Monday, December 16, 2013

Beall's Rant

Jeffrey Beall of Beall's list of predatory scholarly publishers recently made some strident arguments against Open Access (OA) in the journal tripleC (ironically, an OA journal). Beall's comments are part of a non-refereed section dedicated to a discussion on OA.

Michael Eisen takes down Beall's opinion piece paragraph by paragraph. Stevan Harnad responds to the highlights/lowlights. Roy Tennant has a short piece on Beall in The Digital Shift.

Beall's takes a distinctly political approach in his attack on OA:
“The OA movement is an anti-corporatist movement that wants to deny the freedom of the press to companies it disagrees with.”
“It is an anti-corporatist, oppressive and negative movement, [...]”
“[...] a neo-colonial attempt to cast scholarly communication policy according to the aspirations of a cliquish minority of European collectivists.”
“[...] mandates set and enforced by an onerous cadre of Soros-funded European autocrats.”
This is the rhetorical style of American extremist right-wing politics that casts every problem as a false choice between freedom and – take your pick – communism or totalitarianism or colonialism or slavery or... European collectivists like George Soros (who became a billionaire by being a free-market capitalist).

For those of us more comfortable with technocratic arguments, politics is not particularly welcome. Yet, we cannot avoid the fact that the OA movement is trying to reform a large socio-economic system. It would be naïve to think that that can be done without political ideology playing a role. But is it really too much to ask to avoid the lowest level of political debate, politics by name-calling?

The system of subscription journals has an internal free-market logic to it that no proposed or existing OA system has been able to replace. In a perfect world, the subscription system uses an economic market to assess the quality of editorial boards and the level of interest in a particular field. Economic viability acts as a referee of sorts, a market-based minimum standard. Some editorial boards deserve the axe for doing poor work. Some fields of study deserve to go out of business for lack of interest. New editorial boards and new fields of study deserve an opportunity to compete. Most of us prefer that these decisions are made by the collective and distributed wisdom of free-market mechanisms.

Unfortunately, the current scholarly-communication marketplace is far from a free market. Journals hardly compete directly with one another. Site licenses perpetuate a paper-era business model that forces universities to buy all content for 100% of the campus community, even those journals that are relevant only to a sliver of the community. Site licenses limit competition between journals, because end users never get to make the price/value trade-offs critical to a functional free market. The Big Deal exacerbates the problem. Far from providing a service, as Beall contends, the Big Deal gives big publishers a platform to launch new journals without competition. Consortial deals are not discounts; they introduce peer networks to make it more difficult to cancel existing subscriptions. [What if Libraries were the Problem?] [Libraries: Paper Tigers in a Digital World]

If Beall believes in the free market, he should support competition from new methods of dissemination, alternative assessment techniques, and new journal business models. Instead, he seems to be motivated more by a desire to hold onto his disrupted job description:
“Now the realm of scholarly communication is being removed from libraries, and a crisis has settled in. Money flows from authors to publishers rather than from libraries to publishers. We've disintermediated libraries and now find that scholarly system isn't working very well.”
In fact, it is the site-license model that reduced the academic library to the easy-to-disintermediate dead-end role of subscription manager. [Where the Puck won't Be] Most librarians are apprehensive about the changes taking place, but they also realize that they must re-interpret traditional library values in light of new technology to ensure long-term survival of their institution.

Thus far, scholarly publishing has been the only type of publishing not disrupted by the Internet. In his seminal work on disruption [The Innovator's Dilemma], Clayton Christensen characterizes the defenders of the status quo in disrupted industries. Like Beall, they are blinded by traditional quality measures, dismiss and/or denigrate innovations, and retreat into a defense of the status quo.

Students, researchers, and the general public deserve a high-quality scholarly-communication system that satisfies basic minimum technological requirements of the 21st century. [Peter Murray-Rust, Why does scholarly publishing give me so much technical grief?] In the last 20 years of the modern Internet, we have witnessed innovation after innovation. Yet, scholarly publishing is still tied to the paper-imitating PDF format and to paper-era business models.

Open Access may not be the only answer [Open Access Doubts], but it may very well be the opportunity that this crisis has to offer. [Annealing the Library] In American political terms, Green Open Access is a public option. It provides free access to author-formatted versions of papers. Thereby, it serves the general public and the scholarly poor. It also serves researchers by providing a platform for experimentation without having to go through onerous access negotiations (for text mining, for example). It also serves as an additional disruptive trigger for free-market reform of the scholarly market. Gold Open Access in all its forms (from PLOS to PEERJ) is a set of business models that deserve a chance to compete on price and quality.

The choice is not between one free-market option and a plot of European collectivists. The real choice is whether to protect a functionally inadequate system or whether to foster an environment of innovation.

Monday, December 2, 2013

Amazon Floods the Information Commons

Amazon is bringing cloud computing to the masses. Any individual with access to a browser now has access to almost unlimited computing power and storage. This may be the moment that marks the official beginning of the end of the desktop computer, which was already on a downward slide because of the rise of notebooks, netbooks, tablets, and smartphones.

For managers of computer labs, this technology eliminates a slew of nitty gritty management problems without good solutions. When a shared computer is idle, do you take action after 5, 10, or 15 minutes? If you wait too long, you annoy users who are waiting for their turn, and you invite unauthorized users to sneak into someone else's session. If you act too soon, you ruin the experience for the current user. Should you immediately log off an idle user or do you lock the screen for a while before logging off? Again, you balance the interests of the current user against those of the next user. Which software do you install where? Installing all software on every computer is usually too expensive. But if each computer in the lab has its own configuration, how do you communicate those differences to the users? The ultimate challenge of the shared computer is how to let students install software that they themselves are developing while keeping the computer relatively secure, usable to others, and free from pirated software.

Amazon has solved all of this and more. With cloud-based computers, there is no such thing as an idle computer, only idle screens. Shutting down a screen and turning it over to another user does not ruin a session in progress. It is more like turning over a printer. The cloud-based personal computer is configured for one user according to his or her requirements. Students and faculty can install whatever software they need, including their own research software. As to the usual suite of standard applications, cloud services like Adobe Creative Cloud, Google Apps, and Windows Azure have eliminated software installation and maintenance entirely.

The potential of cloud computing in the Information Commons is more than substituting one technology with another. Students and faculty suddenly have their own custom computing laboratory with an unlimited number of computers over which they have complete control. One can imagine projects in which cloud-based computers harvest measurements from sensors across the globe (weather-related, for example), read and analyze the news, and data mine social networks. All of this data can then be fed to high-performance servers running research software for analysis and visualization.

Currently, retail pricing for a cloud-based personal computer starts at $35 per month. This is already a very good price point, considering that it eliminates the hardware replacement cycle, software maintenance, security issues, etc. One can also add and drop computers as needed. Moreover, this is a price point established before competitors have even entered the market. 

When computing and storage become relatively inexpensive on-demand commodity services, computing labs are no longer in the business of sharing computing devices, storage, and software; they are in the business of sharing visualization devices. Currently, Information Commons provide large-screen high-resolution monitors attached to a computer. As large-scale, high-performance, big-data projects grow in popularity across many disciplines, there will be increasing demand for more advanced equipment to visualize and render the results. Today's computing labs will morph into advanced visualization labs. They will provide the capacity to use multiple large high-resolution screens. They may provide access to CAVEs (CAVE Automatic Virtual Environment) and/or additive-manufacturing equipment (which includes 3-D printing). The support requirements for such equipment are radically different from those for current computer labs. CAVEs need large rooms with no windows, multiple projectors, and a sound system. Additive manufacturing may be loud and may require specialized venting systems.

For managers of Information Commons, it is not too early to start planning for this transition. They may look forward to getting rid of the nitty-gritty unsolvable problems mentioned above, but integrating these technologies into the real estate currently used for computing labs and libraries will require all of the organizational and management skills they can muster.

Tuesday, November 5, 2013

Cartoon Physics

When Wile E. Coyote runs off a cliff, he starts falling only after he realizes the precariousness of his situation.

In real life, cartoon physics is decidedly less funny. Market bubbles arise when a trend continues far past the point where the fundamentals make sense. The bubble bursts when the collective wisdom of the market acts on a reality that should have been obvious much earlier. Because of this unnecessary delay, bubbles inflict much unnecessary damage. We saw it recently with the Internet and mortgage bubbles, but the phenomenon is as old as the tulip bubble of 1637.

We also see cartoon physics in action at less epic scales. Cartoon physics applies to almost any disruptive technology. The established players almost never adapt to the new reality when fundamentals require it or when it is logical to do so. Instead of preparing for a viable future, they fight a losing battle hanging onto the past. Most recently, Blackberry ignored the iPhone thinking its serious corporate clients would not be lured by its gadgetry. There is a long line of disrupted industries whose leadership ignored upstart competitors and new realities. This has been the topic of acclaimed academic studies and popularized in every possible venue.

The blame game is a significant part of the process. The recording industry blamed pirates for destroying the music business. In fact, their own neglect to adapt to a digital age contributed at least as much to the disruption.

The scenario is well known, by now too cliché to be a good movie. Leaders of industries in upheaval should know the playbook. Yet, they keep repeating the mistakes of their disrupted predecessors.

Wile E. Coyote finally learned his lesson and decided to stop looking down.

PS: Cartoon physics does not apply to academic institutions, which are protected by their importance and seriousness.

Wednesday, October 9, 2013

Where the Puck won't be

“I skate to where the puck is going to be, not where it has been.”

The academic library has, by default, tied its destiny to a service with no realistic prospects of long-term survival. It has become a systems integrator that stitches together outsourced components into a digital recreation of a paper-based library. This horseless carriage provides the same commodity service to an undergraduate student majoring in chemistry, a graduate student in economics, and a professor of literature. Because it overwhelms the library's budget, organizational structure, and decision-making processes, this expensive and inefficient service hampers innovation in areas that are the library's best hope for survival.

A paper-based library gradually builds a collection of ever-increasing value, and its overhead builds permanent infrastructure. Its digital recreation never builds lasting value. It is a maintenance service, and its overhead is pure inefficiency. This overhead, duplicated at thousands of universities, starts with the costs of preparing for and conducting near-futile site-license negotiations. To shave off a point here and there, the library spends countless staff hours on usage surveys, faculty discussions, consortium meetings, and negotiations with publishers and their middlemen. But the game is rigged. If 15% of a campus wants Journal A, 15% competing Journal B, 10% wants both A and B, and the rest wants neither, the library is effectively forced to rent both A and B for 100% of the campus. This is why scholarly publishers were able to raise prices at super-inflationary rates during a time when all other publishers faced catastrophic disruption. After conducting expensive negotiations, after paying inflated prices, the library must still pay for, build, and maintain the platform that protects publishers' interests by keeping unwanted users out.

Many academics and librarians hope that Open Access efforts will provide an exit from this unsustainable path. If successful, Green Open Access will lead to price reductions and journal cancellations. Gold Open Access seeks to replace site licenses with author page charges. Either strategy reduces the efficiency of library-mediated digital lending by spreading its fixed overhead costs over fewer and/or less expensive journals. New business models for journals, alternative metrics that give scholarly credibility to unbundled works, and any other innovation that competes with site licenses will reduce efficiency even further. All of these factors hasten the demise of an unsustainable service that is already collapsing under its own weight.

Traditionally, a library adapts in response to changing user behavior, attitude, and opinion. However, the Wayne Gretzky quote became a cliché for a reason. When trends have become obvious and users have moved on, it is too late for strategic restructuring.

At the other extreme, an angel investor bets on someone with a compelling idea, accepts the risk of failure, and is prepared to move on to the next player who knows where the puck will be. The library does not have that luxury. It is an institution, not a venture.

The library must maintain sufficient institutional stability to ensure its archival mission. While Open Access is a given, the service portfolio of the future library is far from settled. We must create budgetary and organizational space for new services. We may not know where precisely the puck will be, but we can still move the team out of a field where there is no game to be played.

When canceling site-licensed journals today, the only legally available alternatives are individual subscriptions, pay per view, and self-archived versions of individual papers. This stands in stark contrast with the digital-entertainment universe, where there is a competitive market for providers of personal digital libraries. Services like Apple ITunes, Google Play, Amazon Kindle and Prime, Netflix, Pandora, Spotify, etc. compete on the basis of price, content, usability, convenience, and features. There are many scholarly-communication organizations that could launch analogous services. Within months, Thomson Reuters, EBSCO, publisher alliances, scholarly societies, and even some research libraries could provide a wide selection of options. This will never happen without starving publishers of site-license revenue. Instead of subsidizing publishers, subsidize students and faculty. They are quite capable to choose for themselves what information services they need. After a messy, but short, transition, a competitive market will blossom.

The only thing more terrifying than phasing out a core service is the prospect of outside forces triggering a sudden disruption. Libraries have the choice to disrupt or to be disrupted, to organize their own restructuring or to be restructured by a crisis manager. This is the perfect time to redirect resources away from digital-lending overhead and towards building a scalable, robust, and permanent infrastructure of open scholarly information (refereed papers, technical reports, lab reports, and supporting data). Björn Brembs wants to go even further; he wants libraries to take over all of scholarly communication.

We do not have to wait for Open Access to work its disruptive magic, which may or may not happen at some undetermined time. By forcing the disruption, the rationale for Green Open Access becomes much more straightforward: It creates a permanent public archive of culturally important content that is now controlled by private companies. As a public option to the publishers' walled garden, it may help keep prices in check. That role is much less important, however, when prices are set in a truly competitive market.

Publishers do not think Green Open Access has the power disrupt. They believe they can compensate lower revenue from Gold Open Access by increasing the number of papers they publish. Should site licenses be disrupted anyway, publishers stand ready to compete with libraries.

Publishers are well prepared for any scenario.

Is your library?

Tuesday, August 6, 2013

The Empire Strikes Back

Publishers may soon compete with libraries. The business case for enticing users away from library-managed portals is simple, compelling, and growing. As funding agencies and universities enact Open Access (OA) mandates and publishers transition their journals from the site-license model to the Gold OA model, libraries will cease to be the spigots through which money streams from universities to publishers. In the Gold-OA world, the publishers' core business is developing relationships with scholars, not librarians. For publishers, it makes perfect sense to cater to scholars both as authors and readers.


Current direct-to-scholar portals provided by publishers do not live up to their potential. Each portal is limited to content from just one publisher. Without interoperability, each publisher portal is an island. Only scholars covered by a site license can afford to use them, and those scholars have access to a gateway for all site-licensed content irrespective of publisher: their library web site. In spite of these near-fatal flaws, publishers invest heavily in their direct-to-scholar portals.


These portals are opportunities for future growth. The model is well established: Thomson Reuters' Westlaw is the de-facto standard for legal research in the US, and it is able to command premium pricing for structured public-domain information. It may take a long time for scholarly publishers to duplicate Westlaw's success. Yet, even without access fees, publishers might be able to unlock significant marketing and business-intelligence value from their systems. Knowledge from managing the publishing process combined with usage data from their portals will give publishers unprecedented insight into every aspect of scholars' professional lives in education, research, and development.


For publishers, the transition to Gold OA is rather tricky. They hope to maintain their current level of revenue while replacing the income stream from site licenses with an equivalent income stream from author page charges. This goal, implausible just a few years ago, now seems realistically within their grasp. The outcome remains far from certain, and publishers are hedging their bets by fighting Green OA and lobbying hard for embargo periods. As long as site-license revenue is their main source of revenue, publishers cannot afford to compete with libraries and journal aggregators, their current customers and partners. This calculation will change when Gold OA reaches a certain critical point. This is the context of proposals like CHORUS, an attempt to take over Green OA, and Elsevier's acquisition of Mendeley, a brilliant social-network interface for scholarly content.


Publishers, indexing services, journal aggregators, startups, some nonprofit organizations, and library-system vendors all have expertise to produce compelling post-OA services. However, publishers only need to protect their Gold OA income, and any new revenue streams are just icing on the cake. All others need a reasonable expectation of new revenue to develop new services. This sets the stage for a significant consolidation of the scholarly-communication industry into the hands of publishers.


As soon as the Gold OA shock hits, academic libraries must be ready to engage publishers as competitors. When site licenses disappear, there is no more journal-collection development, and digital lending of journals disappears as a core service. This is a time that requires major strategic decisions from leaders in academia. With its recently released new mission statement, the Harvard Library seems to pave the way: “The Harvard Library advances scholarship and teaching by committing itself to the creation, application, preservation and dissemination of knowledge.” The future of the academic library will be implemented on these pillars. While the revised mission statement necessarily lacks specifics, it is crystal-clear in what it omits: collection development.

Wednesday, June 19, 2013

Chudnov's Mission

Library mission statements are pablum intended to placate everyone and offend no one. It could be different, as I recently found out because of a tweet and a blog from Lorcan Dempsey, which led me to the personal mission statement of Dan Chudnov:


How refreshing!

Chudnov blogged this in 2006, the year in which Time Magazine's person of the year was “You.” Youtube had just exploded into our consciousness. Social networking was hot. This was the end of broadcasting and the beginning of narrowcasting. Time Magazine realized then that new web technologies would center around the individual and his or her personal needs and wants. The world embraced this idea.

Libraries could have aligned with this fundamental shift. But seven years later, libraries remain rooted in the concept of providing services for the average user of a particular community. Chudnov's mission is a radical departure from this model and an ambitious goal. Give to the masses what not so long ago was a rare commodity of only the most privileged: a personal library that archives all the information one has created, has consumed, is consuming, and intends to consume.

In 2013, parts of this vision have been realized. Unfortunately, libraries were largely on the sidelines. A slew of commercial enterprises provide aspects of personal digital libraries, either free of charge or at relatively low cost. Google is organizing the world's information, but its personalized services put the individual front and center. Browsers keep track of the information we have consumed, and they let us bookmark the information we intend to consume. Netflix keeps track of our movies, the Kindle store of our books, iTunes of our music, and Gamefly of our games. We archive our writings, our observations, our pictures, our videos in social networks, cloud-based storage, and blogs. Amazon, Facebook, Flickr, Google, Microsoft, Tumblr, Twitter, Yahoo, and many others would love to provide as many services as possible to each of us as part of their corporate strategy. The current situation is chaotic and messy. Yet, the last thing we should strive for is an orderly, easy, convenient information landscape dominated by a few commercial entities and governments. We should wish for more chaos and more providers competing with one another.

We take it increasingly for granted that we can experience our entertainment on our terms. We want to watch our movies and TV shows when the time is right for us, not when a network decides we should watch it. Unlike DVD rentals, streaming services never sell out no matter how many of our neighbors rent the same video. Yet, when it comes to academic libraries and professional information needs, researchers still accept that their individual requirements are subject to community compromises. Researchers whose information needs are much different from those of average library users are effectively relegated to second-class status.

How can a community-based library adapt? What is its role in an environment increasingly dominated by commercial enterprises? What are the specific steps it can take to help its users develop a personal library? What kind of help do its users need? Should the community library provide alternatives for commercial services? Or, should it merely supplement them? How do these new services fit with institutional traditions and commitments? Should the library help its users regain control of the information they ceded to for-profit companies in a Faustian bargain? If yes, what are the concrete steps that can accomplish this? Should the library help its users regain control of search engines dominated by commercial priorities? If yes, how?

Chudnov's mission statement leaves considerable freedom for interpretation. Like all good mission statements, it sets a direction. It provides a long-distance view. It crystallizes what is important in a time of information overload: focus on the real information needs of individuals. Libraries ignore this at their peril.

Tuesday, May 21, 2013

Turow vs Everyone

According to celebrated author, lawyer, and president of the Author's Guild Scott Turow, the legal and technological erosion of copyright endangers writers. (New York Times, April 7th, 2013) His enemy list is conspiratorial in length and breadth. It includes the Supreme Court, publishers, search engines, the Hathi trust, Google, academics, libraries, and Amazon. Nevertheless, Turow makes compelling arguments that deserve scrutiny.

The Supreme Court decision on re-importation. (Kirtsaeng v. John Wiley & Sons, Inc.)
This 6-3 decision merely reaffirmed the first sale doctrine. It is highly unlikely that this will significantly affect book prices in the US. If it does, any US losses will be offset by price increases in foreign markets. More importantly, the impact will be negligible because paper books will soon be a niche market in the US.

Publishers restrict royalties on e-books.
Publishers who manage the technology shift by making minor business adjustments, such as transferring costs to authors, libraries, and consumers, underestimate the nature of current changes. Traditional publishers built their business when disseminating information was difficult. Once they built their dissemination channels, making money was relatively easy. In our current world, building dissemination channels is easy and cheap. Making money is difficult. Authors may need new partners who built their business in the current environment; there are some in his list of enemies.

Search engines make money of referring users to pirate sites.
Turow has a legitimate moral argument. However, politicizing search engines by censoring search results is as wrong as it is ineffective. Pirate sites also spread through social networks. Cutting off pirate sites from advertizing networks, while effective, is difficult to achieve across international borders and requires unacceptable controls on information exchange. iTunes and its competitors have shown it is possible to compete with pirate sites by providing a convenient user interface, speed, reliability, quality, and protection against computer viruses.

The Hathi trust and Google scanned books without authorization.
Hathi and Google were careless. Authors and publishers were rigid. Experimentation gave way to litigation.

Some academics want to curtail copyright.
Scholarly publishers like Elsevier have profit margins that exceed 30%. Yet, Turow claims that “For many academics today, their own copyrights hold little financial value because scholarly publishing has grown so unprofitable.”

Academics' research is often funded in part by government, and it is always supported by universities. Universities have always been committed to research openness, and they use published research as means for assessment. This is why academics forego royalties when they publish research. The concept of research openness is changing, and many academics are lobbying for the idea that research should be freely available to all. The idea of Open Access was recently embraced by the White House. Open Access applies only to researchers funded by the government and/or employed by participating universities and research labs. It only covers research papers, not books. It does not apply to independent authors. Open Access does not curtail copyright.

Legal academics like Prof. Lawrence Lessig have argued for stricter limits on traditional copyright and alternative copyrights. Pressured by industry lobbyists, Congress has repeatedly increased the length of copyright. If this trend continues, recent works may never enter into the public domain. Legislation must balance authors' intellectual property rights and everyone's (including authors') freedom to produce derivative works, commentaries, parodies, etc.

Amazon patents a scheme to re-sell used e-books.
This patent is a misguided attempt to monetize the human frailty of carrying familiar concepts from old technology senselessly into the new. It is hardly the stuff that made this forward-looking company formidable.

Libraries expand paper lending into digital lending.
Turow demands more money from libraries for digital lending privileges. He is too modest; he should demand their whole budget.

When a paper-based library acquires a book, it permanently increases the value of its collection. This cumulative effect over many years created the world's great collections. When a community spends resources on a digital-lending library, it rents information from publishers and provides a fleeting service for only as long as the licenses last. When the license ends, the information disappears. There is no cumulative effect. That digital-lending library only adds overhead. It will never own or contribute new information. It is an empty shell.

Digital lending is popular with the public. It gives librarians the opportunity to transition gradually into digital space. It continues the libraries' billion-dollar money stream to publishers. Digital lending have a political constituency, but it does not stand up to rational scrutiny. Like Amazon's scheme to resell used e-books, digital-lending programs are desperate attempts to hang on to something that simulates the status quo.

Lending is the wrong paradigm for the digital age. Instead, libraries should use their budgets to accumulate quality open-access information. They should sponsor qualified authors to produce open-access works of interest to the communities they serve. This would give authors a choice. They could either produce their work commercially behind a pay wall, or they could produce library-funded open-access works.

Monday, May 13, 2013

Calculus for Jocks


In 1983, I taught my first course. The teaching assistants called it “Calculus for Poets”, but the incident I remember is about a basketball player. It gave me a glimpse into the mindset of college athletics departments in the US, and it cemented my opinion that competitive sports do not belong on campus.

I had come to the Courant Institute of Mathematical Sciences at New York University a year earlier. Being on a computer-science research fellowship, I did not teach that first year. After switching into the mathematics department, I was a teaching assistant for one year, and I was assigned the Calculus undergraduate course.

Near the end of the first term, the basketball coach came, unannounced, to my (shared) office. He explained that this student athlete, while not a star player, was very important to the team. If the student failed calculus, his Grade Point Average (GPA) would be too low to continue on the team. This would surely hurt the team, and he needed my help.

The coach promised to work with the student in the second term and make up for the failed first term if I would just give him a passing grade now. According to the coach, this wasn't really dishonest. After all, this was calculus not history: When you fail a term of history, you have not learned what happened in a particular era, and you cannot make it up without taking that term over again. However, if you succeed in the second term of calculus after failing the first term, you really have made up your lack of knowledge of the first term, because you could not succeed in the second term without making up the knowledge you missed in the failed first term. So, I should let him pass now in anticipation of second-term success.

I had been in the US long enough to have heard about the power of coaches on campuses, and I told him I would think about it. I immediately walked over to the office of the faculty member overseeing the course. He was not available, but I described the incident to his secretary, and I asked to schedule an appointment. The appointment never materialized. Instead, I received a short note to handle the situation as I saw fit.

I felt badly about being brushed off like that. With the passing of time, it feels even stranger than it did at the time. The Courant Institute was an extremely social place. Graduate students and faculty mingled in the 13th floor lounge every day for coffee breaks and lunch. Courant faculty took their graduate students seriously, and I do not remember any other incident where a faculty member was dismissive of my concerns. Yet, in this case, a faculty member refused to examine a matter that could easily have mushroomed into something serious. In retrospect, I suspect the supervising faculty member may have been affiliated with the School of Arts and Sciences and not the Courant Institute. This would explain why I do not remember his name or face and why I have no recollection of ever meeting him, even for an orientation session.

I had to deal with the coach on my own. After another meeting, I did not give the student a passing grade, but I acquiesced to the next request from the coach. I let the student drop the course. NYU’s rules allowed for late drops, provided they were approved by the instructor. By dropping the course, the student barely satisfied the GPA requirement. Unfortunately for the team and fortunately for the student, his grades did not satisfy the parental requirement. After seeing his son’s results, the student's father stepped in. He pressured his son to stop playing basketball and start studying. The student newspaper lamented this turn of events.

This was a small incident. Many much worse incidents occur every day on every campus. Yet, NYU basketball was, and remains, a small-time program (NCAA Division III). In 2010, the Wall Street Journal thought NYU is neglecting its basketball program and should raise more money for it. A student athlete on this team, no matter how good a player, has absolutely no prospects to play professionally after college. Yet, this coach thought it totally justified to sacrifice the player's education for a few team wins.

I can only imagine what happens at a university that participates in NCAA Division I, where top players have more realistic hopes for a pot of gold at the end of their college career. Does that justify a fraudulent degree? I can only imagine how the coach of such a team deals with lowly teaching assistants, instructors, and assistant professors. His multi-million dollar paycheck, which far exceeds the paycheck of the university’s president, depends on a winning record. Given where his incentives are, how much does that coach care about his or her players and their education?

One of my fellow teaching assistants explained the system at his alma mater, a top research university in NCAA Division I. On the books, athletes took identical courses with identical numbers and the same number of credits. Somehow, they would be enrolled in special sections held at times and in locations that fit the athletes’ schedules. These sections were not listed in the catalog and unavailable to regular students. Were those sections just as hard? Were they graded on an athlete's curve? No one knew, not even the athletes. This old rumor may carry no weight as evidence, but it shows how easy it is to fix things. With millions of dollars on the line, do you trust that everyone will act ethically all the time?

Monday, April 22, 2013

The Sibyl of Cumae


“The seventh was of Cumae, by name Amalthaea, who is termed by some Herophile, or Demophile and they say that she brought nine books to the king Tarquinius Priscus, and asked for them three hundred philippics, and that the king refused so great a price, and derided the madness of the woman; that she, in the sight of the king, burnt three of the books, and demanded the same price for those which were left; that Tarquinius much more considered the woman to be mad; and that when she again, having burnt three other books, persisted in asking the same price, the king was moved, and bought the remaining books for the three hundred pieces of gold: and the number of these books was afterwards increased, after the rebuilding of the Capitol; because they were collected from all cities of Italy and Greece, and especially from those of Erythraea, and were brought to Rome, under the name of whatever Sibyl they were.”
The myth of the Sibyl of Cumae from: The Divine Institutes, by Lactantius (b. ca. A.D. 250), Book I, Chapter VI.

Publishers select, prepare, market, and disseminate information. They developed their selection processes at a time when it was expensive to prepare and disseminate information. As these costs decreased, they could publish more and be less selective. However, the selection process endows information with gravitas, a valuable commodity for marketing. Today's publishers must balance two conflicting interests: increase revenue by publishing as much as possible vs. increase profit margins by selectively publishing high-value information. Scholarly publishers found a way to do both.

Where the Sibyl of Cumae burned some books to increase the value of the remaining books, a scholarly journal rejects a certain number of papers for each paper it publishes. Many of the rejected papers may be interesting, but they do not fit the journal's mission. For the publisher, this is an opportunity to spawn new journals in the wake of its successful journals. Such portfolios of journals are less selective than their individual journals. Of course, if one considers the scholarly publishing industry at the macro level, the notion of selectivity virtually vanishes. Papers are submitted and re-submitted until an outlet is found.

The Sibyls of Scholarly Publishing perform an elaborate dance with pyrotechnic effects that give the illusion they burn papers. In fact, each Sibyl takes in new and rejected papers, packages some of them in a journal, and pretends to burn the rest before handing them off to her sisters. Each Sibyl maximizes the price in her respective corner of the universe. Academia repeatedly acts like King Tarquinius, who thinks the woman mad and pays the price she demands.

It may take years and several turnovers of the editorial board before an established journal that covers a large domain accepts papers in an emerging field. This has created a seemingly insatiable demand for new highly specialized journals. Each successful journal serves its publisher by raising revenue, its editorial-board members by raising their research prestige, and its authors by providing an avenue for dissemination of material without a natural home in existing journals. Many of these journals cater to such a small cadre of specialists that they subvert the single largest scholarly benefit of the refereeing process: a critical reading by someone with a different point of view and background. Even when run with the best of intentions, these narrow journals are echo chambers for group think. Emerging fields need some breathing room, particularly in the early developmental stages, but they should not be immune from outside criticism. Do these journals really serve the cause of good scholarship? Are they worth the super-inflationary cost increases, which they help create?

Open Access may not reduce the cost of scholarly communication as originally hoped. A large-scale conversion to Gold Open Access would shift the costs from universities to governments. Once university administrations no longer feel the budgetary pain and the costs are baked into government budgets, publishers would be free to continue the super-inflationary trajectory. There would not be any market forces that limit the introduction of new journals, the growth of existing journals, or the price charged per paper published. The access problem would be resolved by hiding, compounding, and postponing the cost problem. In the end, the scholarly-communication market would remain as dysfunctional as ever.

Technology has eroded the foundation of the current scholarly-communication system. It assumes that there is a scarcity of dissemination, and it uses that scarcity for the purpose of gatekeeping. In fact, dissemination is abundant and nearly free. The scarcity and associated gatekeeping are marketing illusions.

The reluctance to change is understandable. A scholarly-communication system is a delicate balancing act. It must be fair, but critical. It must discourage poor research, yet be supportive of new ideas, including ideas that challenge established views. Because scholarly communication is tied to research assessment, any changes to the system must gain wide institutional acceptance.

Ultimately, we have little choice but to accept today's reality. Anyone has the power to disseminate any information, regardless of quality. No one has the power to be a gatekeeper. At most, editorial boards have the power of influence in their respective communities; they can highlight important achievements and developments. But even this power to influence may soon be challenged by crowd-sourced quality labels of alternative metrics. (Perhaps not.)

We should be elated about the recent successes of the Open Access movement. We should also recognize that Open Access is not an end point. It is only the first step in the reinvention of scholarly communication.

Tuesday, March 26, 2013

Open Access Politics

The Open Access (OA) movement is gaining some high-level political traction.

The White House Open Access memorandum enacts a national Green OA mandate: Most US funding agencies are directed to set up OA repositories for the research they fund. This Green OA strategy contrasts with the Gold OA strategy proposed by the Finch report in the UK. The latter all but guarantees that established publishers will retain their revenue stream if they switch their business model from site licenses to Author Page Charges (APCs).

The White House memorandum is likely to have the greatest impact. As its consequences ripple through the system, the number and size of Green OA repositories is likely to grow substantially over the next few years. Large-scale validation of altmetrics and the development of new business models may lead to the emergence of new forms of scholarly communication. Green OA archivangelist Stevan Harnad hypothesizes a ten-step scenario of changes.

There are also reasons for concern. As this new phase of the OA movement unfolds on the national political stage, all sides will use their influence and try to re-shape the initial policies to further their respective agendas. The outcome of this political game is far from certain. Worse, the outcome may not be settled for years, as these kind of policies are easily reversed without significant voter backlash.

At its core, OA is about an industry changing because of (not-so-)new technology and its accompanying shift in attitudes and values. In such cases, we expect established players to resist innovation by (ab)using politics and litigation. The entertainment industry lobbied and litigated against VCRs, DVRs, every Internet service ever launched, and now even antennas. In the dysfunctional scholarly-communication market, on the other hand, it is the innovators who resort to politics.

To understand why, suppose university libraries were funded by user-paid memberships and/or service fees. In this scenario, libraries and publishers encountered the same paper-to-digital transition costs. When library prices sky rocketed, students and faculty created underground exchanges of scholarly information. They cancelled their library memberships and/or stopped using their services. The publishers' revenue streams collapsed. Only the most successful journals survived, and even they suffered. Publishing a paper became increasingly difficult because of a lack of journals. This created an opening for experiments in scholarly publishing. This bottoms-up free-market transition would have been chaotic, painful, and forgotten by now.

We do not need to convert our libraries and research institutions into free-market enterprises. We do not need to abandon the fundamental principles on which these institutions are built. On the contrary, we must return to those principles and apply them in a new technological reality. Rebuilding the foundations of institutions is hard under the best of circumstances. When users are shielded from the external incentives/hardships of the free market, it is near impossible to disrupt, and continuity remains an option far beyond reason.

Green OA is an indirect approach to achieve fundamental change. It asks scholars to accept a little inconvenience for the sake of the larger principle. It asks them to deposit their papers into OA repositories and provide free access to publicly-funded research. It is hoped that this will gradually change the journal ecosystem and build pressure to innovate. It took dedicated developers, activists, advocates, and academic leaders over twenty years to promote this modest goal and create a movement that, finally, seems to have achieved critical mass. A growing number of universities have enacted OA mandates. These pioneers led the way, but only a government mandate can achieve the scale required to change the market. Enter politics.

Scholars, the creators and consumers of this market, should be able to dictate their terms. Yet, they are beholden to the establishment journals (and their publishers), which are the fountain of academic prestige. The SCOAP3 initiative for High Energy Physics journals shows how scholars are willing to go to unprecedented lengths to protect their journals.

Market-dominating scholarly publishers are paralyzed. They cannot abandon their only source of significant revenue (site licenses) on a hunch that another business model may work out better in the long term. In the mean time, they promote an impossible-to-defend hybrid Gold OA scheme, and they miss an opportunity to create value from author/reader networks (an opportunity recognized by upstart innovators). This business paralysis translates into a lobbying effort to protect the status quo for as long as feasible.

Academic libraries, which enthusiastically supported and developed Green OA, now enter this political arena in a weak position. The White House memorandum all but ignores them. Before complacency sets in, there is precious little time to argue a compelling case for independent institutional or individual repositories preserved in a long-term archive. After all, government-run repositories may disappear at any time for a variety of reasons.

The Gold OA approach of the Finch report is conceptually simpler. Neither scholars nor publishers are inconvenienced, let alone disrupted. It underwrites the survival of favored journals as Gold OA entities. It preempts real innovation. Without a mechanism in place to limit APCs, it's good to be a scholarly publisher in the UK. For now.

Wednesday, February 13, 2013

Crowd Sourcing Gun Control

Background checks deter only those with prior offenses. For more comprehensive gun-violence prevention, we need systems that help law enforcement focus on high-risk individuals before their first offense. Such prevention systems must operate within important legal and political constraints: they must be compatible with the second amendment, they must be relatively effortless for the vast majority of low-risk gun buyers/owners, and they must respect the privacy of all citizens.


Graph theory makes it possible to extract strong information from a large set of weak relationships. It is the foundation of companies like Google and Facebook. It is the basis for alternative metrics that assess the quality of scholarly papers. Graph theory also applies to gun control. As individuals, we absorb all kinds of information about others and decide to like or dislike, befriend or avoid, trust or distrust them. These individual assessments are rather unreliable, but they become a powerful resource when connected with one another in a graph. For gun control, imagine a system in which gun owners recommend one another. This creates a graph on which to unleash the power of graph theory.

Consider a concrete example (no graph-theory knowledge required):

  1. Every gun buyer/owner registers three Qualified Recommenders (QRs). Without assuming any legal liability, QRs state that the gun buyer/owner is not presently a danger to himself/herself or to society and that they will withdraw as QR if their opinion changes or if they should lose contact with the gun owner.
  2. Most legal gun owners may act as a QR. However, law enforcement may revoke anyone's QR status at any time for any reason. Such action prevents this person from submitting new recommendations, and it voids his/her existing recommendations. Revoking the QR status carries no other consequences. One's own QR status does not affect one's right to own guns.
  3. A gun buyer who is unable to obtain three QRs is subject to a background investigation that may include interviews with family, friends, and neighbors.
  4. A gun owner may lose recommendations because (1) QRs may withdraw their recommendation, (2) QRs may die, or (3) QRs may lose their QR status. A gun owner who does not maintain three QRs is subject to increased scrutiny.
Maintaining three QRs is relatively effortless for low-risk gun owners, who tend to have strong ties to their communities. If they should become high risk (through mental illness, for example), their QRs are likely to withdraw their recommendations, thereby creating an early warning.

The impact on high-risk gun owners is more substantial. Very few law-abiding gun owners would be willing to sponsor someone they consider dangerous because of mental health, behavioral issues, or criminal connections. Convicted criminals cannot be QRs, because they cannot be legal gun owners. This reduces the pool of available QRs for their non-convicted associates. Law enforcement can reduce that pool even further: when they arrest someone, they could revoke the QR status of the arrestee and of his/her QRs (under the theory that these QRs are unreliable).

Consider a few scenarios.

Scenario 1. Joe is a gun owner. His QRs are his wife (Mary) and two friends (John and Pete).

Scenario 1a. Joe develops a mental-health problem and is in therapy. Mary convinces Joe to dispose of the guns. Mary withdraws her recommendation and, perhaps, approaches John and Pete to withdraw theirs. There are no legal proceedings of any kind.

Scenario 1b. Joe develops a mental-health problem, refuses therapy, and develops violent tendencies. The QRs and the police develop a plan to withdraw their recommendations and, simultaneously, to impound Joe's weapons. After the emergency intervention, Joe's fitness to own guns is evaluated.

Scenario 1c. Joe is arrested under suspicion of committing a violent crime.

Upon his arrest, Joe's QR status is revoked immediately. Joe's gun rights depend on the resolution of the criminal case against him. Mary, John, and Pete are not legally liable for failing to intervene in time, but their QR status is revoked.

Mary, John, and Pete remain legal gun owners. If they acted as each other's QRs, they have to find replacement QRs. For most gun owners, these circumstances are an extremely rare occurrence, and replacing QRs is a low hurdle. For criminal gangs, every arrest eliminates QRs from their environment and increases their difficulty of replacing QRs.

Scenario 2. Bill sells his recommendations. As soon as he is identified as a shady operator, the police subjects his network to heightened scrutiny. With the first arrest in Bill's network, he loses his QR status, and his recommendations are voided. As Bill's clients find replacement QRs, they leave useful network trails.

Scenario 3. Mary, a law-abiding citizen living in a gang-infested neighborhood, wants a gun for protection. Some of her relatives and friends have arrest records, and she cannot find three QRs. She is a high-risk gun buyer because of the circumstances of her life: her gun is at much higher risk to be stolen and/or misused. She is able to buy a gun after passing an in-depth background investigation.

Scenario 4. Bob is a criminal and does not care about gun laws. Because his guns are illegal, he faces a much harsher sentence if arrested and convicted.

Bob's guns are most likely stolen or obtained from sources that claim their guns were stolen or lost. If one of his guns is traced back to Mary, her QR status is revoked immediately. If Mary is an innocent bystander, losing her QR status is no big deal. If she is a member of a network that traffics guns illegally, this incident reduces the pool of available QRs.

Other Scenarios. The QR system implements a form of social control that occurs organically in small communities where everyone knows their neighbors (a rarity nowadays). The process of finding QRs provides opportunities for a community to intervene in a non-confrontational manner and to help prospective gun buyers make a realistic assessment of their own risk profile.

The QR system shows that gun-violence prevention does not require storing detailed personal information. It shows that an alert and involved community does not need to become a society of snoops that report every eccentricity to the police. Yet, the QR system is just one example. Graph theory provides a lot of flexibility to respond to legal and political concerns.

I call on the graph-theory community to contribute to the gun-control effort. It is up to you. Criticize. Amend. Propose. Do something. Anything. Save lives.

Monday, January 14, 2013

MOOCs Teach OA a Lesson

Just four years ago, Massive Open Online Courses (MOOCs) were tentative experiments promoted by a handful of professors. (Wikipedia, New York Times, Chronicle of Higher Education) Today, universities across the world are rushing in, and millions of students are enrolling. Contrast this with the Open Access movement (OA). More than twenty years after the introduction of the hep-th database (which became arXiv), OA remains a struggle. There have been significant OA advances, but universal open access to the scholarly literature remains a distant promise, probably requiring many more years.

Why did OA never reach the kind of momentum MOOCs seem to have?

Because successful MOOCs serve many thousands of students, their per-student costs are extremely low when compared to traditional teaching. Yet, the cost of producing a series of high-quality large-scale interactive multimedia events is significant. Compared with MOOCs, the start-up cost of OA is almost negligible. After an institutional repository is set up, the only barrier to OA is a few key strokes per scholarly paper.

Why were academic leaders so concerned about the minimal costs of OA? Why are they not concerned about the far more significant costs of MOOCs?

MOOCs have the potential of disrupting thousands of teaching positions. MOOCs are a threat to admissions offices and a system of university reputations based on rejection rates. On the other hand, universal OA would primarily disrupt libraries, publishers, and their middlemen, not academics. Yet, academic leaders are enamored with MOOCs, and they treat OA like a chore for which there is always some excuse to postpone. If MOOCs really prove to be as disruptive as hoped or feared, they figure it is preferable to be on the side of the disrupters.

Why do academic leaders not make the same calculation with respect to OA? Why do they fear the potential of OA-caused disruption? Why do they embrace the potential of MOOCs-caused disruption?

In my search for answers, I arrived at four tentative conjectures.

Conjecture 1. MOOCs are in their infancy. The wave of initial excitement will pass, and the hard MOOC work lies ahead. OA is further along in its evolution. Passed its own wave of initial excitement, OA is now in the slow process of building its infrastructure. Some form of OA will soon emerge as the inevitable path.

This conjecture provides cover to continue on the current path.

Conjecture 2. With MOOCs, first movers have a clear advantage. They have the most time to develop the know-how for producing successful MOOCs. With little competition, they can afford to make mistakes and learn from them. With OA, first movers provide a service to those on the sidelines and get little in return. (This perverse incentive explains, in part, the need for OA mandates.)

This begs for initiatives that reward scholars who make their works OA.

Conjecture 3. With MOOCs, faculty control their work, and they do what they do best: they innovate an area in which they are experts. OA feels like an external imposition. To add insult to injury, some repository managers have turned simple light-weight OA repositories into a bureaucratic mess with useless policies that turn faculty off. And it is not just repository policies. Scholars are increasingly awash in conflicting and confusing OA-related policies from funding agencies, publishers, universities, and libraries. Discussions about OA mandates do not help the cause either. It is irrelevant that OA mandates require very little effort when enacted; the discussion itself is a turn-off.

This is an argument to reduce the heavy-handedness of current OA approaches. Eliminate the bureaucracy, and replace institutional repositories with self-managed individual repositories. These may not eliminate all institutional policies, but they give scholars a greater degree of control and flexibility. Individual repositories are also portable when scholars move from one institution to the next. There are at least two options that make it easy for scholars to manage their own individual repository: academia.edu and myopenarchive.org. ORCID, the recently launched initiative to manage the identities of scholars, could also evolve into a system of individual repositories.

Conjecture 4. OA is not sufficiently disruptive. Hoping to minimize resistance to OA, OA advocates tend to underemphasize the disruptiveness of OA. Gold and Green OA leave the scholarly-communication system essentially intact. When presented in a minimalist frame, they are minor tweaks that provide open access, shift costs, and bend the cost curve. Such modest, even boring, goals do not capture the imagination of the most effective advocates for change, advocates who have the ears of and who are courted by academic leaders: venture capitalists. This is a constituency that seeks out projects that change the world.

This conjecture is an argument to pursue disruptive OA. What if OA completely erased the cost of all scholarly communication? That would reduce the cost of education and/or research by at least as much as some of the most disruptive MOOC scenarios.

PeerJ falls in the category of disruptive OA. PeerJ is a new model for open-access journals with peer review. PeerJ charges authors a one-time $99 membership fee and eliminates the per-paper publication charges of Gold OA journals.

One could, of course, dispose of journals altogether. Combine individual repositories with open evaluation and alternative metrics. The field of altmetrics has developed various impact measures based on usage statistics of individual papers. This fine-grained analysis is far superior to the rather coarse and often misleading Journal Impact Factor. To succeed, open evaluation and altmetrics must win over the entrenched interests that control academia's prestige machine.

Perhaps, none of these conjectures fully explain differing attitudes towards MOOCs and OA. Perhaps, it is a combination of all four. Perhaps, there are other factors at play. If so, what are they, and how should those factors influence our approach to OA?