EPUBs are an experimental feature, and may not work in all readers.

Abstract

Ever-increasing journal subscription prices continue to stress the academic community in both economic and intellectual terms. Deploying newly available tools and approaches to article production in a collaborative manner offer dramatic reductions in cost, up to two orders of magnitude. This level of cost saving practically mandates a re-thinking of the entire business model for scholarly publishing. Open access—providing content globally at no cost—should be both economically viable and sustainable.

Introduction

In the worlds of scholarship and scholarly publishing, technology and the Internet have served as disruptive forces, both negative and positive. On the downside, the ease with which both students and faculty can commit plagiarism is of great concern. On the upside, the ease with which lectures can be prepared, course materials made available, and new knowledge shared globally is a strong positive force in education. Extending this power to scholarly publishing inevitably leads to a frank discussion about the real costs that have long burdened the ways that scholars communicate with each other.

The discussion boards and listservs of the scholarly publishing and library worlds abound with frequent, repetitive, ping-pong-like debates about subscription costs, mandatory subscription bundling, alternative subscription pricing models, cancellation rates, open archiving, and definitions of copyright or fair use. The issues raised are the symptoms of an enterprise that appears to have numerous maladies. It costs more than it should, it is held hostage by commercial interests, it is dysfunctional in its efforts to disseminate information, and it is reluctant to adopt new approaches, methods, and technologies.

Many in the academic world consider the mechanism to be either broken, or at the least not functioning optimally, and in need of repair. Outlining and analyzing the steps that lead to the production and dissemination of scholarly knowledge can be both startling and revealing. No longer is the publishing enterprise held hostage by the need for a printing press and ink. Yet, rarely do the participants in this debate step back from the heat of the argument for an overview of the processes to:

  • diagnose the problems
  • propose rational solutions, and
  • manage effective and successful implementation of those solutions.

“Does the publishing process need to cost what it does?”

Does the publishing process need to cost what it does? If not, why has the world not adopted newer, more cost-efficient approaches? This raises the greater issue: who or what is impeding progress? How can we overcome these hurdles? Only if we do can we determine whether the costs are fair, intellectual integrity (the peer review process) maintained, access enhanced, and posterity (archiving) and the future (new knowledge) served. Even among the over 3,300 scholarly journals listed in the Directory of Open Access Journals [DOAJ], many are still burdened by traditional and unnecessarily high production costs.

The publishing process seems too self-evident, rooted in history, and logical to require a detailed description. Yet that is precisely what we need to do, in light of new technologies that cover tasks from manuscript preparation throughout the editorial review process and then to display, dissemination, and preservation. Re-examining the steps involved reveals the relative efficiencies or inefficiencies of the system, allowing us to determine what is broken or antiquated and how it might be fixed or updated.

The cost of intellectual goods

Scholarly communication begins with the preparation of a manuscript, either research-oriented or didactic. Sharing of knowledge within the academic community furthers both intellectual inquiry and careers (retaining a position or advancement). In nearly all cases, faculty or industry salaries and grants of some kind have covered those information development costs—the scholarship or research itself, any associated data analysis, essential illustrations, and other supporting materials. While the work may have begun in handwritten form, a safe assumption is that it will have been converted into a word-processed document - an electronic format of some kind—in preparation for submission. [Sosteric] The cost at this point in the publishing process, the cost of goods, is zero. In no other industry does a community produce material of great value, give it away to others outside of their community, and then allow them to sell it back to the community that created it - and at a high price. [Bergstrom]

Peer review costs

While there are rare exceptions, the rule has long been that peer review is a contributed and collaborative effort within a scholarly community. Each reviewer dedicates an amount of time (again covered by salary or grant, or simply donated in available free time) to evaluate submissions in that individual’s field. It is a socialistic enterprise of sorts, benefiting the field and colleagues: from each according to his ability, to each according to his need. [Odlyzko1] The cost to the scholarly publication is again customarily zero.

There are, however, some minimal expenses involved in preparing the manuscript for distribution to section editors and reviewers. Recall that the originally submitted manuscript may have arrived in any of several electronic formats. Reviewers need a common cross-platform standard for viewing, such as the Portable Document Format (PDF), an open file format developed and maintained by Adobe. Recent versions of Adobe Acrobat allow file permissions to be set to allow reviewers to comment directly within a file using the free Adobe Reader software. Other document formats (Tex or LaTex) have common currency in such specialized communities as physics or mathematics and can substitute for PDFs (without conversion).

Converting documents from a variety of formats into PDF is neither difficult nor costly, with software tools ranging in cost from free (the open source Open Office program as well as dedicated PDF converters, with some limitations) [Google1] [Wikipedia] to under a hundred dollars (a single academic software copy of Adobe Acrobat Professional). The conversion process to PDF takes a minute or two in minimally trained hands and should cost no more than a dollar or two in time and materials. The resulting PDF can be sent to the reviewer as an e-mail attachment or made available for direct download (with e-mail notification) from an editorial management system. In the exceptional case that reviewers require paper documents, the cost of postage plus the secretarial function cost add another dollar or two to the publication cost.

Editorial management costs: the workforce

Editorial oversight for incoming manuscripts involves both processes and people. Process control requires managing the flow of manuscripts among authors, editors, and reviewers. The people responsible for editorial progress are the editors, associate or section editors, managing editors, and editorial assistants who have traditionally controlled the flow of paper and now electronic files. At the end of the process stream are the copy editors, layout editors, and the staff who format and tag content according to pre-set standards, for display, indexing, and archiving.

The editorial staff contributes both time and effort to the production of a finished article. That effort encompasses oversight of the editorial workflow process, editorial assignments for peer review, editorial decision-making, and copy editing for intellectual content, style, and overall appearance.

“For all but the largest-circulation (and often most lucrative) journals, senior editorial staff have long donated their time and effort to the cause as an integral part of their academic positions.”

For all but the largest-circulation (and often most lucrative) journals, senior editorial staff have long donated their time and effort to the cause as an integral part of their academic positions. In select cases, societies or publishers have contributed a stipend of varying amount to cover editorial, managerial, or secretarial time. Economic pressures and dramatic cost-cutting by for-profit publishers (as they raise subscription costs) have caused these support funds to evaporate. Fortunately, the amount of time required for workflow oversight has diminished with the newer and cleverer publishing software now in use.

An editor of a low-volume journal can readily handle submissions in the 100 per year range with a 25% acceptance rate and spend no more than an hour a week on administrative functions in an automated manuscript management system. Experience of numerous journals using the Open Journal Systems software implemented on the Scholarly Exchange site[1] demonstrates that functions formerly handled in a paper-based world by editorial assistants or managing editors can be managed efficiently by a single editor or editorial assistant. A dozen or more publications could be produced collaboratively if they were to share administrative support. [SE][Crow]

Editorial decision-making and article-editing functions require considerably more time. Like the authors’ and reviewers’ time, these are customarily subsumed in the salaries or grants that faculty receive. Again, some society and for-profit journals continue to provide limited financial support for these tasks.

A working platform

Editorial workflow software to manage the movement of electronic documents is now plentiful. It permits authors or editors to submit electronic documents directly to the flow-control system and at the same time enter the necessary descriptive data (metadata) to uniquely identify the work. Submitting authors can provide guidance to the editorial office (keywords and other information) that facilitates the review process and contributes to effective indexing and archiving. Such systems contain the submissions, reviews, editorial decisions, and correspondence—in a word, everything of and about the manuscript—in one place. They record the pathways that articles have followed and often the related correspondence. Even for journals that still use paper in some aspect of the process, these software packages prove helpful, if only as clever electronic tickler files and content hoppers.

These software programs far exceed in complexity the previous generation of systems that merely tracked the physical presence of paper-based manuscripts. Despite startling differences in acquisition cost that run from free (the open-source Open Journal Systems and DPubs, among others [OJS] [DPubs]) to tens or hundreds of thousands of dollars (and even into the tens of millions)[2] for the proprietary commercial and not-for-profit publishing systems [ScholarOne] [EditorialManager] [HighWire] [NYTimes] [Atypon], the functionalities are remarkably similar, to judge from marketing materials and trial versions.

There are costs for the implementation and support of such all-in-one platforms. Here again, the costs for essentially similar services range from free [SE] to under a thousand dollars [OJS] to many orders of magnitude greater, with only slight differences in the services provided. An editorial platform already set up and made available to multiple journals ready to use (the application service provider [ASP] model) is far more effective than an individually installed and configured system.

The experience at Scholarly Exchange is that sharing the platform reduces cost to a minimum by spreading the costs over multiple journal titles. It enables almost immediate startup for groups interested in starting a journal on the SE/OJS platform. Setup time for SE is 5 minutes [SE] and for the journal staff another hour or so [OJS].

Putting shape to content

“One open access publisher reduced its costs by making authors responsible for their own copy editing, advising that they retain the services of a professional copy editor for accepted articles as needed”

The only inescapable costs at the end of the editorial review process relate to copy editing and tagging. Journal editors may already have spent considerable (free) time working with the author on content presentation, and many editors consider their work the final stage for an accepted manuscript. Others insist on the more traditional refinement that professional copy editors bring to bear, adding semantic clarity to content. One open access publisher reduced its costs by making authors responsible for their own copy editing, advising that they retain the services of a professional copy editor for accepted articles as needed. [BioMedCentral] Other publications have begun to follow that model.

The cost of copy editing depends on the area of specialization, manuscript length, and regional pay rates. Nonetheless, allocating from $25-$75 for a 5-10 page manuscript seems a conservative yet realistic number.

While many journals are content to present the PDF format as the final viewable product, with a consequent cost of perhaps another dollar or two to convert the final edited version of the manuscript, such archives as the US National Library of Medicine’s PubMedCentral require conversion of submitted material into highly structured and tagged ASCII text. Both metadata and data (content) must be fully identified at beginning and end of each manuscript element.

A Google search [Google2] yields a range of software programs in the multi-hundred dollar range to perform some degree of tagging, often to a user-configurable XML standard. Even the cleverest programs, attempting to digest variably entered word-processed documents and convert them to the simplest XML document definitions, tend to make endless mistakes. Human oversight and intervention are necessary. Costs may mount to an extent that defeats the benefit of the automation.

Another Google search [Google3] yields, as a viable alternative, a number of commercial outsourcing options, companies that offer re-keying or optical character recognition (OCR) and tagging services. Balancing the automation followed by human intervention-and-cleanup against simple outsourced human re-keying often favors the latter. More complex document tagging definitions (such as the US National Library of Medicine DTD) can be accommodated through the outsourcing option, with re-keying and tagging costs at slightly over a dollar per standard page of text.[3] The final product from the process (double key entry and proofing) may be a PDF, an HTML-tagged document, and an XML-tagged document, all for the same minimal page fee.

While none of the Scholarly Exchange journals has yet opted for tagging, either for HTML display or XML archiving, the quoted costs from various off-shore vendors have averaged just over a dollar a typescript page, with a five-page article encoded for archival storage costs priced at barely more than an elaborate Starbucks coffee preparation.

Scholarly Exchange itself does not offer the conversion services but directs its client journals to make arrangements with established cost-effective services. The transfer of documents occurs as either e-mail attachments or uploads from journal to conversion service and back. The requirements for conversion are those specified by an archiving organization, such as PubMedCentral, and are generally quite familiar to experienced conversion services.

Displaying and storing the final product

The final product is no longer the monolithic printed page but rather a collection of tagged ASCII text, with a wide range of supporting files (sound, video, and program files, to name but a few). They comprise the content and related metadata to be stored, converted, and displayed in a variety of ways and many more ways in the future.

The cost of preparing a final common pathway for a manuscript need no longer be prohibitive. The most page-like appearance for many journal readers is the PDF, whether viewed on-screen or printed out—titles centered, citation at the top of each page, all the niceties of the original printed product. Enhancing the value of PDF is its ability to be fully searchable electronically and its archival nature in an open file format environment that can be understood electronically and up-converted to new and evolving storage standards. [PDF/A].

PDF documents accommodate Web hyperlinking. While higher-resolution images can be embedded in a PDF file, there are inherent limitations in the viewable resolution of Web displays. Linking thumbnail or low-resolution images to higher-resolution or enlarged versions and including sound or animation may be done in PDF or may necessitate an HTML-based display file. Moving beyond PDF to this tagged-and-formatted state of the content entails some additional work. Fortunately, the preparation cost associated with this degree of structure in an article has dropped dramatically over recent years through the combination of technology and outsourcing.

Indexing, archiving, and sharing

Costs to include published articles in specialized indexed collections or archives are often not included in traditional estimates of first-copy production costs. The very nature of the Web and the plummeting costs of electronic storage open possibilities and profoundly reduce costs for wide information dispersion and distribution.

Scholarly Exchange (http://www.scholarlyexchange.org), the organization I helped found in 2002, actively encourages its journals to participate in electronic archives, wherever possible through university or public archives. Indexing or archiving costs are orders of magnitude less than they used to be, and for most publications they are no longer prohibitive. For a cost per page of barely more than a dollar, easy-to-find and reliable outsourcing firms output three file formats, PDF, HTML, and XML, the last of which is sufficient for institutional archiving.

The CrossRef digital object identifier (DOI), once seen as the only standard for identification and searching, is being replaced by rapidly evolving academic literature-search technologies from Google, Yahoo, and Microsoft, among others. The increasing trend toward self-archiving or university- or government-based digital archives leads to multiple copies of articles kept at different sites. This has been formalized with the open-source LOCKSS (Lots of Copies Keep Stuff Safe) software, enabling a journal in a digital archive at one institution to be replicated and kept fully synchronized with one, a hundred, or a thousand other digital archives, and essentially at trivial cost to the institutional archive. [LOCKSS]

The current state of digital archiving is another fluid entity with a variety of initiatives competing for attention and longevity. While for-profit publishers have under some duress agreed to deposit archival versions of their publications in selected university library archives, in the event of financial failure or other disappearance [Yale], a number of scholarly publishers have combined to promote Portico, a system of fee-based secure storage. [Portico] Whether this project will share with CrossRef the challenge of fees too high for small independent publications is yet to be seen. Equally uncertain in the future is CrossRef’s ultimate fate in the face of Google/Yahoo/Microsoft and their efforts at scholarly content indexing and discovery.

In the meanwhile, university librarians have been aggressively pursuing such independent projects as DSpace [DSpace] and Eprints [Eprints], both based on commonly available and open-source standards such as the Open Archive Initiative [OpenArchive]. These more democratic and less commercial resources may, in conjunction with the Internet Archive [Archive], offer a viable solution to the problem of archiving in the electronic environment.

Funding the enterprise

The ongoing debate about the cost of publishing a scholarly article is often subsumed in the question of who pays. Today the familiar subscription model, in which the reader pays for publications, shares the economic landscape with a newer author-pays model. BioMedCentral and the Public Library of Science both began with low author-pays fees. Their fees have risen, though not to the level that commercial publishers (who have adopted the author-pays concept to some extent) have set their fees.

It is difficult to determine definitive industry averages, since various forms of support and discounting apply to both not-for-profit and for-profit reader-pays models. Not-for-profit fees that began under $1,000 now appear to have edged up to $1,500, while commercial publishers levy a $3,000 or greater fee on authors who wish to make their articles freely available at the outset.

There are other models under consideration and in test. Scholarly Exchange has begun experiments with a number of models to promote sustainability. Since a journal’s startup period, the first months to a year, may be the most fraught in terms of recruiting content as well as editorial and reviewing assistance, SE provides the initial year of participation for a journal at no cost. This financial encouragement facilitates publishing experimentation at no risk, for a new publication or to transfer what was a paper-based product to paper-and-electronic or electronic-only. SE has also been testing the acceptance and viability of various forms of advertising, ranging from automatically back-fed topic-related advertising to display advertising solicited by the journal or society and on-screen acknowledgement of journal sponsorship. SE has developed the capability for donation opportunities on-screen in association with article downloads to support journal activities. [SE] While the amounts gathered may be relatively small in relationship to the typical revenues from subscriptions, the costs that have to be offset under the Scholarly Exchange model are trivial.

Real costs

While studies have pointed to first-copy paper or paper-plus-electronic production costs per article in the $4,000–$5,000 range [King] [Odlyzko2] [Bot] [ARL], the current experience that Scholarly Exchange has had with its hosted electronic-only journals points to dramatically lower costs.

“neither authors nor reviewers nor editors (in most cases) contribute directly to the costs”

Recall that neither authors nor reviewers nor editors (in most cases) contribute directly to the costs. There are minimal costs in the range of several dollars for the preparation and delivery of a reviewer-acceptable PDF document. There are similar costs for the final display and formatting of a PDF. Per-article costs for optional professional copy editing (after author-provided copy-edited text) and conversion with tagging can add in the range of $50 and $5 respectively to the cost. The platform upon which review, production, and display occur can add anywhere from nothing to $50 per article to the cost, depending on the platform chosen and the volume of articles processed.

My estimate is that a journal with 50 articles in a year could be published for under $4,000; double the number of articles, and the cost goes up to just over $7,000. At 250 articles a year, the cost is under $17,000. If the journal chose not to provide copy editing or XML conversion and tagging—two of the larger costs—the totals would be $1,200, $1,650, and $3,000 respectively. See table 1 for details (for distribution of costs).

Table 1: Out-of-Pocket Cost Estimates with an Open-Source-Based Publishing Initiative
Cost* Estimated cost / year / journal
100 submitted/50 published articles 200 submitted/100 published articles 500 submitted/250 published articles
STANDARD COSTS
Supported software platform 750.00 (1st year free) 750.00 (1st year free) 750.00 (1st year free)
Intellectual goods 0.00 0.00 0.00
Peer review (PDF creation $2/ ms) 200.00 400.00 1,000.00
Editorial management 0.00 0.00 0.00
Online display (PDF creation $2/ ms) 100.00 200.00 500.00
Total for required items 1,050.00 1,350.00 2,250.00
Required cost/published article 21.00 13.50 9.00
OPTIONAL COSTS
Optional: copy editing @ $50 / article 2,500.00 5,000.00 12,500.00
Optional: XML conversion and tagging 250.00 500.00 1,250.00
Total with options 3,800.00 6,850.00 16,000.00
Cost with options/published article 76.00 68.50 64.00

*All costs in US dollars.

“if one reduces publishing costs substantially, by two orders of magnitude, the question of who pays—reader or author—becomes irrelevant”

Hypothetically, if one reduces publishing costs substantially, by two orders of magnitude, the question of who pays—reader or author—becomes irrelevant. A journal article with a first-copy production cost of $4,000 and another with a cost in the $60 range have vastly different funding requirements. The double-digit article is no less valid in terms of content, appearance or archivability—overall value—than the quadruple-digit one. For a year’s worth of double-digit articles in a low-volume specialty journal (50 articles yearly, a not-uncommon figure), the out-of-pocket expense for a Scholarly Exchange–hosted journal is $1,000 rather than $200,000 through a commercial publisher.

The complexities of funding melt away, with a society, a research grant, a departmental fund or a library able to cover costs indefinitely and painlessly. Alternatively, a journal could levy a publication fee in the range of $50-$100 that would adequately cover expenses—well below the fees currently charged by many not-for-profits and a minuscule fraction of those charged by the profit-making entities.

In effect, the cost of one subscription for one library or one author publication fee practically covers the out-of-pocket costs for a journal with 50 articles distributed globally at no cost to readers.

The economic argument is controversial but supportable, and functioning examples are already in place [SE2]. The commercial world and many of the open access journals still wedded to old and established technology have difficulty accepting these newer and dramatically lower cost models. Beyond that, there is resistance, perhaps better described as reluctance and doubt, from the very community that could benefit most, academics.

Facilitating change

Promoting change requires an acknowledgement or an admission that the scholarly publishing mechanism is dysfunctional or broken. Many in the scholarly community fail to see the process in those terms, because the system appears to them to be functioning satisfactorily, and therefore the sought-after change has not occurred. [Harnad] They choose, for reasons of career advancement, to submit to journals that have an established reputation, are affiliated with their professional or scholarly society, or specialize in their field of expertise, not taking economics into account.

The existing system limps along at a cost that is not evident to academic users but is painfully apparent to administrators and librarians. The increasingly exorbitant prices are hidden in higher tuition fees, steeper institutional overheads charged for research grants, ever-longer lists of subscriptions cancelled, manipulative bundling arrangements, and restrictions on acquisitions of other reference resources—in effect the denial of access to intellectual resources.

Therein lies the irony: scholarship, even that published electronically in this Internet age, has decreasing visibility and availability. [Johnson]

The motivation to change has begun with the involvement of a growing chorus of librarians, professors, deans, and presidents, from many disciplines. It will be encouraged by emerging research demonstrating the value of open access publishing. [Eysenbach]

The principle of open access, giving away intellectual products of great value, requires that the means and cost of production be reduced to the barest minimum, as enterprises like Scholarly Exchange have now demonstrated for journals over several years.

It behooves universities and research institutes to rethink their criteria for advancement and tenure, to consider new journals produced at lowest cost and maximal free distribution as recognized venues for scholarship of the highest grade along with the established brands.

Combining the motivation with the means requires that universities, philanthropic entities, and governments make a commitment to support innovative and cost-effective approaches to scholarly communication. They must commit the minimal funding, and acknowledge the value of work appearing in these new venues by literature citation and academic promotion. The simple tools now exist, and there must be the support to begin using those tools. There need to be collaborative efforts that help journals make the transition from old technology to new—or re-invent themselves completely. [Crow] [SE] [Okerson]

Scholars need to understand that low cost of production does not imply low quality of content, that low cost enables greater distribution of knowledge and enhanced academic prestige. Editors require help and encouragement in using these tools. Centers of learning, governments, and philanthropic organizations have a new role: to facilitate the adoption of newer, simpler, less costly systems and methods of publication in the electronic environment. Only then will journals be capable of producing at least as high quality a product as before while sharing it more widely and at dramatically lower cost to society. [Willinsky]



Julian Fisher, M.D., has been peering into the future of information and medical informatics for several decades. He has been instrumental in the development of new products, from early decsion support tools for electronic medical records to the first CD-ROM textbook in medicine and public kiosks with health information for consumers. He developed one of the most widely used peer review software products for academic journals and co-founded the not-for-profit publishing innovator, Scholarly Exchange, of which he is the Managing Director. He is a consulting neurologist at the Beth Israel Deaconess Medical Center, Boston, on the neurology faculty at Harvard Medical School, the medical director of NeuroMatrix, Inc., and a district medical consultant to the U.S. Department of Labor. Dr. Fisher graduated with a B.A. cum laude from Yale University and with an M.D. from Johns Hopkins. He trained in pediatrics in Philadelphia and Lima, Peru, and neurology in Toronto and Rochester, N.Y. He may be reached at fisher@scholarlyexchange.org.


Acknowledgment

I thank my fellow director at Scholarly Exchange, Yuwei Shi (Fisher School of International Business, Monterey Institute), for the extended discussions that have led to this re-thinking of the entire publishing process. Portions of this paper have previously appeared in Science and Technology Libraries 27, no. 4 (2007): 63–78.

Notes

    1. Personal communications with Scholarly Exchange journal editors during 2006.return to text

    2. Amount Elsevier invested in 2000 to put journals online: $60 million. [NYTimes]return to text

    3. Personal communication, Mitchell Freiberg, NewGen Imaging Systems (India), http://www.newgenimaging.com/, November 29, 2007.return to text

    References

    [Archive] Internet Archive home page, http://www.archive.org/

    [ARL] “Graph 2, Monograph and Serial Costs in ARL Libraries, 1986–1999,” ARL Statistics 1998–99. Washington, D.C.: Association of Research Libraries, 1999, p. 9.

    [Atypon] Atypon Systems home page http://www.atypon.com/

    [Bergstrom] Bergstrom, Theodore C. “Free Labor for Costly Journals?” Journal of Economic Perspectives 15, no. 3 (2001): 183–98; author preprint, http://www.econ.ucsb.edu/~tedb/jep.pdf.

    [BioMedCentral] BioMed Central, “Using copy-editing services,” http://www.biomedcentral.com/independent/develop/copyediting

    [Bot] Bot, Marjolein, Johan Burgemeester, and Hans Roes. “The Cost of Publishing an Electronic Journal: A General Model and a Case Study.” D-Lib Magazine 4, no. 11 (November 1998), http://www.dlib.org/dlib/november98/11roes.html

    [Crow] Crow, Raym. “Publishing Cooperatives: An Alternative for Non-profit Publishers.” First Monday 11, no. 9 (September 2006), (http://www.firstmonday.org/issues/issue11_9/crow/index.html

    [DOAJ] Directory of Open Access Journals home page, http://www.doaj.org

    [DPubs] Cornell University Library. DPubs home page, http://dpubs.org/

    [DSpace] DSpace home page, http://www.dspace.org

    [EditorialManager] Aries Systems. Editorial Manager home page, http://www.editorialmanager.com/

    [Eprints] Eprints home page, http://www.eprints.org/

    [Eysenbach] Eysenbach, Gunther. “Citation Advantage of Open Access Articles.” PLoS Biology 4, no. 5 (2006): e157, [doi: 10.1371/journal.pbio.0040157]

    [Google1] “free web conversion to pdf adobe” (in quotation marks)

    [Google2] “XML conversion” (in quotation marks)

    [Google3] “XML outsourcing” (in quotation marks)

    [Harnad] Harnad, Stevan. “Free at Last: The Future of Peer-Reviewed Journals.” D-Lib Magazine 5, no. 12 (December 1999), http://www.dlib.org/dlib/december99/12harnad.html

    [HighWire] Stanford University Libraries. HighWire Press home page, http://www.highwire.org/

    [King] King, Donald W., and Carol Tenopir. “Economic Cost Models of Scientific Scholarly Journals.” Paper presented to the ICSU Press Workshop, Keble College, Oxford, UK, 31 March to 2 April 1998, http://www.bodley.ox.ac.uk/icsu/kingppr.htm

    [Johnson] Johnson, Richard K. “Open Access: Unlocking the Value of Scientific Research.” Paper presented at The New Challenge for Research Libraries: Collection Management and Strategic Access to Digital Resources, Conference sponsored by the University of Oklahoma, March 4–5, 2004, http://www.arl.org/sparc/bm~doc/OpenAccess_RKJ_preprint.pdf

    [LOCKSS] LOCKSS home page, http://www.lockss.org/

    [NYTimes] Kirkpatrick, David D. “As Publishers Perish, Libraries Feel the Pain; Mergers Keep Pushing Up Journal Costs.” New York Times, Business section, Nov 3, 2000, http://query.nytimes.com/gst/fullpage.html?res=9F05E4D71F30F930A35752C1A9669C8B63&sec=&spon=&pagewanted=all

    [Odlyzko1] Odlyzko, A. M. “Peer and Non-peer Review.” Peer Review in Health Sciences, 2nd ed., Fiona Godlee and Tom Jefferson, eds., pp. 309–11. London: BMJ Books, 2003.

    [Odlyzko2] Odlyzko, A. M. “The Economics of Electronic Journals.” First Monday 2, no. 8 (August 1997), http://www.firstmonday.org/issues/issue2_8/odlyzko/index.html/, reprinted in Journal of Electronic Publishing 4, no. 1 (September 1998), http://hdl.handle.net/2027/spo.3336451.0004.106. Definitive version in Technology and Scholarly Communication, Richard Ekman and Richard E. Quandt, eds. (Berkeley, Calif.: University of California Press, 1999), 380–93.

    [Okerson] Okerson, Ann. “Back to Academia? The Case for American Universities to Publish Their Own Research.” Logos 2, no. 2 (1991): 106–12. http://www.library.yale.edu/~okerson/case.html

    [OJS] Public Knowledge Project. Open Journal Systems home page, http://pkp.sfu.ca/ojs/

    [OpenArchive] Open Archives Initiative home page, http://www.openarchives.org/

    [PDF/A] “ISO 19005-1. Document management - Electronic document file format for long-term preservation - Part 1: Use of PDF (PDF/A),” http://www.digitalpreservation.gov/formats/fdd/fdd000125.shtml

    [Portico] Portico home page, http://www.portico.org/

    [ScholarOne] Scholar One home page, http://www.scholarone.com

    [SE] Scholarly Exchange home page, http://www.scholarlyexchange.org

    [SE2] Scholarly Exchange journals listing, http://www.scholarlyexchange.org/hosted.html

    [Sosteric] Sosteric, Mike, Yuwei Shi, and Olivier Wenker. “Electronic First: The Upcoming Revolution in the Scholarly Communication System.” Journal of Electronic Publishing 7, no. 2 (December 2001), http://hdl.handle.net/2027/spo.3336451.0007.203

    [Wikipedia] Wikipedia, s.v. “Portable Document Format.” http://en.wikipedia.org/wiki/Pdf

    [Willinsky] Willinsky, John. The Access Principle: The Case for Open Access to Research and Scholarship. Cambridge, Mass.: MIT Press, 2006.

    [Yale] Yale University, “Yale Library to Plan Digital Archives with Elsevier Science,” news release, February 23, 2001, http://www.yale.edu/opa/newsr/01-02-23-02.all.html