This article was submitted 11 Dec 1997 and published without substantive change in American Entomologist 44: 135-138 (Fall issue, 1998).
The future of scientific journals: free access or pay per view?
Thomas J. Walker
Journal publishers are preparing for a revolution in how they deliver their product. For more than 300 years, they have distributed the articles they publish in bound batches (issues) to subscribers and as multiple copies of the same article (reprints) to authors. Researchers then acquire copies of articles they are particularly interested in by photocopying (in person or by proxy) from issues or by soliciting reprints from authors. Recently many publishers, including ESA, have begun testing a third means of deliverynamely, the Internets World Wide Web (WWW). Eventually publishers will quit the first two means in favor of the third, but how they will then recover the costs of publishing is unresolved. This article describes two competing views of cost recovery and concludes that scientific societies, such as ESA, can lead the way to a publishing system that gives free, convenient access to all primary scientific literature.
How can traditional journal articles be delivered via the Web?
Web delivery of journal articles is noteworthily simple. In the past decade, nearly all publishers have switched from cut-and-paste to electronic page layouts that they then use to make the plates for their printing presses. The digital files that produce the plates can also be used to produce a Portable Document Format (PDF) file for each article. The process is as easy as sending word-processor files to a printer, allowing publishers to make PDF files of journal articles for ca. $1 per page.
Once the PDF file is made, it need only be put on a WWW server to make the article globally accessible. WWW server space costs less than a dime per article, and if the publisher chooses to put entire issues online with free access, research libraries will provide the space and perpetual care for free. The Florida Center for Library Automation has done this for Florida Entomologist since 1994 (http://www.fcla.edu/FlaEnt/).
If you want to view a PDF file, you must install the free Acrobat Reader on your computer. Your Web browser can then automatically display the article when you click on the link to its PDF file. From this display, you can browse, search, read, or print the article. Printing produces the equivalent of an excellent photocopy.
A PDF file delivers a copy of the original article, not a reformatted version of it. If this surprises you, it is because you are used to Web documents posted in HyperText Markup Language (HTML). HTML is the language of the Web, and access to PDF files is ordinarily via HTML documents. At significant expense, publishers sometimes translate traditional journal articles into HTML. This allows them to optimize on-screen viewing and interactivity. However, these added features are useless when the articles are printed, so the publishers post PDF files as well. (For a more detailed comparison of PDF and HTML, see http://tjwalker.ifas.ufl.edu/pdfvhtml.htm.)
Why will central printing cease?
Central printing of issues and reprints will cease because, compared to traditional distribution of articles, Internet distribution is more convenient for users and more economical for all parties.
Considering only convenience, rank these three means of getting a copy of an article in a journal that neither you nor any near-by colleague subscribes to:
A different and very important way that Internet delivery makes access to journal articles more convenient is full-text searching. An index file compiled from the text of posted articles allows retrieval by any word or phrase. To illustrate, all volumes of three traditional journals published by Ecological Society of America are online from their beginning dates through 1994 (http://www.jstor.org/). An online search of these 142 journal volumes for Aphytis retrieves, in about 10 s, the 35 articles that refer to that genus of parasitoid wasps. Similarly, a search for competitive displacement retrieves 99 articles. Searching for Aphytis and competitive displacement finds the 4 articles that contain both. A search for Morris and Miller 1954 finds the 8 articles that cite those authors classic article on ecological life tables.
In economy, Internet delivery overwhelms its rivals. Whereas Web distribution is nearly free, distribution via issues and reprints entails large costs to three groups:
The total cost of traditional distribution of one years issues of an average journal surely exceeds $200,000, yet the total cost of putting and keeping one years PDF files on a research librarys Web server for 30 years is less than $1,000 (based on estimates by J. F. Corey, Director, Florida Center of Library Automation, personal communication, 1997; details at http://tjwalker.ifas.ufl.edu/access$$.htm). Web distribution of journal articles costs less than one-half of 1% as much as traditional distribution!
If Web distribution is more convenient than traditional distribution and if switching to Web distribution saves more than 99.5% of the very large costs of the present distribution system, why havent the affected parties done more to effect the change? Each party has its own reasons. Most researchers havent yet realized that they can have the same familiar product much more conveniently. Although librarians are the ones most concerned about the costs of the present system, they have the least leverage to change it. And, most importantly, publishers, including many scientific societies, have grown accustomed to large profits that may be hard to maintain if they end central printing and switch to Web distribution. Nonetheless, publishers do foresee the switch and are trying to envision how they will recover editing and composing costs and maintain profits.
Two visions of cost recovery with e-distributed journals
What publishers see as they look to the future is either a cost-recovery system that retains research libraries as the largest contributors and requires some users to pay per view, or a system that is so radically different that institutions do not subscribe and access is free.
In this vision all users get free access to journal articles, with libraries buying no subscriptions but facilitating Web access by contributing space on their servers. The revenue comes from page charges paid by authors or their institutions. Before you say No way!, consider these two features of a Web-based free-access system. First and most important, the inclusive costs of such a publication system are remarkably low compared to the (very high) inclusive costs of the current system. Second, costs of the present publication system are paid principally by authors and their institutions, as page charges, reprint costs, library subscriptions, and library facilities and services. Thus authors and their institutions, which already pay for publication, will continue to do so, but the total paid will be very much less, even if no savings are made in editing and composing.
Status quo vision
To understand the status quo vision, you must understand the status quo, and it may be different from what you think. For at least the first 300 years of journal publication, scientific societies were the dominant publishers. However, in the past 40 years, commercial publishers have expanded their very minor roles to become the dominant players. They have done this by aggressively starting new journals as new areas of research have sprung up or blossomed and by eschewing page charges while generating nearly all of their revenues from library subscriptions. These subscriptions, priced to maximize profits, have resulted in exponentially growing income for some publishers, allowing them to buy or start yet more journals. The archetypal example of this is Reed Elsevier, which now publishes more than 1200 journals with sales of ca. $880 million, an estimated pretax margin of nearly 40%, and a fund of $1.7 billion to acquire other companies (Hayes 1995, Frank 1997). Libraries pay $103 per issue for Reed Elseviers Journal of Insect Physiology. (In contrast, libraries pay $32 per issue for ESAs JEE.) The subscription prices that have led to the financial success of for-profit publishers have caused a crisis for research libraries (Cummings et al. 1992). Librarians buy a smaller and smaller portion of what their patrons need, as they seek to cope with mounting subscription costs by canceling subscriptions, seldom starting new ones, and buying fewer books. The not-for-profit publishers (mostly scientific societies) have priced their institutional subscriptions more modestly and supplemented their journal income with page charges and membership dues that include compulsory subscriptions. Nonetheless, societies often depend on profits from their journals to pay for other member services, and the income from their journals is heavily dependent on institutional subscriptions. For example, in 1996, ESA made 53% of its $2.3 million net income from its journals, and more than half of its journal income was from institutional subscriptions. In short, the status quo requires a large revenue stream from libraries.
To keep this stream flowing, the status quo vision of cost recovery depends on Subscriptions (bought by individuals, often as part of dues to a society), Site Licenses (bought by libraries for their patrons), and Pay Per View, for those who neither subscribe nor belong to an institution that has a site license. As is evident, S/SL/PPV differs from the present system mainly by requiring unaffiliated non-subscribers to pay for access to individual articlesi.e., in its pay-per-view feature. All who access the articles must pay or belong to an institution that pays or the system fails, and that is its Achilles heel. With S/SL/PPV, authors must not distribute electronic versions of their articles, although e-publication and the Internet make it easy for them to do so and difficult or impossible for publishers to prevent.
Why access will be free
Authors/researchers want free access. Researchers get no royalties from the journal articles they publish. Moreover, they pay high prices for reprints (or make cheaper photocopies) and take the time and trouble to distribute them to whoever expresses an interest in having one. In addition, when they choose a society-published journal for their articles, they (or their grants or institutions) generally pay substantial page charges. Researchers have always wanted all interested persons to have convenient access to their published results and have been willing to pay for it. Until now, the best they could do was to publish in widely circulated journals and to buy and mail paper copies of their articles. The advent of Web distribution offers a near-perfect solution that is affordable.
Institutions will want it. Presently, research institutions provide access to the journal literature through libraries that spend massively, though inadequately, to acquire and maintain journals and to make the journals available many hours a week. For articles published in journals they cannot afford to buy and shelve, they provide expensive interlibrary loan or document delivery services. When administrators see that they can cut their institutional budgets for access services as much as 50%, they will become strong advocates for a free access system.
Societies will provide it. Scientific societies will provide free access because their authors and members will vote to have it and to pay for it. As described in the last section of this article, societies can substantially increase their publication profits during the transition to all-electronic access by selling electronic reprints. Furthermore, by providing free access to their authors, societies can gain an advantage in their competition with commercial publishers for the best manuscripts.
Commercial publishers will be forced to follow. As societies increasingly give free, unlimited Web access to their journals, commercial publishers will see their share of submitted manuscripts decline. At the same time they will have problems in preventing redistribution of articles posted on their restricted-access Web sites. For these reasons, commercial publishers will eventually switch from S/SL/PPV to free access. To cover their costs, they will initiate page charges, thereby giving up an important advantage they have long had in competing with societies for submissions.
Concerns about archiving. As long as Web distribution of articles is an adjunct to traditional distribution, concerns about archiving the digital files are unimportant. Although it would be lamentable to lose the files, they could be recreated from traditional paper copies. The JSTOR project is doing just that as it digitizes and places on the Web entire back runs of long-published journals, such as Ecology and Ecological Monographs (http://www.jstor.org/). However, when central printing ceases and master copies of articles are digital rather than ink-on-paper, concern will be justified. A prestigious 21-member task force recently reported on the problems of digital archiving (TFADI 1996). Among other things, they concluded that digital technology is changing so rapidly that digital information will have to be transferred periodically to new media that use new hardware and software. Such transfers are likely to be much easier and cheaper than recent efforts to move ink-on-acid-paper information to microfilm or digital formats. As archiving paper-published journals wanes, research libraries should be able to apply some of their released resources to archiving digitally published journals.
The end of PDF. Once central printing ceases, PDFs ability to distribute faithful copies of paper-published articles becomes irrelevant, and articles are likely to be formatted only in HTML, which by then should be much improved.
New opportunities. Once journal publication is entirely electronic, many improvements are easily accomplished. For example, articles can be published as soon as they are refereed, edited, and composed, because they no longer need to be bundled into issues; color figures can be incorporated as cheaply as black-and-white ones; space will no longer be dear, allowing full data sets and supplementary information to be appended; audio and video clips and interactive models can be incorporated; commentaries and responses to commentaries can be appended to the original article; and links to subsequent articles that refer to the present one can be added. (To be competitive in grasping these new opportunities, ESA is developing a new, all-electronic journal, Entomological Techniques.) [ESA should soon post information on its plans for this new journal, and that URL should be added here.]
Perfection within reach? Perfect access to primary scientific literature (however its content may evolve) must be fast, convenient and free; articles must be retrievable by any aspect of their content; and all previously published articles, as well as all currently published ones, must be included. Web distribution is fast and can be convenient and free. Full-text indexing allows retrieval by any combination of character strings connected by any of a toolkit of logical operators. At modest cost, old articles can be digitized and made full-text searchable (described below). Something close to perfection is within reach.
How ESA can lead the way
By taking the three steps described below, ESA can help lead the way to near-perfect access to all journal literature. In fact, ESA has taken the first step and its relevant committees have endorsed the second. ESA should take these three steps because they improve its services to its authors and other members and because they will prompt other societies and commercial publishers to move toward free and convenient access, thereby benefiting all researchers, their sponsors, and the public.
Step #1. Sell and promote electronic reprints.
Electronic reprints are Web-accessible PDF files of traditionally published articles. Authors who buy e-reprints make copies of their articles easily available to anyone with access to the Internet. E-reprints do not have to be stored or mailed by authors; users may acquire them by going to online tables of contents, clicking on the desired articles, and printing one or more copies. This is easier, cheaper, and faster for both authors and recipients than is the traditional reprint system. It can also be much more profitable for publishers, because producing an infinite supply of e-reprints costs only about a fourth as much as producing 100 traditional reprints (http://tjwalker.ifas.ufl.edu/eprints.htm).
As originally proposed (Walker 1996) and as implemented by Florida Entomologist (http://www.fcla.edu/FlaEnt/), authors paying for electronic reprints buy the permanent posting of their articles with no limit to the number of downloads. ESAs Governing Board voted to offer authors e-reprints in December 1996, but, in June 1997, in deciding how to implement such sales, they specified that authors must pay per 100 downloads and not have the option of buying an infinite supply. Thus ESA authors can now buy 100 downloads of the PDF files of their articles for the price of 100 paper reprints or 200 downloads for the price of 200 paper reprints, etc. The Governing Board should be encouraged to reconsider this decision and allow authors to buy infinite downloads for the price of 100 reprints. This would make e-reprints more attractive to authors and easier for ESA to manage. It should increase reprint profits, because more authors would buy e-reprints and ESA profits nearly twice as much when it sells an infinite supply of e-reprints as when it sells 100 paper reprints (http://tjwalker.ifas.ufl.edu/eprints.htm).
Step #2. Post unlimited e-reprints for all articles two years after publication.
The cost of putting articles on the Web is so low that only about 20% of the extra reprint profits anticipated from e-reprints would be required to make PDF files of all articles and to put them on the Web two years later (http://tjwalker.ifas.ufl.edu/esajstor.htm). A delay of two years should be sufficient to forestall loss of subscription income, which is needed so long as traditional distribution is continued. A longer delay should not be imposed without evidence to the contrary. At the ESAs last annual meeting, both the Electronic Publication Implementation Committee and the Publication Committee recommended that the Governing Board consider this step. [This paragraph may need updating after the Governing Board considers the step at its June meeting.]
Since 1994, the Florida Entomological Society (FES) has used income from page charges to post unlimited e-reprints of all Florida Entomologist articles. Thus far it has posted them a few weeks after issues are mailed and has lost few institutional subscriptions. However, if libraries start canceling their subscriptions because their patrons can access all current articles on the Web, FES plans to charge authors for immediate posting of their e-reprints. For authors who choose not to pay, e-reprints will be posted after one or two years. So long as some authors choose not to pay the surcharge, libraries that want to provide their patrons timely access to all Florida Entomologist articles will have to continue to subscribe. Should all authors choose to pay for immediate posting, the sum of the surcharges will approximately equal what is now earned through institutional subscriptions (http://tjwalker.ifas.ufl.edu/impact.html).
Step #3. Begin posting back files on the Web.
Near-perfect access to ESA journals requires that all back issues be on line. To do this would cost more than $100,000 (the Annals and JEE each have 90 back volumes). However, for less than ESA has earmarked for starting its new all-electronic journal, it could put ten years of back issues of its four principal journals on the Webe.g., 1987-1996, ca. 53,000 pages. Once the advantages of fully searchable, freely accessible online archives have been established, ESAs Entomological Foundation should be able to attract donations to complete the project.
That posting all back files is feasible is illustrated by the experiences of the Florida Entomological Society. After entomological researchers and authors had embraced having all Florida Entomologist articles published since 1994 freely accessible on the Web, the FES Executive Committee considered a proposal to spend $12,000 from its reserves to put the 20,000 pages of Florida Entomologist not yet e-published (1917 to 1993) on the Web, with free access and full-text searchable. Within two months of the proposal, $4,000 was pledged from each of three sources and no reserve funds had to be committed. [The details and status of this project are at http://tjwalker.ifas.ufl.edu/backissu.html, but I expect it will be completed by July 1 [requiring this paragraph to be updated prior to publication.]
A final word
It is only a matter of time before Web distribution of journal articles is universal. Because such distribution is so economical, free access to the primary research literature is affordable. At the same time, institutional costs of providing access can be drastically reduced. Scientific societies can and should lead the way toward free accessand their members should make sure they do.
Note: This commentary is to a large extent a summary and an update of a much longer Web-published article that is illustrated and has abundant hyperlinks to relevant literature and examples: The electronic future of scientific journals, at http://tjwalker.ifas.ufl.edu/aedraft.htm.
Thomas J. Walker, an entomologist at University of Florida, began teaching a course in information techniques for incoming graduate students in 1965. When computers migrated to desktops in the early 80s, he decided to keep the course current. He is still trying. In 1994, he helped the Florida Entomological Society put its Florida Entomologist online (http://www.fcla.edu/FlaEnt/). His long-term interest in singing Orthoptera is finding an outlet in a Handbook of Katydids and Crickets.