Changing the paradigm

SH: Here is a quick summary of the points of agreement and disagreement with the UC view  of OA and IRs as described by CC:

(1) UC considers publication reform to be the goal, OA merely a means: I consider OA to be the goal and publication reform merely a hypothetical possibility that might or might not follow from OA.

(2) UC considers providing OA to postprints (i.e., final drafts of published journal articles) a lesser priority for IRs, I consider them to be the first priority.

(3) UC  moved away from Eprints and postprint self-archiving because of the extremely low level of spontaneous uptake by UC faculty, assuming it was because it was "too difficult." In reality it was because UC did not adopt an institutional self-archiving requirement. Those unstitutions that have done so have dramatically higher self-archiving rates.

(4) UC instead outsourced self-archiving to an expensive service that, being a secondary publisher, needs to expend a lot of resources on following up rights problems for each paper; the result so far is that UC is still not self-archiving more than the 15% worldwide self-archiving baseline for postprints.

(5) The other reason UC moved away from Eprints and postprint archiving is because of its publishing reform goals, including university self-publishing (of journals and monographs). I think monographs are (for the time being) a separate matter, and should be handled separately from journal article OA, and that peer review needs to be implemented by a neutral  3rd party, not the author or the author's institution. The immediate priority is postprint OA.

In summary, UC seems to be giving its own hypothetical views on the future of scholarly publishing -- and its own aspirations for the hypothetical new publishing system -- priority over an immediate, pressing, and remediable practical problem: the needless, daily loss of  25% - 250% or more of the usage and impact of 85% of UC research output. Because researchers are relatively  uninformed
and uninvolved  in all this, they do not have a clear sense of the implicit trade-off between (a) the actual daily, cumulative usage/impact loss for their own research output, with its tested and demonstrated remedy,  and (b) the untested hypothetical possibilities with which the library community seems to be preoccupied.

RP: Initially you built the eScholarship Repository with the EPrints software, which was developed at Southampton University in the UK?

CC: Right. We started with Eprints, and the aim was to create what people now call an institutional repository — a repository where faculty could put materials (text and images) that they wanted to disseminate, or actually publish.

RP: You later switched to the bepress software. Why?

CC: We found it so, so, so difficult to get faculty even to test the EPrints software that we abandoned the idea of providing a platform for faculty to individually publish  [sic] their own works.

SH: I think here is where the strategic error occurred. Not in switching softwares (since the software makes absolutely no difference) but in abandoning the goal of 100% OA for UC postprint output. The reason is implicit in the words CC uses to describe it: The self-archiving of already published postprints is not publishing at all, but merely OA-provision -- except if the underlying goal is not OA, but self-publishing!

 CC: Around the same time we serendipitously encountered the bepress software, and right away we could see that it would to allow us to do something much more important. We could see that if we used the bepress software the repository could also support peer-reviewed publications. Consequently, by the time we launched we had switched to a different model, and we had adopted the bepress software.

SH: Again, it is hypotheses about publishing reform and aspirations for UC self-publishing that motivated the change the change of "model." (Model for what, one wonders? OA is not a model. It just means making journal articles free for all online. It is publishing reform that involved models. Better if UC had done the tested, demonstrated part first, by adopting an institutional self-archiving policy, as at least four other universities have since done, successfully, and once the doable part was successfully done, moved on to the hypothetical part...)

RP: How was the model different?

CC: The bepress software allowed different units within the University of California to become the gatekeepers, with all the editorial and administrative ability resting with an academic department, an institute, or a research unit, rather than with individual faculty, or with the library.

SH: There are two issue here: (1) Did the Eprints software allow departments or research units to be their own gate-keepers for self-archiving? Of course it did, either within one Eprints installation, or, optionally, across many, thanks to OAI interoperability. But much more important is: (2) Is local gate-keeping the goal of UC researchers? Has the gate-keeping not already been done by the peer-reviewers for the journal in which it was published? It looks here as if, once again, the hypotheses about publishing reform and UC self-publishing are driving the agenda, not researchers' immediate needs (which are for maximizing research access, usage and impact, via OA).

RP: So where EPrints software assumed that researchers would do the inputting of papers themselves, bepress software was more suited to third-parties depositing them?

CC: That is one difference — although, because the software is difficult to use, Eprints submissions are often managed centrally.

SH: As this is not about defending the Eprints software in particular, I only note in passing that the difficulty was not the software but the fact that UC researchers were not required to self-archive, and hence didn't. In institutions where self-archiving is required, it is done, easily, by researchers themselves, not centrally. The central proxy self-archiving is a start-up strategy, used successfully by some institutions to get the practice started; it is not a feature of the software:

Carr, L. and Harnad, S. (2005) Keystroke Economy: A Study of the Time and Effort Involved in Self-Archiving. 

CC: Additionally, the bepress software lent itself to the size of UC; and it allowed the University to decide exactly what it wanted to put in, and to brand everything in the way it wished.

SH: All the free softwares are likewise configurable in exactly the same way.

RP: You were also able to outsource the hosting of the eScholarship Repository to bepress?

CC: Yes. It is hosted by bepress, but preserved here at CDL. What we are doing is harvesting citations. We then send them to faculty members saying that the listed works may be eligible for inclusion in the eScholarship repository. It is a way to alert them to the repository, and to the fact that they have content that could be placed in it.The message sent to faculty is clickable, and when they click on the link it brings them directly into the repository, where the citation data for the paper automatically fills out the repository metadata fields for them. This, by the way, is the one case where we allow authors to put their content in directly themselves. However, we also allow them to use a proxy — so they can legally assign someone else to put their papers in for them. The aim is to make the process as easy as possible, because time is the biggest constraint when it comes to getting faculty to participate.

SH: So far, this is all excellent practice, and an ingenious start-up strategy (though only if coupled with an institutional self-archiving requirement). But it is the next step that defeats it:

RP: And you have contracted bepress to do rights clearance on the papers?

CC: Right. After the papers are submitted we pay bepress to check the rights on them. That was a concession to the fact that bepress' business would be threatened if they got sued for allowing something illegal to be put into the repository. This part of the process is both onerous and expensive, and we hope we will not need to do it at some point in the future.

SH: So because UC have gotten into a 3rd-party publisher situation, they face rights problems they would not face if it were all in-house self-archiving. They are also incurring great additional expense (at a time when institutions are being deterred from IRs and OA under the false impression that it is expensive). Worst of all, so far the result is still not more UC postprints beocming OA than the global 15% average:

RP: I'm told you have acquired about 1,000 papers in this way... 1,000 postprints is a small drop in the ocean I guess. How many researchers are there within the UC system?

CC: UC is the largest public research university I know of. It has ten campuses and around 16,000 faculty and researchers.

RP: When you ask faculty for a postprint is it a request or a demand?

CC: It is not a demand. Clearly, incentive is the single biggest issue for getting content in. Awareness is another issue, so we are just starting some market research to discover what percentage of UC faculty even know about the repository. I suspect it is less than half.

SH: Perhaps a UC self-archiving requirement would be worth considering after all, since several interational surveys have now reported that 70% - 95% of faculty say they would comply with a self-archiving requierement, and the 4 institutions that have adopted such a policy so far confirm that it works.

Swan, A. and Brown, S. (2005) Open access self-archiving: An author study. Technical Report, External Collaborators, Key Perspectives Inc..

Institutional Self-Archiving Policy Registry

RP: So you still have work to do in publicising the repository?

CC: We do. While we are very excited that we have more than 200 departments participating in the repository we have no idea what percentage of the faculty know about it; and we have no idea what percentage would participate if they did know — because there is no overriding incentive for them to do so today. We need to understand the situation.

SH: The missing element is the institutional requirement to deposit the final accepted, peer-reviewed draft (not the publisher's PDF) as an institutional record-keeping matter: a fulfillment condition for annual review, for research assessment, and for standard CV creation/submission.

RP: As your experience shows, creating a repository is only half the task. You then have to fill it. For that reason there are growing calls for funders to mandate researchers to self-archive their papers. Do you think that that is the best way of filling institutional repositories?

CC: Well, I wouldn’t say that our purpose is simply to fill institutional repositories. We built an institutional repository as one way of providing an alternative to the current publishing system, and to give faculty something to do with that copyrighted material that we keep saying  shouldn’t be given away to publishers.

SH: Filling an OA IR with the institution's annual research article output may not be the only possible goal for an IR, but it is surely the most important priority at this moment for researchers, who need not an alternative to the current publishing system but OA. Copyright renention is not an end in itself for researchers either: OA is. And with OA, copyright retention becomes moot.

CC:  It may turn out that institutional repositories aren’t the way to go however. For that reason we are also interested in encouraging faculty to manage their copyrights differently, and to consider who they give their manuscripts to, and where they commit their editing and reviewing time. So our main focus is in accomplishing that, rather than filling repositories.

SH:  Why all this, when, in and of itself, that is not what faculty want and need? It would be fine if copyright retention were an essential means to an end that faculty do want and need, but it is not. OA is an end in itself, and it does pto require copyright retention when 93% of journals have already given OA author self-archiving their green light:

RP: Do you nevertheless anticipate that  funders will eventually introduce mandates?

CC: Actually we expect that universities will make some sort of a mandate before funding agencies do. In this regard there are a number of white papers floating around the University of California right now. We are waiting to see what happens to those.

SH: But the UC proposal is for copyright retention, whereas what is needed is a self-archiving requirement. Copyright retention requires needless re-negotiation with the 93% of journals that have already endorsed OA self-archiving, and it puts 100% of authors at risk of an unsuccessful re-negotiation, instead of just requiring that 100% of them deposit, leaving the 7% to set access as restricted, pending negotiations, if they wish?

RP: ...Given what you say about rights, I 'd be interested to hear more about the Scholarly Work Copyright Rights Policy white paper. This proposes that UC faculty "routinely grant to The Regents of the University of California a limited, irrevocable, perpetual, worldwide, non-exclusive licence to place the faculty member's scholarly work in a non-commercial open-access online repository." Would this apply only to journal articles or all the works of faculty, including books?

CC: Ultimately it is intended to apply to all works, but starting with journal articles.

SH: Does it make sense to hold back (and weigh down) the sure research benefits of the self-archiving of published journal articles (postprints) for the much vaguer and more controversial case of books?

RP: If it does go ahead would you envisage a postprint mandate following behind it?

CC: Yes.

SH: A postprint mandate should not come behind a copyright blanket retention mandate! That is like making a local emission-reduction plan's adoption contigent on getting all nations to agreeto sign to Kyoto Agreement!

RP: And you would welcome that?

CC: I would. While I don’t find the postprint issue as interesting or exciting as trying to encourage new forms of communication, it is strategically important — because it would allow us to put in place a production-level service capable of managing UC copyrighted material, which would better prepare us for the future.

SH: Then why not adopt a posprint self-archiving mandate immediately, instead of waiting for agreement on the much more demanding and controversial copyright-retention policy?

CC: ... eScholarship Editions are scholarly monographs encoded in XML. ...As you know, the corollary to the serials crisis is that libraries have less money to buy monographs, and so fewer monographs are being published. The fact is, however, that an awful lot of monographs could be published if the UC Press had more editorial bandwidth. So we have been experimenting with empowering UC Press editorials boards, or faculty editorial boards, to become, essentially, publishers. In this way we can extend the work of UC Press.

SH: This is the UC self-publishing agenda, and it is fine, but why is it being coupled with the OA IR issue, and worse, why is it being allowed to hold it back? The (1) UC authors who publish their articles in established peer-reviewed journals may often be the same as the (2) UC authors of monographs, but their situations are very different. The article authors already have publishers (not UC!) and need only OA. The monograph authors may or may not have a publisher, which may or may nto be UC, and they may or may not want OA. Why should the straightforward solution for (1) be constrained by the much less straightfroward solution for (2)?

RP: It's clear you have a very broad view of the role of an institutional repository. Advocates of self-archiving, by contrast, insist that an institutional repository should only ever be viewed as a postprint archive. What's your response to that view?

CC: I think it is unfortunate that the term institutional repository has come to mean something narrower. As I say, the postprint component is the least interesting and ultimately the least important part of this. So while right now it is tactically extremely important to deposit postprints, ultimately I envision a very different arrangement between universities and publishers than we have now.

SH: The reasoning here is unclear: Postprint OA is clearly the heart of the OA movement, and an end in itself (even if there are further ends thereafter). CC agrees that "right now it is tactically extremely important to deposit postprints." Yet UC is not doing what needs to be done to achieve that "narrower" immediate goal. It is instead aiming at the "wider" hypothetical one, and the result is that only 1000 of the "extremely important" postprints have been deposited in the UC IR to date, while "white papers" are being written about retention, publishing reform, and UC self-publishing plans. If the narrower posprint target is indeed an important prerequisite for the rest, then why not make at a concerted affort to reach it first, and leave the more hypothetical phase for afterward? Or at least do it in parallel.

RP: You believe universities should be in control of the publishing process, rather than managing papers that have been published by someone else?

CC: That's right. Eventually I hope all the content will be hosted and managed by universities themselves, and the publishing services would be in the form of added value. So, for instance, a published article would refer back to the raw article in the repository.

SH: This is all fine, but completely speculative. The course that will be taken by journal publishing and monograph publishing, published by universities or published by others, is right now a matter of pure speculation, whereas the course that is taken in access-provision to a university's own postprint output is a practical matter entirely in the hands of the university and its researchers. Why is immediate OA to postprints being held hostage to hypotheses about eventual publishing reform?

RP: What worries self-archiving advocates about this is that if universities try to make institutional repositories too broad in functionality they could delay the transition to an open access environment; that we need to stay focused on the narrower view until OA is achieved. You are arguing that we need to plan for the longer-term future from day one are you?

CC: I think so. Moreover, I don’t see why a broader view would slow OA down. It is a matter of getting the right platform and getting things moving so that faculty can see that there are other things that can be done.

SH: But we have a clear example of "why a broader view would slow OA down"! In 2001 UC adopted Eprints and waited to see whether its IR would fill spontaneously. It did not. So instead of adopting a self-archiving policy (as Southampton, QUT, Minho, and CERN have since done, successfully filling their archives -- Eprints, Eprints, Dspace, and CDSware, respectively), UC adopted another software -- and another agenda instead of OA: publishing reform, copyright retention, and university self-publishing.

RP: I wonder if we might see increasing tension between researchers and librarians over the issue of institutional repositories? I ask because the primary aim of researchers is to achieve maximum impact for their research; librarians, by contrast, are looking to create large digital libraries or even, as in the case of UC, complete publishing systems. Could this threaten the historic relationship between librarians and researchers?

CC: I can see such a tension theoretically: where resources were limited, for instance, the aim of building a digital library could seem to stand in the way of getting publishing out quickly. But ultimately I think you are presenting a false dichotomy.

SH: The only tension is about lost time. UC, the world's biggest university system, 5 years down the line after establishing one of the first IRs, has 1000 postprints of UC published journal articles therein. Meanwhile, tiny Minho has 3297 items, QUT  has 2194, Southampton 7745 plus 9795 for its ECS department alone, and CERN, larger but nowhere near UC in size,  has  75,000 items. Assuming (as at Southampton and CERN) that about 70% of these at least are postprints, it looks very much as if an institutional postprint self-archiving policy has served these other institutions well. Particularly instructive is CERN, which, now that is it well on the road to 100% postprint OA is now, only now, turning to the question of publishing reform. If all other universities and research institutions (including the biggest, UC) were to do likewise, we would already be there (at 100% OA) and in a far better position to contemplate the hypothetical horizons of ensuing publication reform.