Thursday, November 10, 2016

Simpler Times

“The Library of Congress is worried about the exponential growth of the number of journals. By 2025, their shelves will fill up faster than the speed of light. However, a professor of physics assured them there was no problem: exceeding the speed of light is allowed when no information is transmitted.” 

There are references to variations of this joke as far back as 1971. I first heard it in 1983 or 1984, when I was a graduate student. This is how I learned that some academics were concerned about the state of scholarly communication.

In simpler times, the values of publishing and scholarship were well aligned. The number of slots in respected journals was extremely limited, and fierce competition for those slots raised the quality and substance of papers. As publishers became more efficient and savvy, they created more journals and accepted more papers. Scholars competing in the academic job market were always eager to contribute ever more papers. As scholars published more, hiring committees demanded more. A vicious cycle with no end in sight.

It is doubtful that the typical scholar of 2016 produces more good ideas than the typical scholar of 1956. The former certainly writes a lot more papers than the latter. The publish-or-perish culture reduced the scholarly paper to a least publishable unit. The abundance of brain sneeze is correlated with several other issues. Many reported results cannot be reproduced. [A Joke Syllabus With a Serious Point: Cussing Away the Reproducibility Crisis] A growing number of papers are retracted for fraud and serious errors. [Retraction Watch] Clinical trials are hidden when they do not have the desired results. [AllTrials] Fake journals scam honest-but-naive scholars, embellish the scholarly records of fraudulent scholars, and/or provide the sheen of legitimacy to bad research. [Beall's List]

This race to the bottom was financed by universities through their libraries. Every year, they paid higher subscription prices to more journals. In the 1990s, library budgets spiraled out of control and finally caught the attention of university administrators. This was also when the internet grew exponentially. Scholars who realized the web's potential demanded barrier-free online access to research. The Open Access (OA) movement was born.

Good scholarship is elitist: we expect scholars to gain status and influence for getting it right, particularly when they had to fight against majority opinion. Journals are essential components in the arbitration of this elitism. Yet, even well before the OA movement, it was in the publishers' interest to lower the barriers of publishing: every published paper incentivizes its authors to lobby their institutions in favor of a journal subscription.

Gold OA journals [Directory of Open Access Journals] with business models that do not rely on subscription revenue made the problem worse. They were supposed to kill and replace subscription journals. Instead, subscription journals survived virtually intact. Subscriptions did not disappear. Their impact factors did not fall even after competing Gold OA journals scaled the impact-factor ladder. The net result of Gold OA is more opportunities to publish in high-impact-factor journals.

The Green OA strategy had a plausible path to reverse the growth of journals: libraries might be able to drop some subscriptions if scholars should shift their use to Green OA institutional repositories (IRs). [OAI Registered Data Providers] This outcome now seems unlikely. I previously argued that IRs are obsolete, and that the Green OA strategy needs social networks that create a network effect by serving individual scholars, not their institutions. [Let IR RIP] In an excellent response by Poynder and Lynch [Q&A with CNI’s Clifford Lynch: Time to re-think the institutional repository?], we learned how some academic libraries are contracting with Elsevier to manage their IRs. They seem to have given up on Green OA as a strategy to reclaim ownership of the scholarly literature from publishers. They have pivoted their IRs towards a different and equally important goal: increasing the visibility and accessibility of theses, archives, technical papers, lab notebooks, oral histories, etc.

The OA movement tried to accomplish meaningful change of the scholarly-communication system with incremental steps that preserve continuity. I called it isentropic disruption. [Isentropic Disruption] However, scholarly publishers have proven extra-ordinarily immune to any pressure. Just the transition to digital wiped out every other kind of publisher. Scholarly publishers did not even change their business model. They also brushed off reproducibility and fraud scandals. They survived boycotts and editorial-board resignations. They largely ignored Green and Gold OA. Perhaps, the OA movement just needs more time. Perhaps, the OA movement is falling victim to a sunk-cost fallacy.

The current system is financially not sustainable and, worse, is bad for scholarship. Within the shared-governance structure of universities, it is virtually impossible to take disruptive action in the absence of immediate crisis. Universities tend to postpone such decisions until no alternative remains. Then, they inflict maximum pain by implementing unplanned change overnight.

Yet, there are options available right now. With time to plan a transition, there would be much less collateral damage. For example, I proposed replacing library site licenses with personal subscriptions to iTunes-like services for academics. [Where the Puck won't be] Personal digital libraries would be much easier to use than the current site-licensed monstrosities. With scholars as direct customers, the market for these services would be extremely competitive. By configuring and using their personal library, scholars would create market-driven limits on the number of available publication slots. Those willing to consider out-of-the-box crazy approaches can even achieve such limits within an OA context. [Market Capitalism and Open Access]

Academics created the problem. Only academics can solve it. Not libraries. Not publishers. Digital journals are already filling the virtual shelves at the speed of light... The punch line of the joke is in sight.

Sunday, July 24, 2016


The Institutional Repository (IR) is obsolete. Its flawed foundation cannot be repaired. The IR must be phased out and replaced with viable alternatives.

Lack of enthusiasm. The number of IRs has grown because of a few motivated faculty and administrators. After twenty years of promoting IRs, there is no grassroots support. Scholars submit papers to an IR because they have to, not because they want to. Too few IR users become recruiters. There is no network effect.

Local management. At most institutions, the IR is created to support an Open Access (OA) mandate. As part of the necessary approval and consensus-building processes, various administrative and faculty committees impose local rules and exemptions. After launch, the IR is managed by an academic library accountable only to current faculty. Local concerns dominate those of the worldwide community of potential users.

Poor usability. Access-, copy-, reuse, and data-mining rights are overly restrictive or left unstated. Content consists of a mishmash of formats. The resulting federation of IRs is useless for serious research. Even the most basic queries cannot be implemented reliably. National IRs (like PubMed) and disciplinary repositories (like ArXiv) eliminate local idiosyncrasies and are far more useful. IRs were supposed to duplicate their success, while spreading the financial burden and immunizing the system against adverse political decisions. The sacrifice in usability is too high a price to pay.

Low use. Digital information improves with use. Unused, it remains stuck in obsolete formats. After extended non-use, recovering information requires a digital version of archaeology. Every user of a digital archive participates in its crowd-sourced quality control. Every access is an opportunity to discover, report, and repair problems. To succeed at its archival mission, a digital archive must be an essential research tool that all scholars need every day.

High cost. Once upon a time, the IR was a cheap experiment. Today's professionally managed IR costs far too much for its limited functionality.

Fragmented control. Over the course of their careers, most scholars are affiliated with several institutions. It is unreasonable to distribute a scholar's work according to where it was produced. At best, it is inconvenient to maintain multiple accounts. At worst, it creates long-term chaos to comply with different and conflicting policies of institutions with which one is no longer affiliated. In a cloud-computing world, scholars should manage their own personal repositories, and archives should manage the repositories of scholars no longer willing or able.

Social interaction. Research is a social endeavor. [Creating Knowledge] Let us be inspired by the titans of the network effect: Facebook, Twitter, Instagram, Snapchat, etc. Encourage scholars to build their personal repository in a social-network context. Disciplinary repositories like ArXiv and SSRN can expand their social-network services. Social networks like, Mendeley, Zotero, and Figshare have the capability to implement and/or expand IR-like services.

Distorted market. Academic libraries are unlikely to spend money on services that compete with IRs. Ventures that bypass libraries must offer their services for free. In desperation, some have pursued (and dropped) controversial alternative methods of monetizing their services. [Scholars Criticize Proposal to Charge Authors for Recommendations]

Many academics are suspicious of any commercial interests in scholarly communication. Blaming publishers for the scholarly-journal crisis, they conveniently forget their own contribution to the dysfunction. Willing academics, with enthusiastic help from publishers, launch ever more journals.[Hitler, Mother Teresa, and Coke] They also pressure libraries to site license "their" journals, giving publishers a strong negotiation position. Without library-paid site licenses, academics would have flocked to alternative publishing models, and publishers would have embraced alternative subscription plans like an iTunes for scholarly papers. [Where the Puck won't be] [What if Libraries were the Problem?] Universities and/or governments must change how they fund scholarly communication to eliminate the marketplace distortions that preserve the status quo, protect publishers, and stifle innovation. In a truly open market of individual subscriptions, start-up ventures would thrive.

I believed in IRs. I advocated for IRs. After participating in the First Meeting of the Open Archives Initiative (1999, Santa Fe, New Mexico), I started a project that would evolve into Caltech CODA. [The Birth of the Open Access Movement] We encouraged, then required, electronic theses. We captured preprints and historical documents. [E-Journals: Do-It-Yourself Publishing]

I was convinced IRs would disrupt scholarly communication. I was wrong. All High Energy Physics (HEP) papers are available in ArXiv. Being a disciplinary repository, ArXiv functions like an idealized version of a federation of IRs. It changed scholarly communication for the better by speeding up dissemination and improving social interaction, but it did not disrupt. On the contrary, HEP scholars organized what amounted to an an authoritarian take-over of the HEP scholarly-journal marketplace. While ensuring open access of all HEP research, this take-over also cemented the status quo for the foreseeable future. [A Physics Experiment] 

The IR is not equivalent with Green Open Access. The IR is only one possible implementation of Green OA. With the IR at a dead end, Green OA must pivot towards alternatives that have viable paths forward: personal repositories, disciplinary repositories, social networks, and innovative combinations of all three.

*Edited 7/26/2016 to correct formatting errors.