Category Archives: Scholarly Publishing

From control to contempt

I hope it was clear, when I wrote about the press release from Elsevier addressing their new approach to authors’ rights and self-archiving, that I believe the fundamental issue is control.  In a comment to my original post, Mark Seeley, who is Elsevier’s General Counsel, objected to the language I used about control.  Nevertheless, the point he made, about how publishers want people to access “their content,” but in a way that “ensures that their business has continuity” actually re-enforced that the language I used was right on the mark.

My colleague Paolo Mangiafico has suggested that what these new policies are really about is capturing the ecosystem for scholarly sharing under Elsevier’s control.  As Paolo points out, these new policies, which impose long embargo periods on do-it-yourself sharing by authors but offer limited opportunities to share articles when a link or API provided by Elsevier is used, should be seen alongside the company’s purchase of Mendeley; both provide Elsevier an opportunity to capture data about how works are used and re-used, and both  reflect an effort to grab the reins over scholarly sharing to ensure that it is more difficult to share outside of Elsevier’s walled garden than it is inside that enclosure.

I deliberately quote Mr. Seeley’s phrase about “their content” because it is characteristic of how publishers seem to think about what they publish.  I believe it may even be a nearly unconscious gesture of denial of the evident fact that academic publishers rely on others — faculty authors, editors and reviewers — to do most of the work, while the publisher collects all of the profit and fights the authors for subsequent control of the works those authors have created. That denial must be resisted, however, because it is in that gesture that the desire for control becomes outright disrespect for the authors that publishing is supposed to serve.

Nowhere is this disrespect more evident than in publisher claims that the works they publish are “work made for hire,” which means, in legal terms, that the publisher IS the author.  The faculty member who puts pen to paper is completely erased from the transaction.  To be clear, as far as I know Elsevier is not making such a claim with its new policies.  But these work made for hire assertions are growing in academic publishing.

Three years ago I wrote about an author agreement from Oxford University Press that claimed work made for hire over book chapters; that agreement is still in use as far as I am aware.  At the time, I pointed out two reasons why I thought OUP might want to make that claim.  First, if something is a work made for hire, the provision in U.S. copyright law that allows an author or her heirs to terminate any license or transfer after 35 years simply does not apply.  More significantly, an open access license, such as is created by many university policies, probably is not effective if the work is considered made for hire.  This should be pretty obvious, since our law employs the legal fiction that says the employer, not the actual writer, is the author from the very moment of creation in work made for hire situations.  So we should read these claims, when we find them in author agreements, as pretty direct assaults on an author’s ability to comply with an open access policy, no matter how much she may want to.

As disturbing as the Oxford agreement is, however, it should be said that it makes some legal sense.  When a work is created by an independent contractor (and it is not clear to me if an academic author should be defined that way), there are only selected types of works that can even be considered work made for hire; one of them is “contribution[s] to a collective work.”  So a chapter in an edited book is at least plausible as a work made for hire, although the other requirement — an explicit agreement, which some courts have said must predate the creation of the work — may still not be met.  In any case, the situation is much worse with the publication agreement from the American Society of Mechanical Engineers (ASME), which was recently brought to my attention.

ASME takes as its motto the phrase “Setting the Standard,” and with this publication agreement they may well set the standard for contemptuous maltreatment of their authors, many of whom are undoubtedly also members of the society.  A couple of points should be noted here.  First, the contract does claim that the works in question were prepared as work made for hire.  It attempts to “back date” this claim by beginning with an “acknowledgement” that the paper was “specially ordered and commissioned as a work made for hire and, accordingly, ASME is the author of the Paper.”  This acknowledgement is almost certainly untrue in many, if not most, cases, especially since it appears to apply even to conference presentations, which are most certainly not “specially commissioned.”  The legal fiction behind work made for hire has been pushed into the realm of pure fantasy here.

What’s more, later in the agreement the “author” agrees to waive all moral rights, which means that they surrender the right to be attributed as the author of the paper and to protect its integrity.  Basically, an author who is foolish enough to sign this agreement has no relationship at all to the work, once the agreement is in place.  They are given back a very limited set of permissions to use the work internally within their organization and to create some, but not all, forms of derivative works from it (they cannot produce or allow a translation, for example).  Apparently ASME has recently started to disallow some students who publish with them to use the published paper as part of a dissertation, since most dissertations are now online and ASME does not permit the  writer to deposit the article, even in such revised form, in an open repository.

To me, this agreement is the epitome of disrespect for scholarly authors.  Your job, authors are told, is not to spread knowledge, not to teach, not to be part of a wider scholarly conversation.  It is to produce content for us, which we will own and you will have nothing to say about.  You are, as nearly as possible, just “chopped liver.”  It is mind-boggling to me that any self-respecting author would sign this blatant slap in their own face, and that a member-based organization could get away with demanding it.  The best explanation I can think of is that most people do not read the agreements they sign.  But authors — they are authors, darn it, in spite of the work for hire fiction — deserve more respect from publishers who rely on them for content (free content, in fact; the ASME agreement is explicit that writers are paid nothing and are responsible for their own expenses related to the paper).  Indeed, authors should have more respect for themselves, and for the traditions of academic freedom, than to agree to this outlandish publication contract.

Stepping back from sharing

The announcement from Elsevier about its new policies regarding author rights was a masterpiece of doublespeak, proclaiming that the company was “unleashing the power of sharing” while in fact tying up sharing in as many leashes as they could.  This is a retreat from open access, and it needs to be called out for what it is.

For context, since 2004 Elsevier has allowed authors to self-archive the final accepted manuscripts of their articles in an institutional repository without delay.  In 2012 they added a foolish and forgettable attempt to punish institutions that adopted an open access policy by purporting to revoke self-archiving rights from authors at such institutions.  This was a vain effort to undermine OA policies; clearly Elsevier was hoping that their sanctions would discourage adoption.  This did not prove to be the case.  Faculty authors continued to vote for green open access as the default policy for scholarship.  In just a week at the end of last month the University of North Carolina, Chapel Hill, Penn State, and Dartmouth all adopted such policies.

Attempting to catch up to reality, Elsevier announced last week that it was doing away with its punitive restriction that applied only to authors whose institutions had the temerity to support open access. They now call that policy “complex” — it was really just ambiguous and unenforceable — and assert that they are “simplifying” matters for Elsevier authors.  In reality they are simply punishing any authors who are foolish enough to publish under these terms.

Two major features of this retreat from openness need to be highlighted.  First, it imposes an embargo of at least one year on all self-archiving of final authors’ manuscripts, and those embargoes can be as long as four years.  Second, when the time finally does roll around when an author can make her own work available through an institutional repository, Elsevier now dictates how that access is to be controlled, mandating the most restrictive form of Creative Commons license, the CC-BY-NC-ND license for all green open access.

These embargoes are the principal feature of this new policy, and they are both complicated and draconian.  Far from making life simpler for authors, they now must navigate through several web pages to finally find the list of different embargo periods.  The list itself is 50 pages long, since each journal has its own embargo, but an effort to greatly extend the default expectation is obvious.  Many U.S. and European journals have embargoes of 24, 36 and even 48 months.  There are lots of 12 month embargoes, and one suspects that that delay is imposed because those journals that are deposited in PubMed Central, for which 12 months is the maximum embargo permitted.  Now that maximum embargo is also being imposed on individual authors.  For many others an even longer embargo, which is entirely unsupported by any evidence that it is needed to maintain journal viability, is now the rule.  And there is a handful of journals, all from Latin America, Africa, and the Middle East, as far as I can see, where no embargo is imposed; I wonder if that is the result of country-specific rules or simply a cynical calculation of the actual frequency of self-archiving from those journals.

The other effort to micromanage self-archiving in this new policy is the requirement that all authors who persevere and wish, after the embargo period, to deposit their final manuscript in a repository, must apply a non-commercial and no derivative works limitation on the license for each article.  This, of course, further limits the usefulness of these articles for real sharing and scholarly advancement.  It is one more way in which the new policy is exactly a reverse of what Elsevier calls it; it is a retreat from sharing and an effort to hamstring the movement toward more open scholarship.

The rapid growth of open access policies at U.S. institutions and around the world suggests that more and more scholarly authors want to make their work as accessible as possible.  Elsevier is pushing hard in the opposite direction, trying to delay and restrict scholarly sharing as much as they can.  It seems clear that they are hoping to control the terms of such sharing, in order to both restrict it putative impact on their business model and ultimately to turn it to their profit, if possible.  This latter goal may be a bigger threat to open access than the details of embargoes and licenses are. In any case, it is time, I believe, to look again at the boycott of Elsevier that was undertaken by many scholarly authors a few years ago; with this new salvo fired against the values of open scholarship, it is even more impossible to imagine a responsible author deciding to publish with Elsevier.

Listening to Lessig

Like many other attendees, I was pleased when I saw that the closing keynote address for this year’s Association of College and Research Libraries Conference was to be given by Professor Larry Lessig of Harvard.  But, to be honest, my excitement was mingled with a certain cynicism.  I have heard Lessig speak before, and I am afraid I worried that I would be listening to essentially the same lecture again.

My suspicion was not wholly unwarranted.  In part I think it is the fault of Lessig’s instantly recognizable lecture style.  It is energetic and entertaining, but because its rhythms and conventions are so idiosyncratic, I think it may flatten the message a little bit.

In any case, I sat down in the ballroom of the Oregon Convention Center on Saturday with somewhat mixed expectations.  But what I did not expect was for Lessig to begin his talk by acknowledging that all his public lectures were really the same.  Had he read my mind?  No, his point was a little different.  Over the years, he told us, he has had three major themes – political corruption, net neutrality, and copyright/open access.  But, he told his audience of attentive librarians, those three themes are fundamentally just one theme.  Each is about equality.  Not three themes, but only one — equality.  Equality of access to the political process is the heart of his current campaign against the corruption of our political system by the endless pursuit of money.  Equality of access to the means of communication and culture is key to the fight for net neutrality.  And equality of access to knowledge is what animates the open access movement.

So it turns out that my worry, prior to the talk, was both unfair and, in a sense, correct.  All Lessig’s lectures are very much the same, because the underlying value he is asking us to focus on is the same.

Thinking about this unity-behind-diversity in the messages about political corruption, net neutrality and open access set me thinking about the way my colleagues and I frame our advocacy for the last of those items, open access to scholarship.  Our messages, I think, tend to focus on incremental change, on the benefits to individual scholars, and on not rocking the academic boat too much.  Lessig reminded me that there are good reasons to rock a little bit harder.  Publishing in toll access journals and neglecting open access options or additional means of dissemination is not just short-sighted.  It is dumb, and it is harmful.  We need to say that occasionally.

Publishing exclusively through closed access channels is dumb because it ignores huge opportunities available that can, quite simply, make the world a better place.  And such publishing fails to take full advantage of the greatest communications revolution since the printing press.  Indeed, online toll-access deliberately breaks network technology in order to protect its outmoded and exclusionary business model.  Doing this is simply propping up the buggy whip manufactures because we are afraid of how fast the automobile might travel.  The academy is not usually this dumb, but in this case we are wasting vast amounts of money to support an obsolete technology.  I know that the promotion and tenure process is often cited as the reason for clinging to the old model, but this is simply using one outdated and inefficient system as an excuse for adhering to another such system.  Traditional modes of evaluation are breaking down as fast as traditional publishing and for the same reasons.  Hiding our heads in the sand is no solution.

More to the point, however – more to Lessig’s point – is the fact that this traditional system we are so reluctant to wean ourselves from actually hurts people.  It fosters ignorance and inequality.  It makes education more difficult for many, retards economic progress, and slows development worldwide.  As academics and librarians who by inclination and by professional responsibility should be committed to the most universal education possible, it is shameful that we cling to a system where only the rich can read our scholarship, only the privileged gain access to the raw materials of self-enlightenment.  How can a researcher studying the causes and treatments of malaria, for example, be satisfied to publish in a way that ensures that many who treat that disease around the globe will never be able to see her research?  How can an anthropologist accept a mode of publishing that limits access for the very populations he studies, so they will never be able to know about or benefit from his observations?  Why would a literary scholar writing about post-colonialist literature publish in a way that fosters the same inequalities as earlier forms of colonialism did?

In this wonderful column from Insider Higher Ed., the ever-insightful Barbara Fister writes about what we really mean when we talk about serving a community, and what we might mean by it.  She comments on the “members-only” approach to knowledge sharing that has become an accepted practice, and challenges us to rethink it.  Like Lessig, Fister is calling us to consider our core values of equality and the democratization of knowledge.  She also reminds us of how dumb – her word is wasteful – the current system is.

Perhaps the most vivid example of how subscription-based publishing fosters, and even demands, inequality is found in the ongoing lawsuit brought against a course pack publisher in India by three academic publishers.  Two of the “usual suspects” are here – Oxford University Press and Cambridge University Press (joined, in the Delhi University suit, by Taylor and Francis) – and this lawsuit is even more shameful than the one brought against Georgia State.  The problem, of course, is that the books published by these modern-day colonialists are too expensive for use in India.  I once was told by a professor of IP law in India that a single textbook on trademark law cost over a month’s salary for his students.  Photocopying, whether it is authorized by Indian law or not (and that is the point at issue) is a matter of educational survival, but these publishers want to stop it.  Their rule – no one should ever learn anything without paying us – is a recipe for continued ignorance and inequality.  It is disgraceful.

I use the word colonialists in the paragraph above quite deliberately.  What we are seeing here is the exploitation of a monopoly that is imposed on a culture with the demand that people pay the developed world monopoly holders in order to make progress as a society.  We have seen this too many times before.

The thing I like best in the article linked above – the whole thing is well worth a careful read — is the brief story of how a student group in India began handing out leaflets about the lawsuit at a book fair where CUP representatives were hawking their wares.  They wanted to let people know that buying books from Oxford and Cambridge is supporting a worldwide campaign of intimidation that is aimed at reducing access to knowledge and culture.  Publishing with these presses is a form of colonial occupation that extorts from whole populations a high price to obtain the means of cultural and intellectual growth.  The reaction, of course, was predictable; the publisher summoned the police to protect themselves and others from these unpleasant truths.  But the technique has merit; perhaps we can also find ways to shame these publishers when they attend our academic or professional conference, when they send salespeople to our campuses, and when they recruit our colleagues to write and review for them.  A commitment to equality demands no less.

Copyright, Open Access, and Human Rights

The United Nations Human Rights Council is holding its 28th session this month, and one item on the agenda is discussion about a report from Farida Shaheed, who is a “Special Rapporteur” in the area of “cultural rights.”  Ms. Shaheed is a well-known Pakistani sociologist and human rights activist.  Her report is a remarkable document in many ways, with a lot of things to like for those who are concerned about the overreach of copyright laws.  There are also some points that are troubling, although, on balance, I would love to see this report get attention and action from the U.N.

In some sense, the most remarkable thing about this report is its frank recognition that intellectual property laws are in tension with the fundamental human right of access to science and culture. In only its third paragraph, the report reminds us that since at least 2005, the World Intellectual Property Organization, a U.N. agency, has been mandated (whether effectively or not) to give “renewed attention to alternative policy approaches to promote innovation and creativity without the social costs of privatization.”  In short, the WIPO is charged, whether effectively or not, to find ways to facilitate open access to science and culture.  This charge is made explicit in the recommendations, where the Special Rapporteur directly suggests that “[p]ublic and private universities and public research agencies should adopt policies to promote open access to published research, materials, and data on an open and equitable basis, especially through the adoption of Creative Commons licenses” (para. 113).

When librarians and other open access advocates discuss OA policies with their faculties, perhaps we should recognize that there is a compelling argument to be made that this is not just a “what’s best for academia and for my interests” issue, but a true human rights issue.  Ms. Shaheed’s report makes this case in a concise and compelling way.  And this point also reminds us of why open access that is achieved simply by paying the commercial publishers to release articles is not a solution, because it does not really promote equitable access.  The fees charged are too high for many authors, they are not administered in a transparent way, and, frankly, some of the publishers cannot be trusted to fulfill their end of the bargain.  Barbara Fister discussed some of these problems in more detail in her recent Inside Higher Ed blog post, called “New Predatory Publishing in Old Bottles.”  If we take open access seriously as a step toward a more democratic and equitable culture, we must embrace a wider variety of “flavors” of OA, and not assume that the “usual suspects” can do it for us.

To return to a reading of the Human Rights Council report, there are strong endorsements of the idea that cultural and scientific development depends on restraining the reach of IP protections.  The section on “Copyright policy and cultural participation” is structured around three themes that all begin with “promoting cultural participation through…” and then go on to discuss copyright limitations and exceptions, international cooperation, and open licensing.  Here are some specific recommendations that I found very encouraging:

  • In regard to negotiations that are already underway, the report endorses ratification of the Marrakesh Treaty to Facilitate Access to Published Works for Persons Who are Blind, Visually Disabled, or Otherwise Print Disabled.  On the other hand, Ms. Shaheed expresses concern (para 19) about the Trans-Pacific Partnership Agreement that is currently being negotiated in secret, which is the problem that the report focuses on, calling for all such multinational agreements to be discussed in a transparent way (para 92).  Since the TPP is often defended with the claim that it will benefit developing countries, it is fascinating to see it cited as an example of a “democratic deficit in international policy making.”  From a different article, this editorial about the TPP by a U.S. Senator raises the same type of concern, and together they make a strong case against the agreement.
  • In paragraph 22 the report discusses the potential that a pervasive licensing environment can inhibit artistic self-expression and slow cultural development.
  • The report calls for the adoption of international treaties on copyright limitations and exceptions for libraries and education.  Given the current climate in the WIPO, this seems like a long shot, but is valuable in part because it calls attention to that climate, which is dominated by representatives of commercial interests (including, unfortunately, the U.S. Trade Representative).
  • On the issue of copyright limitations and exceptions, the report specifically points to fair use as a tool for allowing a more “comprehensive and adaptable” approach to unlicensed uses (para 73).  The report notes that most countries take the route of adopting exceptions for specific types of use, which provides more certainty, but adds that that approach may be inadequate in the current environment.  In general the report calls for a flexible approach to “uncompensated use of copyrighted works , in particular in contexts of income disparity, non-profit efforts, or under-capitalized artists” (para 106).  It specifically asserts, in this regard, that member states should not take a rigid approach to the so-called “three-step test” for copyright exceptions that is found in the Berne Convention (para 104), which is often used as a weapon by commercial interests against broadly applicable exceptions.
  • One of the recommendations I like best in this report, found in paragraph 107, is that states should enact laws that would prevent copyright limitations and exceptions from being overruled by contractual provisions, and protect such exceptions from excessive technological interference as well.  The UK has recently adopted the contractual part of this idea, stating that certain uses that are allowed by the law cannot be prohibited by contracts.  As I have said before, this is an idea we need to incorporate into U.S. copyright law, and it is good to see the U.N. special rapporteur endorse it so firmly.

There is an overall emphasis in the report that focuses on copyright as an authors’ right, and it is this focus that gives me some ambiguous feelings about the document.  On the one hand, I agree that a focus on authors and supporting authorship will help re-balance our approach to copyright.  Where we have most often gone wrong in this area is when we have allowed copyright discussions to focus on supporting the business models of intermediary organizations, regardless of whether or not those models really helped incentivize authors and creators.  Throughout its history, copyright has been called an authors’ right and treated like a publishers’ privilege.  Re-focusing on authors is part of restoring copyright to its proper function, and makes sense in a document about human rights.  And yet, it is also true that too much stress on author rights can also become unbalanced.  Copyright cannot benefit society unless it weighs the rights of both users and creators, especially since the former often aspire to become the latter.  Authorial control is an important part of the creative incentive, but it can easily go too far.

One troublesome area where this is a real danger is protection for “traditional knowledge and traditional cultural expressions” — the cultural creations of indigenous peoples.  This is an area where there have certainly been abuses, and it is not surprising to find a concern over TK and TCE, as the international community abbreviates them, in a report about IP and human rights.  Unfortunately, protection for these cultural products raises as many problems as it does solutions.  Should there be a public domain for TCEs?  If not, why not?  And who is the legitimate rights holder in traditional knowledge?  The national government?  In an article back in 2011, David Hansen explored some of these issues and found real incompatibility between traditional knowledge protections and the values that animate IP law.

So my final attitude toward this report is mixed, but still strongly positive.  I think it recommends many of the right steps toward restoring copyright and other IP rights to their proper scope and function.  It rightly places the focus on authors and on economic and cultural development.  It reminds us that all high-level copyright conversations should have a human rights perspective.  Where I have concerns, I see a chance for continued conversation.  But at least those conversations would take place with the proper grounding, if the report is taken as seriously as it deserves.

Resistance is Futile

This is a guest post by Jeff Kosokoff, the Head of Collection Strategy & Development for the Duke University Libraries.

In an outstanding example of Buzzword Bingo, EBSCO’s Friday press release announcing their acquisition of YBP from Baker & Taylor (B&T) says that they are assembling the tools “to truly streamline and improve administrative (‘back end’) services in ways that optimize the impact these services have on the end user experience” (EBSCO PR:  https://www.ebsco.com/news-center/press-releases/ebsco-shows-major-commitment ).

If B&T was trying to move YBP, then there were likely multiple bidders, perhaps ProQuest, perhaps one or more of the major publishing concerns. After all, YBP is one of a small handful of comprehensive book jobbers still standing and they have well-established relationships with a large percentage of libraries. At its core, this is a business decision, and EBSCO is not a charity. The acquisition is a way to maximize profits at a company that has been very good at doing so. There is real value in creating more seamless and streamlined workflows within libraries, and this is especially true where libraries are facing staffing challenges. This recent acquisition, along with the recent demise of Swets, has certainly allowed EBSCO to extend their customer base and increase engagement.

After the initial shock, I suppose not many in the library community were particularly surprised. YBP faces increasing financial challenges as library print book acquisition expenditures fall. As we would expect, the narrative from the parties was upbeat and filled with promises that nothing would change beyond some wonderful synergies to come. EBSCO is continuing to position itself as the closest thing we have to a soup-to-nuts library information content and services vendor. EBSCO claims to be the largest library discovery service provider (6,000+ customers), the leading e-content provider (360,000+ serials, 57,000+ e-journals, 600,000+ e-books). Over the past 20 years, they have moved from a serials jobber to a reseller of abstracting and indexing, a major serials and book jobber, a comprehensive and state-of-the-art discovery service, a contract publisher and aggregator of a massive amount of full-text content through a diverse portfolio of subject databases, an e-book aggregation platform. EBSCO has a strong engagement with development of the KOHA open source ILS. While the materials outcomes are unclear, EBSCO is even a partner in the Open Access Scholarly Publishers Association. Adding YBP also brings in another major library analytics tool, Collection HQ to join EBSCO’s Plum Analytics.

So, what’s the problem?

  1. Not Enough Competition?

What are the risks associated with giving too much of your business to a single company? What if things go sour? Many librarians have been frustrated with the dwindling choices we have in the jobber market. In the halcyon summer of 2014, my institution felt like it had three choices for our serials jobbing. When Swets went under last fall, we were left with only two choices. The book jobber market has faced so much consolidation over the past 2 decades that those of us who are dissatisfied with our book jobbers do not feel we have many options.

  1. Too Big/Monolithic To Fail?

I tend to agree that EBSCO can leverage this merger to enable efficiencies. Those efficiencies will certainly be best realized if a library commits more resources to EBSCO’s family of library services and systems. Business risk assessment professionals can offer comment here, but doing too much business with a single corporate entity is fraught with obvious risks. The movement by EBSCO and others to promote end-to-end library systems flies in the face of efforts to modularize library systems and services. Many have been working to build interoperable components instead of single systems. Modularity allows one to both mitigate risk and promote flexibility and evolution of the local set of systems over time. How will we feel limited when it becomes difficult or impossible to replace an unsatisfactory component of a monolithic implementation? The more singular one’s commitment to a small number of providers, the impacts of failure and obsolescence grow. And what happens if EBSCO’s corporate parent gets into trouble? Admittedly, EBSCO is probably the most diversified company in our space. In fact, it is a little surprising not to see EBSCO’s Vulcan Industries competing with Demco and Brodart in the lucrative library fixtures space.

  1. Can a wholly-owned subsidiary really remain agnostic?

For those of us in the trenches of emerging e-book workflows, a continuing challenge is metadata supply and alignment. This is an emerging area of practice, and while ProQuest, perhaps the major aggregated e-book competitor to EBSCO, says it will be business as usual (PQ Blog: http://www.proquest.com/blog/pqblog/2015/Support-Continues-with-YBP.html ), it seems like they might be more than a little nervous. Often a book is available on multiple platforms, and one wonders if long-standing meta-data workflow issues between YBP and EBL will receive the attention they deserve. If I were a smaller e-book publisher that does not want to join with EBSCO’s platform, I think would be very anxious about the acquisition.

  1. Can we learn anything from EBSCO’s past behavior?

During my 20-year professional career in libraries, I have witnessed a lot of very aggressive behavior from vendors. In my experience, EBSCO has been one of the most aggressive. In the 1990s, at a time when most discovery providers were content to resell existing databases, EBSCO brought new A&I products to market that directly competed with existing ones in business, health care and the social sciences. While one could argue this drove improvements, the outcome was essentially a doubling of our subscription loads. Combined with exclusive agreements with minor and major periodical publishers such as Time-Warner and Harvard Business Publishing, libraries have sometimes felt that EBSCO was forcing the issue and gaining something of an unproductive advantage. In the discovery space, there are those who feel EBSCO takes an all-or-nothing approach where their systems work great if-and-only-if you also bring along all your subscriptions to their platform.

Consolidation and further amalgamation in the library information services market has been a trend for a few years now. We have fewer ILS vendors, fewer jobbers, bigger players, and lots and lots of financial capital trying to make lots and lots of money in our industry. Perhaps resistance is futile; perhaps it is the only rational choice.

Paying the bills

Open access is not free.  By saying that up front, I hope to confound some of the more extreme critics of the open access movement, who sometimes pretend that all OA supporters are dreamy-eyed and woolly-headed librarians who imagine that all information “wants” to be free.  So I start from the premise that open access costs money, but I do immediately need to qualify that statement.  By definition, open access does not charge the consumers of information any fee for access.  It is free in that sense, but not in the sense that there are no costs involved in producing the works or making them freely available.

Of course, for most academic work, the majority of the “first copy” costs are already paid for outside the subscription system.  Taxpayer money, research grants or private university funds support the research underlying scholarship, as well as the time of the authors who produce it.  The work of reviewing and improving both articles and books is also borne by these sources, in large part.  Commercial publishers, then, “free ride” on this investment, which is a sign of a failure in the market.

One truth that lies behind much of the open access movement is simply that there are more efficient ways to produce scholarship than the way it has been done in the past.  Digital communications have brought down the cost of other types of content, but publishers continue to raise prices, arguing, improbably, that there are no significant savings in the transition from print to online distribution.

So here is a simple assertion — we must find ways to pay for open access to scholarship, but simply paying traditional commercial publishers to do it for us would just replicate a system whose failure in the online environment is all too obvious.  Gold, or hybrid gold, open access is NOT a solution IF it is expected to simply replace subscription revenue and does not provide an opportunity for a more realistic assessment of costs.

There is a really nice article by Kevin Hawkins from the University of North Texas Libraries about “How We Pay for Publishing” in the latest issue of “Against the Grain.”  I don’t think the issue is openly available, but there is a pre-print of Kevin’s article in the UNT Digital Library here.  The article is a concise overview of “how we got into this mess” and “how might we get out.”  Kevin concisely summarizes the various proposals that would shift library investments from the consumption side of the research process — where we simply buy the products that are sold only to a limited set of buyers/subscribers — to the production side, where library funding helps support the creation of knowledge products that are available to all.  As Hawkins points out, there is a potential free-rider problem here as well, and he writes about how different possible models can address it.  When he turns to the impact of these models on university presses, he envisions a more mission-driven view of those organizations that would recognize that it is unrealistic to expect them to pay for themselves at the end of each fiscal year.  Instead, we should look closely at costs and benefits, recognizing that increased access is a benefit, and support university presses as well as other organizations that are prepared to experiment with new models for more efficient production and broader access.

Costs, of course, are the elephant in the living room whenever improving access is under discussion.  In a recent blog post on open access books and double dipping, Martin Eve of the Open Library of the Humanities questions the claim that a particular publisher must recoup $17,500 to make a single monograph OA.  Martin’s concern is that the publisher in question may be “double-dipping” by demanding the full-cost of the publication as an up-front OA charge and selling print copies as well.  It seems that the OA charge could be reduced by the amount earned through print or e-book sales, just as journal subscription costs ought to come down as author’s select “open choice” options.  The fact that neither reduction is occurring, at least in a wide-spread way, is a sign that we need to ask much harder questions about the actual cost of production as we make this transition to supporting OA on the production side.  The study announced by Ithaka last month should help us get a better handle on the true cost of publishing a monograph.

Those nations that use central funding to support open access publication — and there are a growing number of those — especially need to look at costs to be sure they are not wasting money.  In this guest post on the Digital Science blog, Jan Erik Frantsvag from Norway considers the path that would simply pay article processing charges (APCs) to commercial publishers in order to achieve open access, and lays out some concerns.  Most especially, he explains why the Norwegian Research Council has decided that the funds offered to researchers to fulfill their open access mandate may not be spent on hybrid publications, where only certain articles are made open access after a fee is paid, while a subscription fee is still charged for access to the entire journal.  Frantsvag cites three concerns that caused the Norwegian government to reject this hybrid model.  First, because it double-dips, as described above.  Second, because hybrid journal APCs are higher than those for other OA publications, raising precisely this issue of cost that we have been discussing.  As more and more government funds are used to pay APCs, it seems likely that we will see more probing questions about cost being asked.  Indeed, if U.S. universities are going to develop campus-based funds to support article process charges, we ought to be asking the same questions, and using caps on our funds to encourage more reasonable APCs.  The third reason Frantsvag cites is simply that money for open access should support innovation and experimentation, not simply go to keeping the old business model afloat.

In each of these examples we see that efforts to encourage open access and find new models for supporting research must necessarily look at how much the production of knowledge products — monographs, journal articles, digital visualizations, etc. — actually costs.  In the past these costs have been obscure to universities, hidden in the steady upward spiral of subscription expenses.  But now we are seeing signs that the academic community is beginning to do what it probably should have done long ago — look closely at the actual costs and begin to evaluate the money we are willing to invest in each form of scholarly dissemination.

Who should we trust?

A recent discussion on an e-mail list, about university open access policies, raised the issue of trust.  The participants correctly noted how important it is that there be some level of trust between faculty, administrators and the library (which is usually charged with managing an open access repository once a policy is in place) in order for an open access policy to be adopted and successfully implemented.  Inevitably, the comments brought out the irony that faculty authors often seem very suspicious about administrator motives when debating an OA policy — fearing that someone is trying to “steal their stuff” — yet are perfectly willing to give that “stuff,” in its entirety and for no remuneration, to commercial publishers.  And they do this even though it is obvious that commercial publishers do not share the fundamental values of academia about research and access.

When one looks a little deeper, it is easy to see, I think, that academic authors do not really trust the commercial publishers either; we hear lots of wry comments about how absurd the current system is, followed by a shoulder-shrug expressing resignation.  It is absurd, but it is what we have got.  “Trust” is probably the wrong word for what authors feel as they give away their work.

Perhaps it really is just resignation in the face of how things have been done for hundreds of years.  There was a fine column published recently in the Educause review that encapsulates this scholarly stagnation with a story about a Buddhist monastery where the practice of tying up the monastery’s cat, begun because it disrupted the abbot’s lessons to novices, continued for centuries, through many successive abbots and cats, until it came to be invested with a false sense of significance.  Mark Edington makes the case that we have become unreasonably wedded to a system that once had considerable utility for a set of circumstances that no longer exist, and are reluctant to embrace the changes needed for our new environment.  There certainly is an element of “tying up the cat” in our continued dependence on commercial publishers and our willingness to pay them to do things that, in many cases, we can now do better and more economically ourselves.  It is simply easier to do what we have always done, and to let the dissemination of scholarship remain in the hands of those who have done it for us for 300 years.

It is not that we exactly trust commercial publishers, nor do we exactly distrust them. We may recognize that the values and goals of the commercial publishing business are different from, and even in conflict with, the best interests of scholarly authors and of scholarship itself.  Perfectly nice people, working to advance their own interests as best they can, come in to conflict as the conditions for research and teaching change. And a real ambivalence is created because of how interwoven the parts of the academic enterprise are.  More than just inertia is a work; important aspects of the academic enterprise remain interlocked with traditional forms of publication.

Most obviously, the rewards system in higher education, which by definition has worked well for many of our most influential faculty, is highly entwined with traditional journal and monograph publishing. In the digital age, when we can actually learn so much more relevant information about the uses made of an individual scholarly article, it may be absurd to continue to look at impact factor and the “brand name” of the journal as evaluative measures, but the systems we have built around those factors are well entrenched.  Change seems inevitable, but it will be slow.  One of my major concerns is the cost of this snails’ pace, and the danger that “the usual suspects” will succeed in co-opting the new opportunities and convincing us to pay them for open access and alt-metrics, when we should look for ways to save money and preserve more local control over these new, digital opportunities.  An example of this danger can be found in the “Open Access Roundtable” sponsored by the Copyright Clearance Center that warned about the complexity of article processing charges.  The CCC, of course, sees itself as a solution to that problem, while the role of universities is simply to pay others, through them, for the benefits of open access.

Late last year Jason Schmitt, a professor at a small liberal arts college, posted a column in the HuffPost called “Academic Journals: The Most Profitable Obsolete Technology in History.”  He makes a good point.  It seems odd that academia, which should be in the forefront of leading managed change, should be propping up a technology that has outlived its usefulness.  We would laugh at people who continued to buy buggy whips well into the automobile era, yet we, as academics and librarians, are in danger of becoming those very laughing-stocks. Schmitt focuses on ideas for change laid out by Sam Gershman and John Bohannan and stresses the societal obligation we have to make the knowledge created in our colleges and universities more accessible.

Of course, trust is a key element in moving toward the open access future that Schmitt, like so many others, envisions.  The systems we are moving toward will not be simple; it will involve multiple and diverse business models, and accommodation for lots of different needs.  We will need to decide what parts of the scholarly communication system we can take over on our own campuses, where we should focus collective funding, and what elements might still be most reasonably “outsourced” to commercial firms.  These decisions will be difficult, and different on each campus.  At each stage we will need to decide who we trust; who are the right partners for this part of the journey.

The key to those trusted relationships, I think, is common values.  As we move toward a more open future for scholarly communication, we need to focus on what are the best interests of scholarship and of individual scholars.  The best way to overcome distrust between faculty and administrators, for example, is to ground the conversation on shared or complementary needs and goals — faculty want maximum attention and impact for their work, and administrators want to build and protect the reputation of the institution; open access, implemented thoughtfully, can bring those goals into alignment. Libraries, which are traditionally highly-trusted entities on campus, must keep their focus where, I hope, it has always been, on making teaching and research easier and better.  In all of our conversation about changing the system of scholarly communications, we should keep that goal as our foundation.  Even more importantly, we need to “walk the talk.”  In our decisions about how we spend our money, especially, we should be guide by, and be seen to be guided by, those best interests of our scholars and our institutions. I firmly believe that those interests will lead us, by the shortest reasonable route, to our open future.

Getting into print the hard way

I hope I will be forgiven some self-promotion if I point out that my first book (that’s a little like saying my “first” marriage – so far it is the only one, and no other is anticipated, but one should never say never) has just been published. It is a handbook of intellectual property designed specifically to address the needs and concerns of researchers and teachers, and it was published on December 10 by the Association of College and Research Libraries. It is available for purchase on the ALA Store site here. There is also a PDF made available under a Creative Commons Attribution/Non-commercial license on the ACRL site.  I also expect to have a PDF in the DukeSpace institutional repository soon.

As a last piece of book promotion, I will note the blog post about the book on the Author’s Alliance web site, which was largely written by me.

All that aside, the real purpose of this post is to tell the story of publishing this book, because I think it is instructive for thinking about scholarly communication today. If I had to start with a moral for the whole story, it is that I should have started with the librarians, because they are a group — the ACRL was wonderful to work with — that can publish in the way most consistent with academic values and needs.

The story begins when I was contacted in 2009 by an editor for the University of Chicago Press, who suggested that I should write a handbook on IP for scholars.  I was immediately interested, and, at the editor’s request, wrote first a proposal and then drafts of the first two chapters.  Those chapters were sent out for review and we received both helpful suggestions and a recommendation to continue to develop the book from three reviewers.  With that step complete, the U of C Press and I signed a contract in late 2010, and I set out on a long and difficult journey to finish and publish the book.

The writing process took me considerably longer than I expected; no surprise there, of course.  When I apologized to my editor I was assured that it was OK; they would be interested in the project, they said, whenever it was complete.  During this time, however, I got the first indications that this would not be a smooth relationship.  I sent in chapters as I finished them, as the editor encouraged me to do, but did not get any of the promised feedback as I progressed.  For a couple of years I was essentially writing in a vacuum.

I finished writing in March 2013 and, in May of that year, finally sent the Press the completed draft of the manuscript, which they sent out for review.  The Press’ policy is that they must have two reviews of the full manuscript that recommend publication before they refer the book to their board for a final thumbs up.  So I began to wait.  After six months I checked in and was told that the Press had sent the manuscript to three reviewers.  One had sent in a report with a recommendation to publish.  One had reviewed just two chapters of the book, asserting that they were only qualified to comment on a portion of the subject matter.  The editor had not yet heard from the third reviewer, and without two full review that recommended publication, they could not move ahead.  The editor told me that the third reviewer would be reminded and that the Press would also look for an alternate reader.

Five months after this, having still heard nothing, I inquired again.  My e-mail went unanswered.  A second e-mail did elicit a response, which was that they were still waiting for that third reviewer.  There was no sign they had ever sought that additional reader.

At this point I was pretty frustrated.  Five people had read at least part of the manuscript, and all had recommended publication.  All of them had also made really helpful suggestions for improving the text which I tried to adopt. But the U. of C. Press was still passively waiting for that one final reader so they could comply to the letter with their self-created rules.  What bothered me most was the apparent lack of effort and commitment I was getting from the Press.  After talking with several academic authors about my situation, I decided to withdraw the book from the U of C Press.

When I told my editor about this decision, the first reaction was to try to talk me out of it and ask me to have patience.  The second reaction was to remind me of our contract and to suggest that I had no alternative but to wait it out with the U. of C.  In response, I suggested two concessions from them that would make the waiting easier.  First, I asked (for the first time, I am embarrassed to admit) for some form of open access to a version or a part of the book.  Second, I asked to be paid half of the advance we had agreed on as a sign of the Press’ commitment to the project.  I was told, however, that the first request was impossible because they needed to protect potential sales for the book, and that the second was contrary to their policy.  In regard to the latter issue, I was pointed to specific language in my contract about when the advance would be paid.  So I looked.  It turns out that the contract language was actually different from what the editor told me it was.  It stated pretty clearly that I would be paid the advance when I sent in the manuscript, not, as I had been told, only when and if the board approved final publication.  I have no idea why the contract language was what it was, or if it was really different from what the editor thought it should be.  I had not changed or negotiated that part of the agreement back when it was signed.  Maybe the Press just figured that no one will read their contracts, so they could claim whatever they wanted about them.

Anyway, I brought this discrepancy to the attention of my editor.  That was the last contact I ever had with that person.  A few days later I got an e-mail from the Director of the Press, apologizing to me for the treatment I had received and confirming that the Press had breached our contract by not paying the advance in a timely fashion.  I was told that I would be released from the contract AND that the Press would still pay the advance I was owed.  Over the next few weeks, we signed a formal letter releasing me to publish the book elsewhere and I received a check for the long-delayed advance.

On the plus side of this relationship, I did get quite a lot of very helpful feedback for my manuscript from the readers (although, of course, I am solely responsible for whatever errors remain).  Also, when faced with the clear evidence that they had not lived up to their side of the bargain, the Press did the right thing (maybe just to get rid of me!).  But on the negative side, I got almost no help from the editor in actually developing the manuscript, the “management” of peer-review was lackadaisical, and I wasted over a year after the book was finished waiting for the Press to get its act together before I asserted my freedom to start looking for another publisher.

After this experience, I spoke to one other commercial publisher, and this time made the expectation of an open access copy a condition from the beginning.  That led to a very short conversation.  So finally I did what I should have done from the start — work with the ACRL.  That experience was smooth, collaborative, and focused on the best way to make my work serve the needs of the audience it was intended for.  The ACRL arranged for excellent and expeditious copy-editing and indexing.  They agreed to immediate open access for the entire book. When there was a problem with the first print run of the book, the ACRL took immediate responsibility and re-printed the whole run.  The process took less than six months from signing a contract to publication.

From this experience I take away three lessons, which I think are worth sharing.  First, the claims about how much effort publishers put into a new book, and the help they provide to authors, are at least sometimes exaggerated.  I was working with a prestigious university press on a book they had solicited in the first place, and yet did not get nearly the kind of assistance or interest that I sometimes hear promised.  Second, it is extremely important to read, negotiate and save a copy of the publication contract.  That my book is now in print is largely due, I believe, to the fact that I could go back to the actual language of our agreement in order to convince the U. of C. to release me to find another publisher.  And finally, as I said at the beginning, libraries and library organizations, in my experience, understand the needs and goals of scholarly authors better than commercial presses.  As the future of academic publishing unfolds, I strongly hope that more and more of it will be in the hands, or under the oversight, of libraries.

 

 

Cancelling Wiley?

Because they were spaced almost a full year apart, I really did not connect the dots when two Canadian universities announced that they were cancelling their “Big Deals” with John Wiley & Sons publisher.  The Times Higher Education reported on the decision at the University of Montreal back in January 2014, while the announcement made by Brock University came only a few weeks ago.  I would not have considered this a trend worth commenting on had it not been for conversations I had last week at the Fall CNI Membership meeting.  During that meeting, two different deans of large university libraries told me, unbidden and in separate conversations, they they were also considering ending their deal with Wiley.  I was struck by the coincidence, which caused me to remember these two announcements from Canada and to begin to ponder the situation.

Two different questions occurred to me when I thought about these four significant cancellations or potential cancellations, all directed at a single publisher.  First, why was Wiley the focus of this dissatisfaction?  Second, what is the next step?

As for what the complaints are about Wiley, the answer is pretty much what it always is — money.  The THE article and the Brock University report both tell us that exchange rates have made the annual “higher than the inflation rate” price increases for these packages even harder to bear than usual.  They also point to another problem.  Pricing is based on the large number of titles included in these package deals, but many of those titles are not very useful.  The Brock post notes that the Wiley package has a significantly higher cost per use than does their Elsevier package, which presumably reflects the fact that many of the titles the University is paying for in the package simply do not get used.  The same reality is probably behind the fact noted by THE that Montreal would subscribe to less than 25% of the titles that had been included in the package it was cancelling.  It would be interesting to find out, a year on, how much those other titles have been missed.

In my conversations with the two library deans, much the same thing was said about Wiley — demanding a large price increase, being inflexible in negotiation, and selling “a lot of junk that I don’t need” in the package.  Libraries are beginning to discover that they do not need to put up with those tactics.  Publishers often tell us that they are publishing so many more articles, which justifies their price increases, and they tell us how selective their flagship journals are.  But when we look at these big deals, it is clear that selectivity is not an across-the-board approach; many articles that are not very useful just slide down the hierarchy to get published in journals whose main purpose is to pad out a “big” deal.

To me the more important question is “what now?”  Unfortunately, many times when a library makes this kind of decision there is actually little money saved, since the funds simply go into re-subscribing to a smaller, selected list of titles from the same publisher.  But presumably some of these cancellations result in dollars saved.  And when they don’t, I propose that libraries ought to reexamine their approach.  When you have cancelled a dross-laden package, think twice before reinvesting all of that money in as many individual subscriptions from the same publisher as possible; make a careful decision about where the division between useful titles and unnecessary ones really lies.  Because here is the thing — money that can be saved and reinvested in open access projects will give us a higher return on our investment, because those projects will provide greater access.

It seems clear that, over time, libraries will need to move more and more of their spending away from the consumption side of scholarly production and do much more to support the creation and dissemination of knowledge directly.  Commercial publishers hope to capture those dollars as well, but one of the real benefits of supporting open access can and should be more freedom from businesses addicted to 30% profits.  I would like to challenge libraries to consider, when they have to cancel, using the money to support non-profit or lower profit open access projects.  Work with a society to provide subvention for a scholarly journal to become OA.  Work with your university press to support OA monographs.  Finally, even if not compelled by immediate budget realities, think about making some strategic cancellations in order to take these kinds of steps.  We know that open access is our future, and it is vital that we take control of that future before others take it from us.

I don’t know if Wiley is the worst offender amongst the large commercial publishers, or whether there is a real trend toward cancelling Wiley packages.  But I know the future of scholarship lies elsewhere than with these large legacy corporations.  The process of weaning ourselves from them will be slow and drawn-out.  But especially when the cancellations are going to happen anyway, we should have the idea of using the funds to advance the transition to open access foremost in our minds.

For a similar, but likely better informed, perspective on the idea of cutting subscriptions to support open access, please read Cameron Neylon’s post on “Letting it go — Cancelling subscriptions, funding transitions,”which ties the idea in his title to the discussion going on in the Netherlands about Elsevier’s big deal.

Public access and protectionism

By now many folks have commented on the announcement from Nature Publishing Group early this week about public access to all of its content and most have sussed out the fairly obvious fact that this is not open access, in spite of the rah-rah headline in the Chronicle of Higher Education, nor even public access as it is defined by many national or funder mandates.  Just to review quickly the major points about why this announcement actually gives the scholarly community so much less than is implied by either of those terms, consider these limitations:

  1. A putative reader can only get to an article if they are sent a link by a subscriber, or the link is present in a news article written by one of the 100 news organizations that NPG has chosen to “honor.”
  2. Articles can only be read inside NPG’s proprietary reader
  3. No printing or downloading is possible, so a non-subscriber hoping to use one of these articles to further her own research better have a darn good memory!
  4. No machine processing will be possible; no text or data mining.

In short, all of the inconveniences of print journals are preserved; what NPG is facilitating here is essentially a replica of loaning a colleague your copy of the printed magazine.  Or, at best, the old-fashioned system whereby authors were given paper “off-prints” to send to colleagues.  Although, honestly, off-prints had more utility for furthering research than this “now you see it, now you don’t” system has.

If this is not open or public access, what is it?  I like the term “beggar access,” which Ross Mounce applied to NPG’s scheme in a recent blog post, since it makes clear that any potential reader must ask for and receive the link from a subscriber.  Some suggest that this is a small step forward, but I am not convinced.  There is nothing public or open about this “ask a subscriber” model; all it really does is prevent scholars from downloading PDFs from their subscription access to NPG journals and emailing them to colleagues who lack a subscription.  In short, it looks like another stage in the ongoing comedy of fear and incomprehension about the way digital scholarship works on the part of a major publisher.  But Mounce’s post suggests that the move is more than that; he points out ways in which it may be designed to prop up digital business that Nature and its parent Macmillan have invested in — specifically ReadCube and AltMetric.com.  The byzantine scheme announced by Nature will drive readers to ReadCube and will generate data for AltMetrics.com, helping ReadCube compete with, for example, Elsevier and their proprietary reading and sharing tool, Mendeley.

That is, this looks like another move in the efforts by the large commercial publishers to buy up and co-opt the potential of open access. On their lips, open access does not mean greater potential for research and the advancement of science; it means a new market to exploit.  If administrators, researchers and librarians allow that to happen, they will have only themselves to blame.

My colleague Haley Walton, who recently attended OpenCon 2014, told me about a presentation made by Audrey Watters that included the idea of “openwashing,” which Watters defines like this:

Openwashing: n., having an appearance of open-source and open-licensing for marketing purposes, while continuing proprietary practices.

This is exactly what is happening in this announcement from NPG; old business models and awkward exploitation of new markets are being dressed up and presented as a commitment to access to scholarship, but the ruse is pretty transparent.  It may quack like a duck, or be quacked about, but this plan is really a turkey.

If NPG really was committed to better access for scientific research, there is a simple step they could take — put an end to the six-month embargo they impose on author self-archiving.  Much of their competition allows immediate self-archiving of an author’s final manuscript version of articles, but Nature does not.  Instead, they require a six-month embargo on such distribution.  So this new move does only very little to ameliorate the situation; the public still cannot see Nature-published research until it is old news.

Speaking of news, at Duke we have a relationship between the office of Scholarly Communications and that of News & Communications whereby we are notified of upcoming articles about research done at Duke.  In many cases, we are able to work with authors to get a version of the article in question into our repository and provide an open link that can be included in the article when it is released, or added shortly after release.  Our researchers find that putting such links in news stories leads to much better coverage of their discoveries and increased impact on their disciplines.  We always do this in accordance with the specific journal policies — we do not want to place our authors in a difficult position — which means that we cannot include Nature-published articles in this program.  To be frank, articles published in Nature remain highly valued by promotion and tenure committees, but relatively obscure in terms of their ability to advance science.  NPG seems to understand this problem, which is why they have selected a small number of news outlets to be allowed to use these tightly-restricted, read-only links.  They want to avoid increasing irrelevance, but they cannot quite bring themselves to take the necessary risk.  The best way they could advance science would be to eliminate the six-month embargo.

It is interesting to consider what might happen if Nature embraced a more comprehensive opportunity to learn what researchers think about open access by tying their “get a link from a subscriber” offer with an announcement that they were lifting the six-month embargo on self-archiving.  That would demonstrate a real commitment to better access for science, and it would set up a nice experiment.  Is the “version of record” really as important to researchers as some claim?  Important enough to tolerate the straightjacket created by NPG’s proprietary links?  Or will researchers and authors prefer self-archiving, even though an earlier version of the article must be used? This is not an obvious choice, and NPG might actually win its point, if it were willing to try; they might discover that their scheme is more attractive to authors than self-archiving.  NPG would have little to lose if they did this, and they would gain much more credit for facilitating real openness.  But the only way to know what the real preference among academic authors is would be for Nature Publishing to drop its embargo requirement and let authors choose.  When they make that announcement, I will believe that their commitment to finding new ways to promote research and learning is real.