying, extracting copyright material over the internet.

You are not talking merely about *depositing* but about OA-depositing
(i.e., depositing *and* setting access to the full-text immediately to
OA rather than CA). So *if* thus making the full-text freely available online
would violate the copyright agreement with the publisher, the only thing
the FRPAA can do, if it wishes to mandate immediate OA-deposit, is to require
the fundee not to sign such an agreement (as the Wellcome Trust requires, for agreements with embargoes exceeding 6 months).

But such a requirement would in turn be open to the author/institutions
objection that it constrains the author's free choice as to where
to publish.

So the solution is to require deposit only (not OA-deposit), and merely to
encourage OA-access setting whenever and as soon as possible.

There is no other current solution: only the endless pre-emptive
debating about what the publisher would/should/could require and what
the author would/should/could do -- in other words, the effective pre-emptive
"embargo" on the very adoption of a self-archiving mandate that we have been
stalled in now for three years! Continuing that debate simply invites more years
of lobbying, delay, and accumulating access/impact loss.

What is needed is an immediate keystroke mandate, immune to debate or delay.
(Then continue the debate while the keystroke mandate is having its natural

This ground has already been covered, over and over, many times before,
in the past three years. There is no resolution, only inaction. It is
time for action, and the Keystroke Mandate is the requisite action.

> So the solution to this problem cannot be within the technological systems
> of OA, [e.g., the email-eprint button of the IR software], but the OA Act
> would have to make exceptions for such cases. The FRPAA does not mention such
> exceptions - hence it needs to be rectified from this pitfall. A legal
> expert would be able to throw some light on it, I believe.

No, a (lawyer-dictated) inventory of provisos and exceptions will only
make the mandate, confused, confusing and ineffectual: Immediate deposit
of full text and metadata. No exceptions. No delays. And as access-setting
is not at issue, publishers, hence author journal-choice, are both out
of the loop.

Words stand in our way: The purpose of a Keystroke Mandate is of course Open
Access. But call it an "Open Access Mandate" instead and you are up against the
inventory of provisos and exceptions (that will in any case not add up to 100%
OA). It is the tail wagging the dog. Mandate the immediate keystrokes, now, and
then go back to debating the provisos and exceptions.

Nor has technology anything to do with it, other than the online medium
itself, that new medium on which the very concept and possibility of OA are
predicated (Open *Online* Access). Read the opening words of the BOAI
statement: "An old tradition and a new technology have converged to make
possible an unprecedented public good:

OAI-compliance is highly desirable, and easily feasible too, but need
not be mentioned in any law, if you think it will produce a paralysing
technology-dependence debate. Same for the trivial eprint-request button,
implemented in both of the major free softwares for OAI IR creation
(Eprints and Dspace)...

Just mandate the immediate keystrokes and nature will take care of all the

Stevan Harnad
Received on Sat May 27 2006 - 20:50:16 BST

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:48:21 GMT