Friday, April 22, 2005

The role of digital rights management in Open Access

Growing conviction that scientific progress will significantly benefit if scholarly articles and research papers are made freely available on the Web has given rise to the Open Access (OA) movement. While there is some awareness that OA articles may require digital rights management (DRM), there is currently only low-level interest in the topic, with many OA advocates maintaining that it has no relevance to OA.

The issue is complicated by the fact that there are currently two ways in which research papers are made OA, each of which has different implications from a rights point of view.


Richard Poynder

OA has gained a lot of traction over the last year, but it has also attracted considerable resistance from commercial and society publishers. Since they currently generate substantial incomes from selling subscriptions to their journals scholarly publishers fear that if research is made freely available on the Internet these revenues will be significantly threatened.

Given the consequent struggle simply to make Open Access happen many OA advocates argue that worrying about DRM today could prove a distraction from the more important task of "freeing the refereed literature."

Since many also view DRM as synonymous with the use of "technical measures" designed to restrict access, rather than as a broad set of tools for managing rights in a digital environment, there is a tendency to see DRM as an issue for proprietary interests alone.

The danger is, however, that if the OA movement fails to engage with the topic those proprietary interests may set the DRM agenda, to the possible detriment of OA.

Nevertheless, some preliminary work on DRM is being done by the OA movement, and the growing success of the Creative Commons may encourage OA advocates to take a greater interest in the topic.

What is DRM?

Any discussion of DRM in the context of OA has first to seek to define the term. The continuing controversy surrounding P2P and illegal file swapping, for instance, has led many to conclude that DRM amounts to little more than "locking up" content with electronic padlocks. Indeed, since this perceived emphasis on restricting access is viewed as the very antithesis of OA, DRM has become the béte noir of many OA advocates.

What this overly narrow view of DRM overlooks, however, is that digital rights management implies something broader than access control alone. It can also be used, for instance, to ensure correct author attribution, to certify document integrity and provenance, to prevent plagiarism, and indeed to enable creators assert their rights in ways that encourage — rather than restrict — access.

It may be helpful in this regard to view DRM as a two-layered cake. In this model the first layer consists of metadata that define the usage rules (rights) associated with the content. Then on top of this can be placed an (optional) second layer of software-imposed limitations on copying, printing, viewing etc. (i.e. technical measures) in order to enforce the usage rules.

Some OA advocates argue that neither layer is relevant in an OA environment. After all, they say, the aim of OA is to make research papers available to everyone, without restriction.

It may be that the use of technical measures — even for apparently harmless purposes such as ensuring document integrity — will prove "politically" unpalatable for the OA movement (although Frederick Friend's INDICARE article appears to demur on this). There are, however, strong reasons for arguing that the use of rights metadata does have an important role to play, and will for this reason be the main focus of this article.

What authors require

It is clear, for instance, that in making their research freely available on the Web researchers have no intention of giving away their IPR. Their only aim is to allow others to read and build on their work without facing the obstacle of the toll-barriers represented by increasingly expensive journal subscriptions.

In fact we know researchers want to maintain control over their work on the Web because they have told us so. In 2002, for instance, when the JISC-funded Rights MEtadata for Open archiving (RoMEO) Project asked researchers for their views 55 percent of those surveyed (both OA and non-OA authors) said they wanted to limit usage of their works to certain purposes — e.g. educational or non-commercial.

And while over 60% were happy for third parties to display, print, save, excerpt from and give away their papers, they wanted this to be on condition that they were attributed as the authors and that all copies distributed were done so verbatim.

What RoMEO made clear, says Steve Probets, a lecturer in information science at UK-based Loughborough University who was involved in the RoMEO Project, is that "authors are interested in maintaining some form of control over who can do what with their articles."

As Brian Simboli, a science librarian at Lehigh University in Bethlehem, PA puts it: "The shift from toll-access to open access may (illogically) encourage people to assume that the whole concept 'intellectual property' has or should undergo some sort of sea change. Intellectual property is still intellectual property, regardless of how it is accessed."

Some rights reserved

What the RoMEO survey also revealed, however, is that the "all rights reserved" model of classical copyright is more than most researchers want. "[T]he protection offered [to] research papers by copyright law," the report concluded "is way in excess of that required by most academics."

In other words, when releasing their work on to the open seas of the Web OA authors are interested in asserting only some of the rights of traditional copyright (e.g. the right to be named as author), while waiving other rights (e.g. the right to copy or make derivative works). That is, their wish is to make their papers available on a "some rights reserved" basis.

But if researchers don’t make clear to their readers on what basis a paper has been released, how will their readers know? They may mistakenly assume, for instance, that a paper has been made available without any restriction on its use and reuse, as if it had simply been placed in the public domain. Alternatively, they may feel constrained about using a paper in the more liberal way the author intends, for fear of legal reprisal

Consequently, if they dismiss DRM out of hand OA authors risk depriving themselves of a useful mechanism for specifying on what basis they are making their work "freely" available.

Expression of rights

For this reason, in 2002 Project RoMEO began developing an XML-based system designed to express rights and permissions in an OA environment. These issues are not unique to OA authors however. Motivated by the same desire to provide greater licensing flexibility for web-based content, for instance, in 2002 a number of intellectual property lawyers, including Lawrence Lessig and James Boyle, founded Creative Commons (CC).

By separating out the basket of rights provided by classical copyright Creative Commons aims to give creators greater flexibility to mix and match those rights they wish to assert, and those they want to waive.

Selecting from eleven different CC licences creators can, for instance, stipulate that anyone can make whatever use they want of a work, so long as the author is credited; that they can make whatever use they want, provided it is not done for commercial purposes; that they can use it, but not create any derivate works, or verbatim copies; or the work can be offered on a “share-alike” basis — thereby allowing others to make derivative works, but only on condition that the resulting work is then distributed under the same share-alike terms.

Once they have chosen a licence that meets their needs, creators can cut and paste the appropriate HTML code from the CC site and place it on their web page. This puts a visible CC logo on the page, as well as inserting machine-readable metadata into the page's HTML describing the usage conditions associated with the licence. Both passing surfers and search engines can therefore quickly establish what permissions apply to the content attached to the page without needing to contact the creator.

The applicability of Creative Commons to OA was immediately apparent to the Project RoMEO team, who incorporated CC licences into the work they were doing. Explains Probets: "[T]he feelings of the Romeo Project were that the Creative Commons licences would be sufficient to specify the majority of restrictions/conditions required by authors (e.g. that authors are attributed, or that derivative works or commercial uses are allowed)."

Probets, however, questions whether inserting rights metadata into OA papers can be classified as DRM. "I'm not sure that I would regard these licences as a DRM solution", he says. "[They] indicate the ways the work can be used; they do not technically enforce that these conditions/restrictions are applied."

This, however, is surely too narrow a view of DRM. How better to describe the process of inserting machine-readable rights information into digital content in order to control how it is used than "digital rights management"?

Others argue that utilising rights metadata without any means of enforcing their prohibitions is pointless. By the same reasoning, however, we might conclude that it is a waste of time creating any rule, or law, unless it can be physically enforced at the point of potential infringement. We also know that anyone happy to infringe copyright law can circumvent most if not all the electronic padlocks devised to date.

Two roads to OA

For researchers wanting to better manage the rights in their papers, however, there is a more immediate problem than enforcement — namely how they establish and define their rights in the first place. And since there are two ways in which researchers can make their papers OA a one-size-fits-all approach is not currently possible.

For researchers using the "Gold Road" to OA matters are relatively straightforward: they can simply publish in one of the new-style scholarly journals produced by OA publishers like BioMed Central (BMC) and the Public Library of Science (PLoS). By reversing the traditional subscription model and charging authors (or more likely their funders) a fee to publish, rather than charging readers to read, golden publishers are able to make research papers freely available on the Web without any access costs.

More importantly, by treating publishing as a service provided to the author, rather than as a property transaction in which the publisher acquires copyright in return for publishing a paper, both BMC and the PLoS are happy to use the Creative Commons Attribution Licence as a default option. The terms of this licence are printed as a copyright notice on all their articles, as well as being inserted into them as machine-readable metadata.

Why that particular licence? Because, explains PLoS' Andy Gass, the CC Attribution Licence best meets the OA criteria outlined in the Bethesda and Berlin OA declarations. These, he says, specify that in making their papers OA authors grant "to all users a free, irrevocable, worldwide, right of access to, and a license to copy, use, distribute, transmit and display the work publicly, and to make and distribute derivative works, in any digital medium for any responsible purpose, subject to proper attribution of authorship … [as well as] … the right to make small numbers of printed copies for their personal use."

But while the Gold Road is the most logical route for researchers wanting to make their papers OA there are today only 1,600 (out of a total of 24,000) golden scholarly journals in which to publish.

For this reason many researchers opt instead for the "Green Road". Rather than publishing with an OA publisher, they continue to publish in traditional subscription-based scholarly journals, but then "self-archive" an electronic copy of their papers, either on their home pages, or in an e-print archive such as their institutional repository or a centrally-based archive like PubMed Central or arXiv.

However, the rights situation on the green road is complex, since traditional subscription-based journals generally insist that authors assign copyright as a condition of publication. As a consequence, researchers relinquish all control in how their IPR is managed. The RoMEO study, for instance, found that in 90 % of cases authors are asked to transfer the copyright in their papers.

Moreover, while 92% of scholarly journals now allow their authors to self-archive it is a far from ideal solution. As authors are not permitted to use the publisher's PDF, for instance, the self-archived version may be somewhat different from the publisher's version.

More problematically, the rights status of self-archived papers is vague and frequently misunderstood. Indeed, there are reasons to believe that general confusion and uncertainty over copyright represents one of the greatest obstacles to self-archiving today, and perhaps explains why still only 15% of authors self-archive. "The fact is that copyright raises its head all the time when authors are asked about OA, and it is acting as a deterrent to self-archiving," says Alma Swan, co-founder and director of UK-based scholarly publishing consultancy Key Perspectives (KPL). "So it can’t be ignored".

The solution, suggests John Ober, director of the policy, planning and outreach office of scholarly communication at the California Digital Library, is for publishers to "turn their publication copyright policies into the appropriate 'set' of Creative Commons elements"

This would clarify the situation over self-archiving, confirm its legitimacy, and so give self-archiving authors the same transparency over rights as is currently available to those publishing in golden journals. As a consequence OA would receive a significant boost.

Reducing the value of self-archiving

Far from helping to facilitate self-archiving, however, most subscription-based publishers today appear more intent on emasculating it. The fact is that as research funders like the National Institutes of Health (NIH) and Wellcome Trust increasingly encourage researchers they fund to self-archive their papers, publishers are becoming more and more concerned that their revenues are under serious threat. In response, they are actively seeking ways in which they can hobble self-archiving.

Having succeeded in persuading the NIH to water down its policy on public access to research, for instance, more and more publishers are insisting that papers are only self-archived on an embargoed basis, demanding delays of between 6 and twelve months between publication and self-archiving. This, say critics, significantly reduces the value of self-archiving, particularly in areas like biomedicine.

Publishers are also insisting that authors provide a link from the archived version to the official version of the article on the publisher's web site, and that they include the article's unique Digital Object Identifier (DOI). The aim is to drive users away from the free version of the article that has been self-archived, to the for-fee version on the publisher's web site.

The next stage in this strategy may be for publishers to change direction and, instead of prohibiting authors to self-archive the publisher's PDF, to actively encourage it. This would give publishers an opportunity to reassert their ownership of the article, to reinforce their brand, and to charge authors in the process. But the real attraction is that the PDF file format is ideally suited to the use of second-layer DRM (technical measures) enabling publisher-determined usage rights to be incorporated into the articles.

The logic here is compelling. After all, as Chris Barlas, a senior consultant at Rightscom points out, to date scholarly publishers have seen little need for DRM. As he puts it: "[M]ost of the STM publishers currently use some kind of subscription system with password protected access to sites as their form of protection." As scholarly papers increasingly leak out of these proprietary databases, however, publishers will surely want to exploit new ways to protect their proprietary interests.

Certainly Springer Science+Business Media, the second largest STM publisher, has begun to go down this road. While it permits authors to self-archive their own versions of papers, Springer now also invites them to self-archive the final published PDF. To do this, explains Springer's executive vice president corporate communications Sabine Schaub, authors can purchase Springer's PDF file from DRM vendor Aries Systems Corporation, to whom Springer has outsourced the function. Aries will then "download the article from Springer Link [Springer's online database], wrap it with a DRM system called DocuRights, and send it to the author for posting or distribution".

Once it is encased in DocuRights, explains Aries' Lyndon Holmes, the article becomes a "pay-per-view object" with usage rules determined by the publisher. "The publisher can, for instance, specify the number of computers a particular PDF can be opened on". Amongst other things, DocuRights also allows publishers to restrict the number of times a paper is printed and/or viewed.

The attraction to researchers is that using the publisher's PDF allows them to offer the final, definitive version of their article, in a clean professional format. Moreover, since today 78 % of authors who have never self-archived are unaware of how to go about it publishers are clearly in a powerful position to persuade them that archiving a PDF reprint is a better way of providing OA.

However, while authors will still be able to provide Open Access (by themselves prepaying for usage) it is not the kind of solution envisaged by OA advocates.

Take the initiative

Confronted by continuous publisher foot dragging over OA some have concluded that, rather than accepting whatever terms publishers impose, it is time for authors to take the initiative over rights. To this end the Scholarly Publishing and Academic Resources Coalition (SPARC) has produced a downloadable Author's Addendum that researchers can print and attach to the publication agreement publishers ask them to sign on the acceptance of their articles.

The aim of the Addendum is to modify the publisher's agreement to make explicit the fact that the author is retaining sufficient rights to self-archive, and to also require that the publisher provide a free PDF version of the article — one, moreover, with no DRM functionality incorporated into it.

More specifically, explains Michael Carroll a law professor at Villanova University who authored the Addendum, it ensures "that the author retains all rights necessary to grant a Creative Commons Non-Commercial-Attribution License". A second version of the Addendum that will allow the author to simultaneously reserve these rights and then grant the Creative Commons license is now in draft, explained Carroll in a recent post to the liblicence mailing list.

Will this prove acceptable to publishers? While agreeing that "the intent of the Addendum is entirely reasonable," Peter Banks, a publisher at the American Diabetes Association (ADA) responded to Carroll's post by cautioning that several clauses in the Addendum were unacceptable. "Were we presented with this Addendum, we would decline to publish the paper. I am quite sure a majority of publishers would do the same".

In reality it is highly unlikely that subscription-based scholarly publishers will allow authors to manage their own rights. Indeed, many have come to see copyright ownership as key to their survival. While they could adapt by converting to an OA publishing model, most publishers view this as far too risky financially, and certainly less profitable.

Publishers' efforts, therefore, appear to be focused on reducing the impact of self-archiving. Embargoes are one way to do that. A more powerful long-term strategy would be to encourage authors to self-archive the publishers' version and arm it with second-layer DRM. As such, the self-archived article would potentially become a Trojan horse capable of transforming OA articles into "pay-per-view objects".

Such doomsday scenarios are no doubt overblown. But they serve to remind us that ignoring rights issues could prove a risky strategy for the OA movement.

For the moment, however, most OA advocates appear happy to sit on their hands. It is, for instance, nearly two years since the funding for Project RoMEO ended. While its work was inherited by the Open Archives Initiative (OAI) rights group, to date most of that group's efforts have been devoted to developing rights expressions for OA records, not for the underling resources!

This means that even where OA publishers and self-archiving authors include rights metadata in their papers there is currently no OA infrastructure able to exploit those metadata to good effect.

Given the continuing scepticism over rights this is perhaps unsurprising. "It is harmless to make rights explicit in metadata, but that's not the priority", says leading OA activist Stevan Harnad. "The priority is the content (for which these metadata would be part of the decoration)".

In other words, until the number of self-archived papers increases there is no point in fussing over rights. But as Swan points out, uncertainties over rights are a major deterrent to self-archiving today — suggesting the movement may face a chicken and egg stalemate.

Moreover, since the 1,600 gold journals can at most make just 5 % of scholarly research OA such a stalemate would represent a significant obstacle to the wider movement. Harnad insists, however, that all that is necessary today is for governments and other research funders to mandate self-archiving. After that, he says, all the other dominoes will "fall naturally (and anarchically) of their own accord".

But is that enough? After all, the NIH's decision not to mandate (but merely encourage) its researchers to self-archive appears to have been partly influenced by uncertainties over copyright. This suggests that until the copyright situation is clarified uncertainty over rights — and how they are managed — will remain a serious obstacle to OA. What better reason for OA advocates to seize the DRM nettle?

Alternative view

As we've seen, one can view DRM in two ways: as a proprietary and totalising means of locking up content and forcing restrictive usage rules on users in order to maximise revenues; or as a set of tools to help creators maximise usage of their work (without ceding ownership) by specifying what rights they wish to retain and what rights they are happy to waive.

While some question whether the use of Creative Commons licences can be classified as "digital rights management" their heavy reliance on machine-readable metadata to control usage suggests it is entirely reasonable to use the term DRM. After all, why should proprietary interests bent only on locking down content have a monopoly on the term? Why should not this overly proprietary definition be challenged?

More importantly, perhaps, the OA movement faces the clear danger that if it does not more actively promote an alternative view of DRM, then proprietary interests may succeed in foisting a more restrictive model on scholarly publishing, with the risk that some of the OA movement's recent gains could be lost.

With luck, the growing success of the Creative Commons — and the recent founding of the Science Commons — may help OA advocates see the relevance of DRM, and encourage them to promote a broader definition of rights management.

At the very least, by assisting researchers to utilise more liberal Creative Commons licences when publishing in traditional journals, OA advocates could introduce greater certainty about the legitimacy of self-archiving. Not only would this provide a boost to the movement, but it would help to demonstrate that digital rights management is not just about "monetising" content, but is part of a larger initiative focused on creating a rights management regime more suited to a networked environment.

"Personally, I think DRM is really important in the context of OA", says Herbert Van de Sompel, a member of the OAI rights group. "It can, indeed, be about protecting authenticity of works, and avoiding plagiarism … [and] … and even CC licences would cover this. But there is another increasingly important aspect. Readers of the future will more and more be robots that will try and make sense of what they 'read' (by mining content), and present their analysis to humans. It is important that such use be explicitly allowed; in the current environment, one really doesn't know whether it is OK to mine content from OA repositories".

Until there is much greater clarity over rights, and how they are managed, the OA movement may struggle to make significant progress. Increasingly it appears that only by grappling with these complex issues can the movement hope to achieve its objectives.

If you would like to comment on this interview please e-mail your views to me at richard.poynder@journalist.co.uk, or to comment publicly hit the comment button below.

This article was commissioned by INDICARE, and is also available on their web site.

5 comments:

Richard Poynder said...
This comment has been removed by a blog administrator.
Stevan Harnad said...

Richard Poynder has written an (as always) thoughtful and informative article:

Richard Poynder: The role of digital rights management in Open Access. Indicare,
April 22, 2005.
blog version

Nevertheless, Richard's article also contains a few prominent internal
contradictions, and once these contradictions are seen and resolved,
it is Richard's thesis -- the seemingly mild thesis that Digital Rights
Management (DRM) needs to be made a priority for Open Access -- that is
left contradicted.

I will analyze these contradictions in a moment; but first,
Richard's article also contains a very important and potentially
misleading error in the form of a rather shrill exaggeration about the number of publishers that have been inspired to back-slide on their prior "green" author self-archiving policies under the influence of the recent announcement of the NIH 6-12-month
Back Access policy: To my knowledge, that number is exactly one publisher, taking one half-step backwards:

Nature Publishing group has back-slid from full-green [immediate
postprint self-archiving] to pale-green [immediate preprint self-archiving plus postprint self-archiving after 6 months]. There has been no other change in the SHERPA/Romeo
policy database since the NIH announcement. The
only publishers mentioning an embargo are the four that were already either gray (American Chemical Society, Yale U. Press, MIT Press) or pale-green (National Research Council of Canada) before the NIH policy.

This leaves the total percentage of green journals at 92%, exactly as before NIH (and up 37% from the prior year), and merely shifts the pale-green/full-green boundary from 13%/79% as before NIH, to 16%/76%. This does not seem to me like a basis for stating that:

"more and more publishers are insisting that papers are only
self-archived on an embargoed basis, demanding delays of between six and twelve months between publication and self-archiving."

But now let us turn to the basic contradiction: Richard suggests
that those OA proponents who insist that the immediate priority is to
self-archive now (for at least 92% of OA's target content, of which only 15%
is being self-archived today) are wrong. Instead of giving priority to self-archiving -- and to adopting institutional
and research-funder policies that require self-archiving -- we
should give priority to Digital Rights Management. So instead of
acting on the green light we have already from 92% of journals, the
priority is instead to go back and try to negotiate something like a
Creative Commons License with those journals.

And this needs to be done for two reasons: (1) to protect our papers from nonattribution, plagiarism or text-corruption (something that our
existing copyright agreements with our publishers already cover!) and (2) to dispel the confusion about copyright and permissions that is holding back self-archiving.

This re-ordering of priorities is being urged, moreover, in full
cognizance of the fact that publishers are highly unlikely to agree
to it (nor, I might add, will authors, who are already sluggish about
self-archiving, fearful that it might entail some risk or burden for them; their fears would instead be amply confirmed by this readjustment of the priorities).

There you have it. I would say that the resolution of this contradiction is quite clear: Self-archived articles published in green journals (92%)
are already protected from nonattribution, plagiarism and text-corruption by their existing copyright agreements; hence the only priority for them
is to be self-archived, immediately. (For the 8% minority published in
the gray journals, authors can, if they like, try to negotiate a CC-like license with their publishers; but surely the priority is the 92%
majority: otherwise it is the tail wagging the dog.)

And the remedy for author sluggishness about self-archiving and confusion about rights is to adopt clear institutional employer and research funder
policies on self-archiving and to provide clear information on the fact that rights are not at issue when supplementing access to one's own green journal articles by self-archiving them.

What we don't seem to have quite realized yet is that the only real obstacle to 100% Open Access is -- and always has been -- not permissions,
property, plagiarism, and all those other p-words, but
KEYSTROKES. It is keystrokes that authors are (for a panoply of reasons, all groundless) slow to perform. So the only thing that is needed is an institutional/funder incentive to perform those keystrokes (via a suitable variant of the publish/perish requirements and rewards that are already in play everywhere today anyway).

For 92% of OA content, the coast is already clear, and green; for the 8% of the shoreline that is still gray, the keystrokes -- for depositing the
full-text plus the metadata (author, title, journalname, date, author's email address, URL for the official version on the publisher's website)
into the author's institutional repository should be required anyway (for internal record-keeping, grant-fulfillment, and performance evaluation
purposes). But the very last keystroke (the one that sets full-text
access as "institutional access" or "open access"), can be left up to the author for now. For the 8% of "gray journal" papers, their metadata will still be visible to all would-be users webwide, and their authors can, if they wish, just keep doing the keystrokes that are needed to send individual copies by email to all eprint-requesters (just as they mailed paper reprints in paper days) -- until they one day tire of doing all those extra keystrokes and simply hit the OA key. Meanwhile the 92%
majority of papers will be OA.

Now some quote/comments on the text:

"DRM... [Digital Rights Management] can... be used... to ensure
correct author attribution, to certify document integrity and provenance, to prevent plagiarism, and indeed to enable creators to assert their rights in ways that encourage - rather than restrict - access."

Indeed it can be -- but for the 92% of articles published in green
journals, there is no need for any further DRM. What is needed is self-archiving! And what authors need in order to be encouraged to
self-archive is less to do, not *more* to do.

"But if researchers don't make clear to their readers on what basis
a paper has been released, how will their readers know?"

By noting the name of the journal in which it was published, just as
they always did, whether they accessed it on paper or on-line. (And
when the full-text is accessible to all, webwide, many "permission" issues are mooted: simply put the URL in your course-pack!)

"...in 2002 a number of intellectual property lawyers, including
Lawrence Lessig (cf. sources) and James Boyle, founded Creative Commons
(CC)."

Yes, but the CC is for a far
broader constituency of creators than just the authors of the 2.5
million annual articles published in the planet's 24,000 peer-reviewed
journals (each and every one of them an author give-away rather than
a royalty-seeking work). Nor is the broader constituency of give-away
work already protected by copyright, as these published journal articles already are. So the broader set needs CC protection; peer-reviewed journal articles already have it.

"...while 92 % of scholarly journals now allow their authors
to self-archive it is a far from ideal solution. As authors are not
permitted to use the publisher's PDF, for instance, the self-archived
version may be somewhat different from the publisher's version."

So what? The problem today is that 85% of the annual 2.5 million
research articles is not accessible at all to many of its potential users. Is the priority to wait till publishers
agree to the self-archiving of their PDF too? The author's
final, refereed, accepted draft is not enough? Let us not keep
losing all that usage and impact out of abstract, content-free
(literally!) idealism...

On the contrary; let there be a clear difference between the author's self-archived postprint and the publisher's official PDF. Let the author's supplementary version point to the publisher's official PDF, both for scholarly rigor and so that those users whose institutions can afford access to it know where to go to get it. But OA is not about the publisher's PDF: It is about the peer-reviewed
content

"[C]onfusion and uncertainty over copyright represents one of the greatest obstacles to self-archiving today, and perhaps explains why
still only 15% of authors self-archive.

The suggestion to add the burden (and risk of failure) of having to
try to renegotiate DRM/CC with their publishers, over and above the
otherwise simple series of keystrokes (that authors already believe,
wrongly, to be a lot
more of a burden to perform than they really are) is not to reduce confusion and uncertainty but to compound it!

Besides, we have already been given the publishers' likely response:

"Were we presented with this [DRM] Addendum, we would decline to
publish the paper. I am quite sure a majority of publishers would do
the same" (Banks 2005).

To repeat: Ninety-two
percent of journals are already green on self-archiving; but to now to go on and ask them for still more -- to insist on re-negotiating
copyright or on exacting the right to self-archive the publisher's PDF
too -- is not only asking for more than what is necessary for 100% OA,
but for more than what is reasonable or fair to ask of a publisher. Far
better to adopt institutional policies that require authors to do what their publishers have already given them their blessing to do.

"The solution, suggests John Ober, director of the policy, planning and outreach office of scholarly communication at the California Digital
Library...is for publishers to "turn their publication copyright
policies into the appropriate 'set' of Creative Commons elements"

Yes, one can state that as "the solution." But whereas right now the
problem (for at least 92% immediate OA) is to get authors to do the
keystrokes, this "solution" would compound that problem by first also requiring publishers to agree to a lot more -- and a lot more that is
completely unnecessary in order to reach 100%. And a lot more improbable
too, than just getting authors to do the keystrokes (which merely calls for an employer/funder policy that requires it).

"Far from helping to facilitate self-archiving, however, most subscription-based publishers today appear more intent on emasculating it"

Then why encourage them by needlessly asking for more, instead of just
going when the going is green?

"Publishers are also insisting that authors provide a link from the
archived version to the official version of the article on the publisher's web site, and that they include the article's unique Digital Object Identifier (DOI) The aim is to drive users away from the free version of the article that has been self-archived, to the for-fee version on
the publisher's web site.

There is nothing wrong with directing (not "driving") those who can
afford it to the publisher's official PDF! The author's postprint is
for those who cannot afford the PDF. And that is what OA is about: the content, not the form.

"The next stage in this strategy may be for publishers to change
direction and, instead of prohibiting authors to self-archive the
publisher's PDF, to actively encourage it. This would give publishers an opportunity to reassert their ownership of the article, to reinforce their brand, and to charge authors in the process. But the real attraction is perhaps that the PDF file format is ideally suited to the use of second-layer DRM (technical measures) enabling publisher-determined
usage rights to be incorporated into the articles.

To repeat. These alarums are all misdirected. The target isn't the
publisher's PDF (and let that be booby-trapped in any way its creator may see fit: who cares?). The target is the content of the author's peer-reviewed final draft (the postprint). And the problem is that 85% of that OA content is not yet being self-archived. The priority, then, is to provide that 85%, as soon as
possible -- not to continue worrying ineffectually about non-issues like
PDF or DRM!

"Were we presented with this [DRM] Addendum, we would decline to
publish the paper. I am quite sure a majority of publishers would do
the same" (Banks 2005)."

QED

"the NIH's decision not to mandate (but merely encourage) its
researchers to self-archive appears to have been partly influenced by
uncertainties over copyright. This suggests that until the copyright
situation is clarified uncertainty over rights -- and how they are
managed -- will remain a serious obstacle to OA."

The only serious obstacle to OA today is the absence of institutional
policies requiring the self-archiving of institutions' own research article output.
Berlin 3 has now recommended such a policy:

And institutions are at last committing themselves to going ahead
and adopting it and registering and describing it
here for other institutions to emulate..

Stevan Harnad

AMERICAN SCIENTIST OPEN ACCESS FORUM

Richard Poynder said...

I thank Stevan for his comments. He makes a number of points, but I think the key question is the extent to which uncertainty about copyright is acting as a deterrent to self-archiving. The first issue confronting the OA movement is that there is currently no definitive answer as to whether researches can legally self-archive without publisher permission. Of course, most publishers do not object to self-archiving. The second issue then is that since evidence suggests that — regardless of its legality, regardless of the existence of publisher permission — many researchers appear to believe that copyright *does* prohibit them from self-archiving how can the movement resolve the matter?

Given that publishers appear unlikely to abandon their proprietary habits voluntarily Stevan justifiably draws attention to my failure to suggest more concrete ways in which those who support OA can achieve clarity on this matter by assisting researchers to obtain more liberal licences. Let me suggest one: research funders could not *only* mandate that researchers self-archive, but *in addition* mandate them to retain sufficient rights to ensure that they are free to self-archive their papers, and to be able to do so immediately on publication. It is not clear to me why they cannot (as specified in the SPARC Author's Addendum http://www.arl.org/sparc/author/addendum.html) also insist that publishers provide a copy of the published PDF (without any DRM security measures incorporated). This might even help to cut down the number of keystrokes that researchers have to make in order to self-archive.

Stevan is clearly right to say that funders could just tell researchers that they are entitled to self-archive, but as we saw with the NIH plan, clarity on the current situation vis-à-vis rights does not exist even amongst funders. Moreover, where publishers give permission to self-archive, but only after an embargo, I am not clear how the current status quo can ever remove researchers' reluctance to ignore such embargoes without a court ruling on the legalities of self-archiving. For these reasons I suggested that the OA movement needs to address rights issues if it wants to move forward.

It is worth noting that three of the recommendations made by the much-applauded UK Select Committee Report (http://www.publications.parliament.uk/pa/cm200304/cmselect/cmsctech/399/39914.htm) addressed the issue of rights. Most notably one of those recommendations stated: "The issue of copyright is crucial to the success of self-archiving. We recommend that, as part of its strategy for the implementation of institutional repositories, Government ascertain what impact a UK-based policy of author copyright retention would have on UK authors. Providing that it can be established that such a policy would not have a disproportionately negative impact, Research Councils and other Government funders should mandate their funded researchers to retain the copyright on their research articles, licensing it to publishers for the purposes of publication. The Government would also need to be active in raising the issue of copyright at an international level. (Paragraph 126)"

This, along with all the other Select Committee recommendations, was of course rejected by the UK Government. It may be, however, that the forthcoming Research Councils UK policy (http://www.rcuk.ac.uk/whatsnew.asp) will pick up this recommendation and act on it in some way. After all, what the Creative Commons has taught us is that copyright does not have to be a winner-takes-all process. All that is required is that researchers retain *sufficient rights to self-archive*, not *all rights*.

I would add that my intention in the article was not to suggest that researchers stop self-archiving in order first to focus their attentions on ways they can avoid plagiarism, or text-corruption etc. in an OA environment; nor was it to recommend that resolving rights issues be given priority over self-archiving. My point was that since copyright appears to be acting as a significant deterrent to self-archiving the OA movement is more likely to achieve its objectives if it addresses the issue of rights, rather than insisting that there is no issue.

Richard Poynder said...

Stevan Harnad said:

> Is it not merely fanning fears (without foundation) to insist
> that the re-negotiating DRM is necessary because "more and more" of
> the 92% green are back-sliding, when to date only one green publisher
> (Nature) has back-slid, and only from full-green (immediate postscript
> self-archiving) to pale-green (immediate preprint self-archiving plus
> postprint self-archiving 6 months after publication)?


However small the overall impact of Nature's back-pedalling may be in itself, it teaches us that relying on "permissions" from publishers is an inherently unstable strategy for the OA movement, since these permissions can be withdrawn at any point in time. Moreover since publishers do not create the material that they publish, a more appropriate way of organising things would be for them to seek permissions from authors, not the other way round.


> > Of course, most publishers do not object to
> self-archiving. The second
> > issue then is that since evidence suggests that --
> regardless of its
> > legality, regardless of the existence of publisher permission --
> > many researchers appear to believe that copyright *does* prohibit
> > them from self-archiving how can the movement resolve the matter?
>
> Very simply. And the researchers themselves have told us exactly how:
>
> Alma Swan's two international, multidisciplinary surveys have reported
> that 79% of authors reply that they do *not* self-archive, and will
> not self-archive, until and unless their employers or funders
> *require* them to do so; but if and when they do require them to do
> so, they reply that they
> *will* self-archive, and will self-archive *willingly*. (Only 4%
> replied that they would not self-archive even if required; I would say
> a 96% resolution counts as a solution, and is vastly preferable to the
> tail wagging -- or holding back -- the dog!)


And as Alma Swan elsewhere says: "The fact is that copyright raises its head
all the time when authors are asked about OA and it is acting as a deterrent
to self-archiving. So it can't be ignored."
(http://poynder.blogspot.com/2005/04/roller-coaster-ride.html).

Indeed, Alma is not the only person to have made this point. Here are the
words of Stevan Harnad:

"[W]hat really needs changing is journals' copyright policies, which are a
perceived deterrent to self-archiving refereed papers (even though they can
be circumvented completely
legally)" http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/1241.html

Since research funders seem equally unclear about the current rights
situation there appears to be considerable confusion all round.


>Richard! You register first, as with all archives, and then you can do the
keystrokes: Please let me know the timing!


I tried it but had some difficulties. Specifically, I could not establish whether the article I was trying to post had been accepted by the system. After 20 minutes trying to work this out I gave up. Perhaps this is because it is only a demo system, but I think the first thing a researcher would want to do is check that his paper has indeed been uploaded — by doing a search to see if it is there. It might also be because I am not as technology literate as I would wish, but then many researchers might be similarly handicapped!


> > Most notably one of those recommendations stated: "The issue of
> > copyright is crucial to the success of self-archiving. We
> recommend
> > that, as part of its strategy for the implementation of
> institutional
> > repositories, Government ascertain what impact a UK-based policy
> > of author copyright retention would have on UK authors. Providing
> > that it can be established that such a policy would not have a
> > disproportionately negative impact, Research Councils and other
> > Government funders should mandate their funded
> researchers to retain
> > the copyright on their research articles, licensing it to
> publishers
> > for the purposes of publication. The Government would also need
> > to be active in raising the issue of copyright at an international
> > level. (Paragraph 126)"
>
> This is actually a very reasonable recommendation, with
> exactly the right priorities. Notice that it does not
> recommend that this further research on the potential impact
> of a potential copyright-retention policy be performed
> *before* or *instead of* or as a *precondition for* the
> recommended self-archiving mandate. It quite sensibly
> recommends this bit of research to be performed in the
> background, in parallel. Nor does it prejudge the outcome; it
> merely says that if the outcome does not prove to be too
> negative, a copy-right retention policy could then be
> mandated too. Who can disagree with that?
>
> But meanwhile, the priority remains an immediate
> self-archiving mandate.


Absolutely. We are in agreement. More clarity is needed.


> The reality is that 92% of journals have given self-archiving
> their green light, yet 85% of authors are still just parked
> and not moving. Prodding them needlessly and irrelevantly to
> move on retaining copyright when they are not even moving on
> self-archiving just seems to be inviting two forms of
> sluggishness in place of one.
>
> I know you weren't recommending that the 15% who *are*
> self-archiving should stop and renegotiate DRM instead! But
> the 15% self-archivers are not the problem: the 85%
> non-self-archivers are. And what is needed to set them in
> motion is an institutional keystroke mandate, not yet another
> thing they can stall on (and this time a 92% unnecessary thing)!


And my point was that many in this 85% group of authors are still parked
because the rights situation is uncertain and they don't want a traffic cop
to pull them in when/if they move on to the highway. So let's clarify the
rights and allow the parked authors to move off into the self-archiving fast
lane - so long, of course, as they have been mandated to self-archive by
their funders. Since funders also have it in their power to clarify the
rights situation - by insisting that researchers retain sufficient rights to
self-archive - they can apparently achieve both aims at one fell swoop.

Richard Poynder

Stevan Harnad said...

On Tue, 26 Apr 2005, Richard Poynder wrote:

RP: "my point was that many in this 85% group of authors are still parked
because the rights situation is uncertain and they don't want a traffic cop
to pull them in when/if they move on to the highway. So let's clarify the
rights and allow the parked authors to move off into the self-archiving fast
lane - so long, of course, as they have been mandated to self-archive by
their funders. Since funders also have it in their power to clarify the
rights situation - by insisting that researchers retain sufficient rights to
self-archive - they can apparently achieve both aims at one fell swoop."


Perhaps this allegory (in which author/researchers are cast
as delivery-truck drivers, delivering research impact for their
employers/funders) will help get the logic of the situation into better
focus:

U-HAUL ALLEGORY

When 85% of delivery-truck drivers are parked motionless facing
traffic-lights, 92% of which are green, one need not (1) seek a court
ruling on the legality of going ahead on green! Nor need one (2) seek,
for each individual driver, the personal legal right to turn the light
green at will of his own accord, when 92% are already green! The
drivers' paymasters can propel them into motion by reminding them
of the logical and legal contingency between green lights and going
ahead as well as the professional and practical contingency between
one's salary and making one's deliveries....


RP: "as
Alma Swan
elsewhere says: 'The fact is that copyright raises its head
all the time when authors are asked about OA and it is acting as a deterrent
to self-archiving. So it can't be ignored.'"


I think that by "can't be ignored" Alma meant mostly (92%, to be exact)
informing drivers about the logical/legal contingencies, as well as
firming up the practical/professional contingencies.

RP: "Indeed, Alma is not the only person to have made this point. Here are the words of
Stevan Harnad
: "'[W]hat really needs changing is journals' copyright policies, which are a
perceived deterrent to self-archiving refereed papers (even though they can
be circumvented completely legally)'


"Since research funders seem equally unclear about the current rights
situation there appears to be considerable confusion all round."


Richard, that quote was from Tue Mar 27 2001 - 19:09:15 BST!

Things have, fortunately, progressed since then -- though more on the
publisher (traffic-light) side than on the author (delivery) side, with
author self-archiving only increasing from about 10% to 15% during that
4-year interval while journal green-lights increased from perhaps <10% in 2001 to 92% today (under pressure from the demand for OA including,
amongst others, the 34,000 authors who signed the PLoS Open Letter
threatening to boycott their journals by September 2001 otherwise!) The
present situation is not without its ironies and internal inconsistencies!
But inconsistencies are there to be pointed out and corrected:
"A Keystroke Koan For Our Open Access Times" (2003)


The second part of my quote, by the way, is
still valid and pertinent
today but it is mooted by the progress on
the first part, 92% of journals now being green; for the remaining 8%,
copyright re-negotiation is one option; the preprint-plus-corrections
strategy is another; and the
"keystroke mandate"
is another:

SH: "to date only one green publisher (Nature) has back-slid, and only from
full-green (immediate postscript self-archiving) to pale-green
(immediate preprint self-archiving..."


RP: "However small the overall impact of Nature's back-pedalling may be in
itself, it teaches us that relying on "permissions" from publishers is an
inherently unstable strategy for the OA movement, since these permissions
can be withdrawn at any point in time. Moreover since publishers do not
create the material that they publish, a more appropriate way of organising
things would be for them to seek permissions from authors, not the other way
round."


When the shade of green for one publisher changes (yet remains green,
the 92% figure unchanged), this is hardly grounds to call for a bottom-up
reconstruction of the traffic grid! Let us reconstruct the grid if and
when we ever need to. The priority right now is making those deliveries!

"32. Poisoned Apple"


RP: "I tried [Demoprints] but had some difficulties. Specifically, I could not establish whether the
article I was trying to post had been accepted by the system. After 20
minutes trying to work this out I gave up. Perhaps this is because it
is only a demo system, but I think the first thing a researcher would
want to do is check that his paper has indeed been uploaded - by doing
a search to see if it is there. It might also be because I am not as
technology literate as I would wish, but then many researchers might be
similarly handicapped!"


Richard, many thanks for that feedback! You are quite right that depositors
require rapid feedback and the reward of seeing of their paper in the archive
(if there are no metadata problems to fix) as soon as possible.

I note that there was a backlog of 49 deposits in the Demoprints buffer.
I have now cleared most for deposit myself (some had errors to fix) and
will ask Chris Gutteridge to make sure someone checks the Demoprints buffer
at least every week.

A week is too slow a turnaround for a real institutional archive, however, so
I hope institutions will make sure that their buffer is processed daily or even
more often, especially in the all-important start-up phase when the idea is
to encourage and reward all self-archivers!

But Richard, your own deposit did not even make it to the buffer: The sequence
is: (1) Register (and wait for confirmation). [I assume you went through this
1-time phase successfully, if you got the the deposit stage.] (2) Deposit the
metadada and the full-text of a paper. (3) Wait for confirmation that it has
been processed (usually after someone has checked that the metadata ar ok and
that the full-text is indeed attached and readable).

You must have left out a step in (2) because your deposit never even reached
the buffer for checking. (Please do try again, or let me know what happened!)

Stevan Harnad