Welcome to the SPARC Open Access Newsletter, issue #164
June 2, 2012
by Peter Suber
In the past few months we've seen a sharp upturn in the momentum for OA on several fronts: researcher activism for OA, researcher criticism of high journal prices, researcher understanding of OA issues, OA policy deliberations at major institutions and governments, and mainstream media news and comment.
Here I want to focus on a relatively neglected front where progress is also accelerating: libre OA. I've been tracking gratis and libre developments for years, and I've never seen such a sudden rise in the number of policies and initiatives going beyond gratis to libre, and going beyond the more restrictive end of the libre spectrum to the less restrictive end. The only pity is that the other sudden spikes of OA progress have tended to overshadow these developments.
(1) Defining the terms
If I'm going to spend time on this topic, I should define my terms. "Gratis" access is free of charge. "Libre" access is free of charge and free for some kinds of further use and reuse. Gratis access is compatible with an all-rights-reserved copyright, which allows no uses beyond fair use (or the local equivalent). Libre access is not compatible with an all-rights-reserved copyright, and presupposes some kind of open license permitting uses not permitted by default. As I've sometimes put it, gratis removes price barriers alone and libre removes price barriers and permission barriers.
There is only one kind of gratis access because there is only one way to make a work free of charge. But because there are many permission barriers that we could remove if we wanted to, libre access is a range or spectrum. When we want to refer to specific types, we can use named licenses. For example, CC-BY and CC0 lie at the upper or most-free end of the libre spectrum. The CC-BY license allows any use provided the user makes proper attribution to the author. CC0 puts a work into the public domain and in that way allows any use whatsoever.
In addition to the spike of recent progress for libre OA itself, there has been a spike of recent discussion of the "gratis" and "libre" terminology. See Postscript 2 below for a longer digression on these terms in response to some of that discussion.
Quick preview: Some want the term "libre" to refer only to the most-free end of the spectrum beyond gratis, not to the whole spectrum beyond gratis. That's a discussion worth having. Meantime, this article covers libre progress in the wider sense, or in the whole spectrum beyond gratis, and includes many developments about libre in the narrower sense (at the CC-BY/CC0 end of the spectrum). Hence, no matter where you stand on the terminology, there's progress here worth noting. We shouldn't let nomenclature disputes hide that fact.
Since I'll also be discussing "green" and "gold" OA, let me recap those definitions as well. Green OA is OA delivered by repositories, regardless of peer-review status, gratis/libre status, funding model, embargo period, and so on. Gold OA is OA delivered by journals, regardless of peer-review methods, gratis/libre status, business model, and so on. It should be clear that the green/gold distinction is not the same as the gratis/libre distinction. Green/gold is about venues or vehicles, while gratis/libre is about user rights. For better or worse, there are four cases to keep distinct: gratis green, gratis gold, libre green, and libre gold. Most of this article is on libre green, with a few remarks on libre gold.
(2) Past paucity of libre green OA
Until recently, libre green OA policies for research articles were relatively rare, and we need to step back for a minute to appreciate that. Here's a quick chronology.
The Wellcome Trust and UKPMC Funders Group broke the ice in 2007 by requiring libre green OA whenever they paid any part of the costs of publication.
But not even this policy required libre when the funder paid for the underlying research without paying the costs of publication, not even when the research costs were more than one hundred times larger than the publication costs. Nevertheless, requiring libre when paying the costs of publication was a natural first step.
When authors or their sponsors pay publishing costs, then publishers no longer need to protect a revenue stream to cover the same costs. Once freed of that business constraint, publishers can and ought to make the work libre OA. Conversely, those paying the costs of publication should demand libre in return for their payments. This should apply to universities an individuals, not just to funders.
Nevertheless, even today, five years later, other funders have been slow to follow the lead of the Wellcome Trust and the UKPMC Funders Group. In my March 2012 comments on the draft new OA policy for the Research Councils UK, I recommended that the RCUK adopt this sensible and well-tested policy. But this recommendation was only necessary because the RCUK was not already proposing to take this step.
The NIH has not taken this step either. Nor have most of the other funders who routinely require gratis OA to the research they fund, even when they are willing to pay the costs of publishing the results.
Meantime, most of the contents in most OA repositories are gratis and not libre. There's a good reason for this. Repositories are not in a position to permit OA on their own and depend on the permissions granted by rightsholders. Nevertheless, libre green OA has been growing for more than a decade.
In 2001, only 7% of the articles deposited in UK PubMed Central (UKPMC) carried open licenses permitting reuse. By 2009, that percentage had grown to 33%, and in 2010 it jumped to 41%. In each of these years, of course, 100% of deposited articles were gratis OA.
The total volume of UKPMC was also growing in the same period, which means that the libre green deposits would have grown in absolute numbers even if their percentages had remained flat or even declined somewhat. In 2009, there about 50,190 articles in UKPMC, in 2010 there were 69,123, and in 2011 there were almost 92,000. To have libre percentage growth on top of that absolute growth is especially heartening.
In the US PubMed Central, the numbers are lower but the curve is still sloping upwards. In his interview with the NIH Office for Extramural Research last month, Richard Poynder reported that there were 2.4 million articles in PMC, and 450,000 in the libre OA subset. That makes the libre portion 18.75% of the whole, presumably for 2012.
(Hunch: The UK PMC percentage is higher than the US PMC percentage precisely because members of the UKPMC Funders Group mandate libre green OA when they pay any part of the costs of publication.)
The Harvard OA policies should be considered libre green policies, starting with the first in February 2008. The policy grants the institution a wide set of non-exclusive rights, and the terms of service for DASH, the institutional repository, take good advantage of them. When an article is in DASH, and when the permission to make it OA comes from the policy, then users are allowed to "use, reproduce, distribute, and display the...Article for: ...personal study; ...teaching (including distribution of copies to students and use in coursepacks and courseware programs); ...research and scholarship (including computational research uses such as data-mining and text-mining); and provision of value-added services (including full-text searching, cross-referencing, and citation extraction)...." That's not CC-BY, but it's way beyond fair use.
There are more than 30 university OA policies based on the Harvard model. All of them are at least potentially libre, by granting the institution enough rights to authorize libre access through the repository. But whether they are actually libre depends on what the repository actually authorizes. Unfortunately, no one has gone through all the repository terms-of-use statements to see how many allow visitors to exceed fair use and in what ways. (But if you're thinking it would be a good idea to check, I agree.)
In November 2008, the University of Liege enhanced its existing OA mandate with a license equivalent to CC-BY-NC-ND for deposits in its repository.
In May 2009, the University of Oregon Library Faculty adopted an OA mandate requiring their publications to disseminated under CC-BY-NC-ND licenses.
In October 2010, a $20 million funding program from the Gates Foundation, the Next Generation Learning Challenges, mandated libre OA under CC-BY licenses for publications arising from the funded research. I wish I could say this was just the start of a string of other libre OA mandates (or even gratis OA mandates) from the Gates Foundation, but I cannot.
In July 2010, eIFL reported that three OA repositories in China, Poland, and South Africa provided libre green OA under CC licenses, and that three others in Botswana, Poland and South Africa recommended it.
In addition from two schools within Harvard, six institutions adopted libre green policies in 2010: the Library Faculty at Arizona State University, Australian National University, the Library Faculty at Northern Colorado University, University of Sassari, Sweden's Royal Library, the Washington State Board for Community and Technical Colleges (SBCTC) on behalf of 34 institutions.
In January 2011, the US Departments of Labor and Education jointly announced the Trade Adjustment Assistance Community College and Career Training (TAACCCT), a four-year, $2 billion funding program for open educational resources (OER) mandating libre OA under CC-BY licenses. This was the first libre OA mandate from the US federal government, and I believe it is still the only one.
(While I only know the one libre OA mandate from the Gates Foundation that I noted above, in April 2011 Gates awarded a grant to Creative Commons to support grantees funded by the TAACCCT libre OA mandate.)
In January 2011, eIFL updated its report on open licenses in developing and transition countries. This time it found 10 repositories offering libre green OA, including one OER repository in South Africa using CC-BY.
In June 2011, the Open Society Foundations (OSF) adopted a policy encouraging grantees to make their funded work libre OA under CC-BY-NC-ND licenses.
Reminder: The CC-BY-NC-ND license used at Liege, Oregon, and OSF is libre in the broad sense (beyond gratis) but not in the narrow sense (toward the CC-BY end of the spectrum). No matter where you'd turn the libre knob, access under this license is still an advance upon merely gratis access. If you think it too weak to include it in a list of libre developments, then you will acknowledge the main point of this section, which is that stronger policies were rare during this period.
If we don't count open data policies, which are libre much more often than OA policies for publications, then the TAACCCT and OSF policies, and the OA policy at one more school at Harvard, were the only libre green policies adopted in 2011. (If I'm overlooking any, please let me know!)
That's the sparse history of libre policies until quite recently. But supplementing these adopted policies were some important recommendations. (I'm limiting the list of recommendations to those from 2009 and beyond; otherwise we'd have to go back to the Budapest statement and the list would be very long.)
In May 2009, a group of major public and private funding agencies convened by the the US Institute of Medicine called on funders of medical research to mandate libre green OA ("without constraints of copyright"). The group included the Gates Foundation, Burroughs Wellcome Fund, Merck Company Foundation, Rockefeller Foundation, US Department of Health and Human Services, US Department of Homeland Security, and US Department of State.
In August 2009, a report from Phil Malone and the Berkman Center recommended that foundations "seriously consider" requiring libre OA under open licenses for funded work.
In November 2009 the National Book Trust of Uganda recommended libre OA books as the "best way for boosting educational quality" for Ugandan students.
Two bills introduced in the US Congress in 2009 would have required libre OA to federally-funded textbooks. Neither bill passed.
In January 2010, the Scholarly Publishing Roundtable stopped short of recommending OA mandates for publicly-funded research in the US, but didn't hesitate to recommend that "the results of research...be published and maintained in ways that maximize the possibilities for creative reuse...."
In January 2010, the joint OA mandate for Finland's 26 Universities of Applied Sciences took effect (it was adopted in October 2009). It didn't require libre green, but recommended it and automatically offered each author the option of attaching the CC license of their choice to each article deposited in the consortial repository.
In February 2010, the Online Guide to Open Access Journals Publishing recommended libre gold OA under CC-BY licenses.
In August 2010, SPARC and Science Commons published a white paper by Simon Frankel and Shannon Nestor, two lawyers at Covington and Burling, on legal issues in university OA policies. Frankel and Nestor recommended the Harvard-MIT model, which allows libre OA through the institutional repository.
In January 2011, the Ghent Declaration to the European Commission called for OA policies to require libre OA under CC-BY licenses.
In October 2011 the FORCE 11 Manifesto (for Future of Research Communication and e-Scholarship 2011) called on researchers to distribute their work in "conformance to OA licenses" and commit "to make all [their] own scholarship as open as possible under the most liberal of those licenses."
(3) Past paucity of libre gold OA
Libre OA through repositories has been rare because most repositories are not in a position to demand it or even to authorize it. Hence, you might think that libre OA through journals would be common because all journals are in a position to do both. But unfortunately that would be wrong. The power of journals to demand and authorize libre OA means that libre gold could be common, and should be common. But scandalously, it doesn't mean that libre gold is already common.
The Directory of Open Access Journals (DOAJ) lists 7,774 peer-reviewed OA journals. (The numbers in this section were current as of May 25, 2012.)
Of these, only 2,208 use some kind of Creative Commons license. That's only 28.4% of the full set.
A few OA journals use home-grown licenses equivalent to CC licenses, but they are relatively rare. For present purposes we can say that roughly 70% of OA journals don't use any kind of open license. Hence, about 70% offer merely gratis OA and use all-rights-reserved copyrights.
Only 917 journals in the DOAJ have the SPARC Europe Seal of Approval, which requires CC-BY. That's only 11.8% of the full set.
To be fair, the SPARC Europe Seal has two requirements: using the CC-BY license and sharing metadata with the DOAJ. Hence, some journals may fail to earn the seal because they fall short in their metadata-sharing practices, not in their licensing practices.
But for present purposes we can say that roughly 88% of OA journals don't use CC-BY.
I've previously argued that OA journals have "no excuse" not to provide libre OA under open licenses, and that the failure of the majority of them to do so "is one of the largest missed opportunities of the OA movement to date."
The failure of 70% OA journals to offer any kind of open license is an embarrassment. It shows that most OA journals don't understand the benefits of libre OA, don't understand their own power to assure it, or both.
The most common response I've heard from merely gratis OA journals is that they wish to block commercial use. But that is not responsive. A CC-BY-NC license would block commercial use while still freeing users to exceed fair use in other respects. The many voices recommending CC-BY (including my own) should not obscure the fact that CC-BY-NC is much friendlier to users and research than an all-rights-reserved copyright.
For the present argument, my main point is that libre gold is rare too, even though it faces none of the impediments of libre green. In fact, the percentage of journals in the DOAJ offering libre gold OA is smaller than the percentage of articles in UKPMC offering libre green --an unexpected and disappointing result. More disappointing: the recent upturn in libre green progress has no counterpart libre gold progress. Libre gold is lower-hanging fruit than libre green, but it remains largely unplucked.
(4) Historical timing of libre green policies
Libre green policies have been scarce for a couple of good reasons, apart from the fact that most repositories are not in a position to authorize it.
First, few publishers are willing to allow libre access. Most green OA, for example, is made possible by permissions from toll-access (TA) publishers, and conversely, most TA publishers permit green OA. But nearly all TA publishers willing to permit *gratis* green OA are unwilling to permit *libre* green OA.
Second, funding agencies and universities have their own reasons to adopt strong OA policies in stages, and to put gratis before libre. They worry that libre green mandates would trigger even higher levels of publisher resistance and opposition than we see today, and make it harder for authors bound such policies to publish their work. This concern is not answered by rights retention. For even when authors retain the right to authorize OA, publishers remain free to refuse to publish any work for any reason.
I think this concern is warranted, or has been warranted, and I've raised it several times over the years. Each time, however, I've urged funders and universities to watch for the moment when they could safely strengthen gratis policies to libre.
http://www.earlham.edu/~peters/fos/newsletter/02-02-09.htm#choicepoints (Sections 4 and the conclusion)
http://legacy.earlham.edu/~peters/fos/newsletter/01-02-11.htm#2010 (Section 2)
https://plus.google.com/109377556796183035206/posts/Y8zPSf5DP5W (Comment #1)
Alma Swan made a similar point in her new book for UNESCO on Policy Guidelines for the Development and Promotion of Open Access (April 2012). From Section 8.2.7 at p. 48: "Requiring libre Open Access is considered a step too far at present, despite its promise for science, as it would make it very difficult for authors to publish in journals of choice because of publisher resistance. It is an issue for future policy, though that future will not be too far ahead."
Publishers who oppose a strong OA policy at a small funder or university could refuse to publish work by those grantees or faculty. But when large funders and universities, and especially when many funders and universites, adopt strong OA policies, publishers can no longer afford to refuse to publish their work.
For example, the NIH is the largest funder of non-classified research in the world, with a research budget larger than the gross domestic product of 140 nations. It's no accident that while many publishers speak against it, not a single surveyed publisher refuses to publish NIH-funded authors.
This is one reason why the libre arc is bending. Some early steps have been taken, some large OA-friendly institutions are warming to libre, many OA-friendly institutions large and small are no longer willing to subordinate their interests to the interests of publishers, and the only players who might have been hurt by premature libre mandates --authors-- are joining the call for stronger OA policies. There's no decisive historical turning point when the concerns that previously held back libre policies are suddenly answered and powerless. So we can't say that the moment has arrived when funders and universities can strengthen green OA policies from gratis to libre. But we can say that the moment is arriving.
"[U]niversities [and funders] can act together without acting as a cartel if critical numbers of them become courageous about seeking their own interests at about the same time. Without critical numbers and critical timing, early requests will simply be rejected. But as soon as some large institutions or clusters of institutions start to win concessions, it will be easier for the next institutions to make the same requests and build on the momentum."
"If only one or two agencies adopted...[a] libre mandate, some publishers might refuse to publish the work they funded. But if this kind of libre mandate became the norm, publishers would have to accommodate it."
(5) Recent rise of libre OA
Finally we can turn to the good news. Here's a catalog of recent libre initiatives --where "recent" means in the past six months or so.
As in previous sections, I'm omitting open-data initiatives, but only because they are more often libre than OA initiatives for texts.
In most cases I've been brief, so that I could say a little about many new developments rather a lot about just a few. But for the first half dozen or so, I couldn't resist the temptation to elaborate.
* Research Councils UK (RCUK)
The RCUK adopted gratis OA mandates in 2006. In mid March of this year they released a draft upgrade to their policies and called for public comments until April 10.
The new draft policy requires that "a user must be able to...re-use the content of published papers both manually and using
automated tools (such as those for text and data mining) provided that any such reuse is subject to proper attribution....The existing policy will be clarified by specifically stating that Open Access includes unrestricted use of manual and automated text and data mining tools. Also, that it allows unrestricted re-use of content with proper attribution – as defined by the Creative Commons CC-BY licence."
Journals publishing RCUK-funded research must either (1) provide "immediate and unrestricted access to the publisher's final version of the paper (the Article of Record), and allow immediate deposit of the Article of Record in a repository without restriction on re-use" or (2) "allow deposit of the published paper (either the version as accepted for publication or the Article of Record) within subject-based or institutional repositories, and allows unrestricted access to those papers after an embargo period no longer than that mandated by the Research Councils." In the former case, the journal may charge a publication fee and the latter case it may not.
Mark Thorley of the Natural Environment Research Council summarized the RCUK deliberations for Nature News in March: "Either green or gold is fine with RCUK, so long as the open papers also have the CC-BY licence."
The RCUK have not published the comments they received or announced which of the proposed policy enhancements they will adopt. But at least some of the public comments supported the libre requirements. For example, my own did, as did some others building on mine.
Other comments were less supportive of the turn toward libre and CC-BY. See in particular the comments submitted by UKCoRR, Stevan Harnad, and Heather Morrison.
The seven funding agencies making up the RCUK control the bulk of publicly-funded research in the UK. Their existing gratis OA mandates already make the UK a leader in national OA policies. If they strengthened their policies to require libre, their size and influence could help elicit publisher adaptation that would help other funders and universities strengthen their policies as well.
* Wellcome Trust
Since 2007 (as noted in Section 2), the Wellcome Trust has required an open license when it paid any part of the cost of publishing an article. But the license it required was CC-BY-NC. This spring Wellcome announced plans to shift to CC-BY.
In an April article for New Statesman, Wellcome's David Carr and Robert Kiley wrote that "We will...ensure that where we pay an open access fee, the content is freely available for all types of re-use (including commercial re-use). This is in line with a recent draft policy published by the UK Research Councils, which we strongly support."
In May, GenomeWeb Daily News reported that "Kiley said, at some time in the near future, probably in 2013, the Wellcome Trust will shift to a creative commons, attribution-only license model, under which it will only make payments to publishers, such as Elsevier or the Public Library of Science, if the publisher makes the research available for use and reuse by the research community. 'We think that is the cleanest, most simple way of insuring that the fruits of our research spending on all of these research articles can be fully built upon and exploited....The best way we can maximize the use of that research is by allowing anyone and everyone to make use of that content,' and to use it for commercial purposes, Kiley explained."
(Free registration required.)
In a private email (quoted with permission), Kiley explained Wellcome's decision this way: "The current policy requires 'libre' OA, when we pay an OA fee, but we accepted that CC-BY-NC counted as libre. We are now of the view that when we pay a fee the only acceptable license will be CC-BY, which does allow commercial reuse, as well as other types of re-use. We need to communicate this change to the publishers used by Wellcome-funded authors, so I anticipate that we will formally introduce this policy change from January 2013, at hopefully the same time as RCUK also move to a CC-BY license."
* World Bank
The World Bank launched an institutional repository and adopted an OA mandate on April 10, 2012.
Here are the libre features of the new policy: when Bank research is published by the Bank itself, then copies must be disseminated through the institutional repository under CC-BY licenses. This policy applies to books as well as articles. When Bank research is published by external publishers, the preprints or working papers must be in the repository under CC-BY licenses. The final versions of the peer-reviewed manuscripts must be in the repository under CC-BY-NC-ND licenses, unless the publisher can be persuaded to allow a more liberal license.
Also see the policy FAQ and repository FAQ.
The new OA policy follows on the bank's open data policy from April 2010.
For the purposes of OA policy, the World Bank is a peculiar institution. It has something in common with universities, funders, and publishers, but it doesn't fall neatly into any of those categories. It hires researchers. It funds research, but so far only by its employees. It publishes some of its own research, but not all of its own research and not research by others. Primarily, it's a bank, and it's a peculiar even for a bank: created by treaty, controlled by the votes of 187 member nations, and funded by pledges from 45 nations. But it's a bank with a research division and a publishing division. Its financial operations serve an overriding mission to reduce poverty and foster development. Its research has the same ends, and its open-data and OA policies are instrumental to those ends.
The RCUK called for comments on a draft policy to require CC-BY, but hasn't yet adopted the policy. The Wellcome Trust announced plans to require CC-BY when it pays publication fees, but hasn't yet made the transition. The World Bank, however, has now adopted an OA mandate requiring CC-BY for a significant portion of the research it funds. The Bank correctly describes itself as "the first major international organization to require open access under copyright licensing from Creative Commons."
The Bank's peculiar structure and mission may have enabled it to be the first institution to take this significant step. But credit also goes to the bank's leadership. In a September 2010 talk, Bank President Robert Zoellick said, "This [new research model] will open the treasure chest of the World Bank's data and knowledge to every village health care worker, every researcher, everyone....Above all, we must look beyond an 'elite retail' model of research. No longer can the model solely be to research a specific issue and write a paper hoping someone will read it. The new model must be 'wholesale' and networked. It must increasingly open information and knowledge to others by giving them the tools to do the economic research themselves."
In an April 2012 interview with Richard Poynder, Bank Publisher Carlos Rossel said, "When the decisions on Open Data and Open Access were made the focus was not on lost revenues but on the cost of not opening our data or adopting an OA publishing model, for our clients and for development."
For several perspectives on the new policy, with frequent emphasis on the benefits of libre OA under open licenses, see the panel discussion at the Bank's May 21 launch event in Washington. (Disclosure: I was a panelist.)
* Willetts initiative in the UK
In May 2012, David Willetts, the UK Science Minister, supplemented the RCUK draft policy with plans for a strong and comprehensive OA policy for the UK.
The details are being hammered out by the Working Group on Expanding Access to Published Research Findings (WGEAPRF), also known as the Finch group after its chair, Dame Janet Finch.
The group will release its report later this month. In the meantime, the Finch Group has released the minutes of its first four meetings.
The minutes make clear that the Group wants to assure some kind of libre OA, even if it is also concerned about its effects on publishers.
(Thanks to Stephen Curry for pointing out the libre theme in these minutes.)
From the minutes of October 17, 2011: "The [Working Group]...confirmed, for the purpose of its work, its understanding about the scope of access: ...that it relates to re-use of published work as well as reading."
From the minutes of December 1, 2011: "There is also a concern that national licensing is not as likely as OA to address re-use."
From the minutes of February 22, 2012: "The sub-group [on author-side payments] had also agreed that author-side payments should bring with them full rights of re-use....The [plenary Working Group] also considered the definitions of and the proposed additions to the success criteria. It was agreed that...assessment of increases in access should include consideration of issues relating to any restrictions on rights of re-use."
From the minutes from April 27, 2012: "David Sweeney reported that [the still-forthcoming position statement from the Higher Education Funding Council for England (HEFCE) on the Research Excellence Framework (REF)] is founded on the Government's innovation strategy. Material submitted to REF for assessment beyond the 2014 exercise should thus generally be made freely available – although the detail, in terms of the extent to which this applies to re-use, has not yet been determined. Publishers feel that this too poses the question about how such a policy might be sustained....There is a need for clarity about what licences should cover, given the distinction between read-only and re-use rights. Here too, sustainability is an important factor. Nonetheless, the Group felt that the policy should stress that there should be as few restrictions as possible for use and re-use, including text and data mining."
When Willetts created the Finch Group last September, its charge was "to look at how UK-funded research findings can be broadened for key audiences such as researchers, policy makers and the general public."
Now it appears that the Finch Group wants re-use rights (libre OA), and will recommend them for UK-funded research. In that sense, its recommendations will complement those already under consideration by the RCUK.
There's another libre implication as well. The April 2012 minutes tell us that the Research Excellence Framework (REF), the latest incarnation of the UK research assessment exercise, "should" require some kind of OA after 2014, but that "the extent to which this applies to re-use, has not yet been determined." Hence, the OA requirement to be built into the REF might require some degree of libre OA as well. This matters because the body of UK-funded research plus REF-submitted research is larger than the body of UK-funded research alone. The REF policy would significantly extend the scope of a libre policy, even if the terms of the REF libre policy were not quite the same as the terms of the RCUK libre policy.
Willetts himself has used similar though less explicit language. "He suggested [to the Times Higher Education] that open access 'could be among the excellence criteria for qualifying articles' for REF rounds beyond 2014. A spokesman for the Higher Education Funding Council for England said that Mr Willetts was referring to a proposal being considered by the funding councils to require articles submitted for the REF to be freely available 'as far as possible'...."
Note the phrase "as far as possible". It might mean that articles submitted to the REF articles will be as open as possible, hence libre. Or it might mean that as many articles as possible will be gratis. The HEFCE report too, like the Finch Group report, will appear later this month, and then we'll know for sure.
In February 2012, the Open Knowledge Foundation launched @ccess, a new initiative focusing on libre OA.
Here's the @ccess mission, in the words of two of its co-founders.
Peter Murray-Rust: "Today we have launched @ccess – a new site, and more importantly a new community – to make scholarly information REALLY LIBRE available....By LIBRE we mean free to use, re-use, and redistribute for any purpose. It's covered by the Open Knowledge Definitions and the actual text of the Budapest Declaration on Open Access 10 years ago."
Tom Olijhoek: "Because open access can range from somewhat restricted (only free reading) to completely unrestricted (completely free for use and reuse) we have proposed to coin the term @ccess for free and unrestricted access to information in accordance with the BBB definition...."
In addition to libre OA, @ccess is committed to networking researchers and interlinking pieces of knowledge. All three goals are present in its first major project, the malaria initiative. @ccess and "MalariaWorld...are developing a comprehensive database of malaria related publications. At the same time we will ask researchers to deposit their manuscripts and data in an open access repository that is linked to the database. This database will also link to open access articles. For restricted access publications we will seek to get as many manuscripts as possible deposited in the database as well. The community will eventually provide open access to all information, provide a platform for collaboration and information exchange and serve as a communication platform for everyone seeking information on, or working on malaria. Other communities can be formed using this model. In this way we would move towards a system of interlinked scientific communities and easy access to pertinent information through these communities...."
* Text mining
In March 2012, JISC released a major report on the value and benefits of text mining.
The report not only points out that priced access blocks text mining, but that permission barriers on gratis works can also block mining. "Current UK copyright restrictions...mean that most text mining in UKFHE is based on Open Access documents or bespoke arrangements. This means that the availability of material for text mining is limited....As several consultees highlighted, ...most text mining is limited to exploring Open Access documents where no additional charges are incurred....Even where text mining is allowed within publisher contracts, licensing terms that require the full attribution of derivative works developed in the text mining process can effectively prevent text mining usage. For example, the Open Access publisher BioMed has such a licence, allowing text mining and the production of derivative works, provided the relevant attribution is made. However, where text mining is used to identify new knowledge derived from cross-article analysis of patterns, it is effectively impossible to identify all relevant attributions that contributed to the new derived knowledge. This therefore means that such text mining cannot be undertaken...."
In April, Heather Piwowar asked Elsevier to allow text mining, and after some negotiation the company agreed.
Piwowar's victory was a breakthrough for text mining, and for uses that not even gratis OA would allow. But her one-off negotiation triggered a larger discussion of how to reach the same result for all researchers and all publishers with out an endless series of one-off negotiations. The buzz quickly moved beyond congratulations to a revolution of rising expectations. For some of the details, see the (OA) SPARC interview with Piwowar on this project, the (TA) coverage in the Chronicle of Higher Education, and the (OA) coverage in The Guardian (with a large trail of reader comments).
A few days after her Elsevier victory, Piwowar asked on her blog whether a text-mining manifesto might be a good idea, and jotted some notes toward a future manifesto. The next month, Peter Murray-Rust posted some additional notes toward a manifesto.
A drafting group is now writing a manifesto. The group isn't ready to release a draft, but it will clearly recommend libre OA as one of the prerequisites. (Disclosure: I'm a time-starved, under-contributing member of the group.)
Although I want to stick to recent developments in this section, any coverage of text-mining breakthroughs should note that the Hargreaves report in May 2011 recommended a UK copyright exception to allow mining, and in August 2011 the UK government committed itself to implement that recommendation.
For the literature on how OA facilitates text mining, from Steve Dickman in 2003 ("Access is a bigger problem than algorithms") to Ross Mounce last week ("I have 20,000+ PLoS articles on my computer right now")...
...see my past blog posts on the subject and the "oa.mining" tag library from the OA Tracking Project.
* Digitization projects
In March 2012, the US National Gallery of Art launched NGA Images, a new collection of downloadable, high-res digital reproductions of public-domain artworks. At the same time the gallery announced an OA policy for the new collection: "With the launch of NGA Images, the National Gallery of Art implements an open access policy for digital images of works of art that the Gallery believes to be in the public domain. Images of these works are now available free of charge for any use, commercial or non-commercial. Users do not need to contact the Gallery for authorization to use these images....The Gallery believes that increased access to high quality images of its works of art fuels knowledge, scholarship, and innovation, inspiring uses that continually transform the way we see and understand the world of art. Works in the public domain are those not subject to copyright protection. The Gallery has launched this open access policy with works it believes to be in the public domain, but is hoping gradually to include additional works whose public domain status is currently uncertain...."
Because the OA policy is new and consciously publicized, my guess is that the gallery will stand by the new policy and bring the older TOS page into conformity with it. But I don't know that. I've alerted the gallery to the contradiction and asked how it plans to resolve it. If I get a reply, I'll blog it.
(Aside: It's remarkable how rare it was until recently to hear a major museum or digitization project explicitly and emphatically announce the principle that digital reproductions of public-domain works are themselves in the public domain. I credit Cornell University with a public statement in May 2009 that raised consciousness about this principle and precipitated a series of other notable statements.
For a fuller discussion of the principle, see my article in the July 2009 SOAN, reprinted with some revisions by Nobel Foundation in April 2011.
For some major post-Cornell affirmations of the principle, including statements from the EU Culture and Education Committee, the Europeana Public Domain Charter, the Australian Collecting Institutions, the EU Reflection Group or Comité des Sages, and Yale University, see my article in the June 2011 SOAN.)
* OA journals
In January 2012, Springer's Open Choice journals converted from CC-BY-NC to CC-BY. The company's hybrid OA journals now use the same license as its full OA journals, including the full OA journals from its subsidiary, BioMed Central.
This is good news, but unfortunately it's the only recent news I have under this heading. Where are the other OA journals moving from more restrictive licenses to less restrictive licenses like CC-BY? For that matter, where are the OA journals moving from gratis access and all-rights-reserved copyrights to libre access under open licenses?
It's possible license upgrades like Springer's happen more often than they are formally announced. But I'm afraid they simply don't happen very often.
In December 2012 State Senator Tom Steinberg proposed two bills to allocate public funds to create OA college textbooks under CC licenses. In May 2012 one bill passed the Senate and moved on to the Assembly. The other bill is still pending.
In February 2012, MIT OpenCourseWare and Flat World Knowledge joined forces to create OA textbooks under CC licenses.
Also in February, the Council on Library and Information Resources (CLIR) and National Institute for Technology in Liberal Education (NITLE) launched Anvil Academic, a new publishing platform for enhanced OA books under open licenses.
In May 2012, Eric Hellman launched Unglue.it, a crowdfunding service to raise money to convert already-published books to OA and distribute them under open licenses.
In March 2012, Ross Mounce released the first draft of a crowd-sourced survey of open licenses (if any) used by various journal publishers. "This spreadsheet is an attempt to survey scholarly communications (primarily research journals) to find out which publishers are producing 'open access' articles/journals, and what each of these publishers *mean* by 'open access' with reference (preferably) to an explicit license statement, or some evidence thereof....Analysis of the data collected so far shows that less than 5% of publishers claiming 'open access' actually provide explicitly BOAI-compliant Open Access."
Just last week, Wikipedia launched an article on "Libre" (on libe in general, not just libre software or libre access). It has a subsection on libre licenses, and another subsection on how the term is used in the OA movement.
* Obama administration policy
In November 2011 the White House Office for Science and Technology Policy called for public comments on federal OA policy. In January 2012 it released the comments it received.
It hasn't released a detailed summary. But it released a brief summary --silent on libre issues-- in April (dated March). There's ground for optimism: "Responses to those RFIs are being analyzed now, [and] initial results show strong public support for increasing access to scholarly publications describing the results of federally funded research and for improving scientific data management and access."
In the call for comments, the White House posed a specific set of questions about OA for publications and OA for data. One question about OA publications asked, "What type of access to these publications is required to maximize U.S. economic growth and improve the productivity of the American scientific enterprise?" Many respondents took this as an invitation to discuss gratis v. libre access. As a result, the White House received many comments urging it not only to extend the NIH policy across the federal government, but to strengthen it from gratis to libre. Here are seven examples:
From the American Library Association and Association of College and Research Libraries: "The complete collection of articles resulting from publicly funded research should be made immediately freely accessible, so that the public can fully use them – (i.e.
text mine, data mine, compute on them, create derivative works) without commercial restriction....The NIH public access policy for federally research has proven to be a good model. However, further mandates should reduce the allowable embargo period for public access to research articles and take steps to enhance the ability of users to reuse the material they find in the repositories"
From the Association of Research Libraries: "[T]wo reports described below provide clear evidence that openly available resources with no reuse restrictions promoted economic growth and created new jobs and markets....Extending public access policies that permit full use and reuse rights with no cost barriers will significantly enhance STEM education, level the playing field, and generate more economic growth and job creation in diverse new areas....In order to maximize the investments in cyber and information infrastructure, advance science, and promote innovation, free immediate access with full reuse rights to federally funded research literature would achieve the most benefits."
From the Berkeley Digital Library Copyright Project: "Most importantly, extensibility requires that licenses to the covered works be tied not to the particular entity hosting the works, but rather to the ultimate use of the work, which is available to all hosting entities (and users). License terms should also be permissive in the uses they allow, being careful to expand and not contract the uses to which libraries and end users may put the covered works. Existing exceptions to copyright, such as fair use or the library exceptions codified in Section 108 of the Copyright Act should not be limited by license terms. Contractual limitations on these exceptions have frustrated library efforts to increase online access for their patrons in the past, and such limitations could similarly bind federally funded research in a way that prevents useful access and that curtails future productive uses of these works. These are precisely the type of activities that existing copyright exceptions (and in particular, fair use) are designed to foster."
From Creative Commons: "Scholarly articles created as a result of federally funded research should be released under full open access. Full open access policies will provide to the public immediate, free-of-cost online availability to federally funded research without restriction except that attribution be given to the source....[T]he NIH Public Access Policy does not go far enough in communicating the rights for reuse that should be available to the public that paid for the development and publication of that research. Releasing the outputs of federally funded research free of cost online should be a baseline, but if downstream users (including researchers within the same domain, scholars from other disciplines, creative entrepreneurs, government employees, citizens) are unclear about their legal right to copy, amend and redistribute the federally funded research, those publications will be reused less. This will significantly diminish the potential impact of the researchand, by extension, the public's investment."
From Harvard University: "One type of public access merely provides research results online free of charge. A second type provides research results free of charge and free of certain copyright restrictions. Only the second type frees research for data- and text-mining, translation, conversion to new formats, integration with other tools and bodies of research, and other value-added services. Hence the second type does far more than the first to amplify the benefits of publicly funded research. Limiting the reuse of publicly funded research limits the return on our investment. If we are serious about maximizing that return on investment, we must lift restrictions on use and reuse, not just restrictions on access for reading. The public access policy at the NIH does just the former....In August , Phil Malone and Harvard’s Berkman Center for Internet & Society released a report evaluating the copyright licensing policies used by certain public and private funding agencies. The report recommended that research funders require the use of open licenses for funded research. It articulates nine benefits of open licenses for researchers and the funders themselves....We recommend the Creative Commons Attribution (CC-BY) license....We recommend against licenses that bar commercial use (such as CC-BY-NC), in part because they would limit the utility of publicly funded research for businesses and industry."
From MIT: "In order to maximize the growth of existing markets and develop new markets associated with peerreviewed publications from federally funded research, the entire corpus of articles that have derived from publicly funded research should be made available for public use and reuse. This means not only being able to read the articles, but being able to create new works from them, and being able to use mechanized tools to analyze them or derive data from them. While making the articles openly available for reading would speed science, making the articles available for reuse would also enable the development of new services and products....Public access policies that include open reuse rights (not simply ‘read-only’ access) allow the mining of information and encourage the creation of new tools and the use of new tools, thus providing access to key scientific knowledge more quickly, and offering faster application of that knowledge. When there is full open access, including reuse rights, machines can become readers, fostering new layers of connection and innovation that are not possible through human-based processing."
From the Oberlin Group of Libraries: "Access should be free of charge and should include a broad range of re-use rights so that users can build on and innovate from the research that they find. Providing open access to federally funded scientific and scholarly research reports – including full rights to re-use or mine these reports – allows more users to stay abreast of cutting-edge ideas, access these ideas quickly, and generate new uses and applications from this research, speeding the launch of new services and products into the marketplace and energizing the economy....We urge full open access as the norm: free, immediate access with full rights of re-use. Restrictions on access or on use simply reduce the return on taxpayer investment – whether that return is in the education of new researchers or the entry of new products and services into the marketplace."
* Postscript 1.
A few notable recent OA initiatives are silent on libre. I don't want to list them all, but I do want to list a few.
The Federal Research Public Access Act (FRPAA) is silent on libre. It would mandate gratis OA for most publicly-funded research in the US, but only gratis OA.
Likewise Europe's Horizon 2020 program is silent on libre. It will mandate gratis OA for most EU-funded research, but only gratis OA.
These major OA mandates are silent on libre for the same reason that other funder mandates are silent on libre. They're about access to peer-reviewed articles based on research arising from certain funders, not about access to articles for which the funder has paid a publication fee. So far no funding agency anywhere requires libre OA to the research they fund unless they also pay the costs of publication. I hope that changes. But we're not there yet.
FRPAA is already drafted and final; the only question is whether supporters can muster enough votes. But Horizon 2020 will not be final until the European Commission votes on its final form in November 2012. What will happen if all seven of the Research Councils UK firmly commit to libre OA for RCUK-funded research in time for the EC to take note, absorb the arguments, and consider amendments? No one can say, today, and the answer may be up to us.
The CostOfKnowledge petition and Elsevier boycott are silent on libre. This includes the short statement on the petition site as well the longer statement from its founders.
I believe the petition is silent on libre for the same reason that many earlier initiatives are silent on libre. When we don't even have gratis OA, it's prudent to start with gratis.
Finally, the White House OA petition is silent on OA. This was deliberate, just as the petition is deliberately silent on the length of embargo periods. As John Wilbanks explains, "The petition is simple because of two reasons. One, you only get 800 characters to work with. That's not something conducive to nuance. Second, it's simple because we want a positive response from the Administration, and by staying simple we allow a little bit of flexibility to them as they respond."
(Aside: If you haven't already signed the petition, sign it now! If you haven't already spread the word about it, spread it now!)
* Postscript 2. More on terminology.
One recent development is the rise in the number of people calling for the term "OA" to mean libre and nothing but libre. In a 2004, I took this position myself.
But by 2008, I had to concede that the term "OA" had drifted to cover both gratis and libre access. (In fact, I was already complaining about drift in the 2004 article.) Moreover, I had to concede the separate point that changing the actual usage of speakers and writers was an unwinnable fight.
Individual writers might use "OA" as if it meant libre and libre only, but others will continue to use it for gratis access as well. This is not a normative question about what a term ought to mean, but a descriptive question about how a term is actually used.
The question is what to do about that drift. In the same 2008 article I argued that the drift didn't hinder us from speaking precisely, recommending the policies we actually want, arguing that libre is superior to gratis, or arguing that one particular stripe of libre is superior to another. I still believe that as well. The drift is completely harmless if we take the trouble to say what we mean. The gratis/libre distinction, and a range of named licenses, allow us to do that.
I sympathize with the argument that the term "OA" should be used as defined in the BBB documents, which all demand libre. I especially sympathize with the argument that the term should be used in the Budapest sense, since the Budapest statement first introduced the term and defined it. I was the chief draft of the Budapest statement, and I haven't retreated an inch in my support for it. But none of this makes the dift in usage go away. To speak about libre OA today, and make ourselves clear to a mixed audience, we need an adjective like "libre" or "Budapest" in front of "OA". I don't mind that. As I said in 2008, "That's more than winnable. It's easy."
Apart from the terms we prefer to use, we must recognize the distinction between works that are gratis and works that are more free than gratis. Moreover, we must recognize the distinctions within the spectrum of licenses that make work more free than gratis. For example, if we limited "OA" to libre access, we'd still need a generic term for works that are at least free of charge (gratis or libre or both). Right now, like it or not, "OA" plays that role. If we limited "libre" to works at the upper or most-free end of that spectrum, we'd still need a term for the wider range of works that are "more free than gratis".
Since 2008 I've used "gratis" to mean free of charge (like the FLOSS community), and "libre" to mean "more free than gratis". But I acknowledge that the FLOSS community uses "libre" for the narrower range at most-free end of the scale. This isn't a disagreement about licenses, policy, or freedom; it's a divergence of diction. In my 2008 article, I emphasized the ways in which the meanings of "gratis" and "libre" were alike in FLOSS and OA. But since then the differences present even then have become more evident to some activists in both domains. If we can find a term that we all agree means "more free than gratis", and another pair of terms for the less-free and more-free ends of that spectrum, I think we'd all be happy. I know I would. As I said in my 2008 article, my decision to use "libre" in the wider sense ("more free than gratis") is "provisional in the sense that I'll continue to look for better terms."
Meantime, I don't think "libre" is a bad choice for "more free than gratis". If you asked 100 randomly picked members of the FLOSS community for a term that meant "more free than gratis", without priming the pump, I'd bet that most would offer "libre". Nevertheless, whether we think the diction similarities between the communities outweigh the differences, or vice versa, the differences remain, and it would be desirable to iron them out. We just have to beware that a person or a committee can't change actual usage just by wishing to change it.
The good news is that we have a useful recourse no matter what happens to usage. As I wrote in 2008, "What's the best way to refer to a specific type of [more-free-than-gratis] OA? With a license. We'll never have unambiguous, widely-understood technical terms for every useful variation on the theme. But we're very likely to have clear, named licenses for every useful variation on the theme, and we're already close. Licenses are more precise than single terms and not nearly as susceptible to misunderstanding or divergent usage."