Panel Discussion on E-Print Servers and Traditional Publishing - A Report from the 49 th IEEE CDC in Atlanta Friday, December 17, 2010 12:00-13:30 Grand Salons A/B, Hilton Atlanta Hotel in Atlanta, Georgia USA As a service to the IEEE Control Systems Society (CSS) and the larger technical community that it serves, a panel of experts was convened at the 49 th IEEE Conference on Decision and Control held in Atlanta over the three-day period December 15 17, 2010. The panel s topic was the relationship between e-print servers (like MIT s DSpace and Cornell University s arxiv.org) and traditional publishers in Science, Technology, Engineering, and Mathematics (STEM). The purpose of the panel was not to recommend policies that might guide this relationship but rather to provide a forum in which STEM publishing stakeholders could discuss matters such as publication timelines, the protection of intellectual property rights, and the sustainability of various publishing business models. Background STEM publishing has seen enormous changes over the last two decades. Although manuscripts are now circulated electronically for peer review, most journals that cover fields related to automatic control and optimization continue to have review times that require six months or longer to complete a review cycle. If the time required for revisions and re-review are accounted for, the data suggests that getting an article published takes longer than a year in most highly ranked journals. Against this backdrop, the World Wide Web has provided a powerful resource for rapid dissemination of new research. It is very common for people who are engaged in STEM research to post both preprints and reprints on personal websites. Recently, several leading universities have asked their faculty to post both published articles as well as preprints on university servers. In addition to such de facto open access channels for research dissemination, there is the phenomenally successful arxiv.org server that is now operated by the Cornell University Library. Because there is growing interest among researchers in these means for rapid dissemination, the IEEE Control Systems Society has taken steps to help its members to take advantage of e-print publication on arxiv. A Systems and Control section of the Computing Research Repository (CoRR) of the arxiv has been recently created. Further details about subscribing and submitting contributions are available on the website http://arxiv.org/corr/home. This section complements an earlier Optimization and Control section established by Eduardo Sontag and currently moderated by Yuan Wang under the Math Repository http://arxiv.org/archive/math.
We have just learned that the arxiv administrators have agreed to cross link the Optimization and Control section of the Math Repository with the Systems and Control Section of the CoRR Repository. This means that at some point in the future, all submission to either cs.sy or math.oc will appear in both places. The decision to organize the panel resulted from several people feeling that the time was right to explore the opportunities and challenges or new publishing models such as e-print servers. People in the community who have been advocates for such exploration include Yuan Wang, Robert Kosut, and Shankar Sastry. Thus the organizers decided to bring together some leaders in traditional publishing along with someone representing arxiv.org to share ideas on the future of STEM publishing and what the future holds for relationships between e-print servers and traditional publishers. Panelists - Christos G. Cassandras, 2010 IEEE CSS Vice President for Publications - G. David Forney, Jr., Past President of the IEEE Information Theory Society - Christopher Greenwell, Publisher, Control and Signal Processing, Elsevier Ltd - David K. Marshall, Publisher, SIAM - Jon G. Rokne, 2010 IEEE Vice President Publication Services and Products - Simeon Warner, arxiv.org and the Cornell University Library Panel Moderators - John Baillieul, Past President of the IEEE Control Systems Society - Roberto Tempo, 2010 President of the IEEE Control Systems Society Panel Highlights Part of the challenge facing all publishers is finding ways to maintain a revenue stream that will cover the costs of doing business while at the same time recognizing that readers have changed their expectations about what they should be required to pay for. Traditional business models that served well in the days of print distribution are generally not easy to adapt to end-to-end digital publishing in which everything from acquisition to peer review to final publication and archiving are done electronically. The numerous experiments with so-called Open Access (OA) publishing particularly in the life sciences have presented financial challenges to both those who are trying to build new publishing businesses based on OA and to the traditional publishers that must find appropriate responses to the new competition. The panelists agreed that e-print servers as well as traditional publishers need to operate using sound and sustainable business models. Dave Forney related some of the history of the IEEE Information Society s involvement with the arxiv.org preprint server. In October of 2004, the IT Society Board of Governors voted to endorse encouraging members and authors submitting to the IEEE Transactions on Information Theory to post their manuscripts on the arxiv server. Specifics of the policy that was adopted include that it would be at the author s option and include preprints only. The IEEE Publications Services and Product Board (PSPB) gave an informal approval as an experiment. Dave also reported that after a period of initial promotion of arxiv posting (between 2005 and 2007), the uptake has been increasing, and in a recent issue of the IEEE Transactions on IT,
around 44% of the papers had also appeared in some form on the arxiv server. A comparable statistic for the larger body of material published on the IEEE Xplore archive is that around 33% of the manuscripts are now freely available on author websites or other points of access. Dave Forney finally concluded that despite the high level of availability of material on arxiv, the IT Society had not observed any adverse financial impact. Christopher Greenwell, Publisher, Control and Signal Processing Journals for Elsevier, reported that Elsevier has adopted a test and learn approach to new technologies and methods of publication. Chris raised a number of points that were repeatedly echoed during the panel discussion emphasizing the need for e-print servers to find viable business models that would ensure their continued existence as electronic archives. He also shared the official policy that Elsevier has adopted regarding e-print servers: Elsevier's current policy on electronic preprints: Elsevier does not consider the existence of an electronic preprint of an article as ''prior publication, nor will it require authors to remove electronic preprints of an article from public servers should the article be accepted for publication in an Elsevier journal. If a paper posted on a preprints server is accepted for publication in an Elsevier journal, the preprint version may be replaced by an updated author's version of the paper but a link to the journal and a full citation should be included. In line with Elsevier policy the final published version may not be submitted to the preprints server. While the other publishers participating in the panel (the IEEE and SIAM) do not have a completely equivalent policy, IEEE Vice-President of Publication Services and Products, Jon Rokne and SIAM Publisher David Marshall indicated generally congenial approaches to authors who wish to post published work on their personal websites and even on e-print servers. Details can be found in the presentations that are attached below. Jon Rokne referred to the specifics of IEEE policies on such matters that are spelled out in the extensive IEEE PSPB Operations Manual. This can be downloaded from http://www.ieee.org/documents/opsmanual.pdf. Simeon Warner discussed some of the business aspects of running the arxiv.org server, and while the data presented suggests that it provides a very cost-effective mode of electronic dissemination, a sustainable financial model is a goal not yet fully realized. Nevertheless, author interest in arxiv.org continues to grow with no plateau in sight. Warner presented results from a study that showed that e-prints have not undermined the usage of journal papers in the astrophysics community (http://dx.doi.org/10.1087/095315107779490661 or http://arxiv.org/abs/cs/0609126). This finding was consistent with the experience of the IEEE Information Theory Society that was discussed in the presentation of Dave Forney. The final panelist to present comments was Christos Cassandras, the CSS Vice President of Publications Activities for 2010. He emphasized the quality assurance associated with peer review and how this introduced latency into the research dissemination process. He reflected a general consensus that e-print servers should be able to collegially coexist with traditional publishers provided everyone respected copyright and intellectual property rights.
After the formal presentations by panelists, the discussion involving comments and questions from the audience, revisited a number of points. Various people expressed concern that e-print servers like arxiv might be abused by those wishing to avoid having their work subjected to the scrutiny of peer review. While this was repeated in different ways by several people, it is countered to some extent by the explicit prohibition of excessive submission to arxiv. The arxiv policy (http://arxiv.org/help/moderation) states: Excessive submission rate. Articles submitted to arxiv must be of refereeable quality, and there is a practical limit to the rate at which appropriate, independent submissions can be produced by any one person. Moderators may request that a particular submitter limit their submission rate if they have a history of many submissions to inappropriate areas or of doubtful refereeability. No one in the audience or on the planel cited any problematic instances of authors using the e- print server to avoid peer review and mislead readers. Concluding Remarks The panel on the evolving relationship between e-print servers and traditional publishers provided a brief look at current modes of operation and operating policies of both the e-print repository arxiv.org and several traditional publishers (the IEEE, SIAM, and the Control and Signal Processing Journals of Elsevier). To varying degrees, business models for all forms of publishing remain in flux. There was a high degree of unanimity among all panelists and members of the audience regarding the value of peer review. There were only minor differences of approach among the publishers represented. While all expressed willingness to allow authors to post final versions of their published papers, Elsevier and the IEEE have adopted the policy that the posted version can be the paper in the form that it is finally accepted, but it cannot be the fully edited version that has been appear in the publishers journals and web archive (Science Direct and Xplore respectively). SIAM, on the other hand, grant authors the right to post an electronic version of the final SIAM file of the work on the author's current institutional internet server It does not grant the right to post preprints or the redacted version to their personal website. Additional details of the panel discussion can be found in the append presentations. --- Respectfully submitted, John Baillieul and Roberto Tempo January, 2011
Panel Discussion on E-Print Servers and Traditional Publishing Organized by John Baillieul and Roberto Tempo IEEE Control Systems Society
Systems and Control Section of the arxiv arxiv: repository owned and operated by Cornell Systems and Control section (under CS) was recently initiated http://arxiv.org/corr/home Fast distribution of results to the control community No review process Moderators verify if the paper is relevant to Systems and Control Moderators: Marco Lovera, Rick Middleton, RT IEEE Control Systems Society
Acknowledgments Eduardo Sontag who initiated the Optimization and Control section of arxiv (under math) Yuan Wang moderator of OC for several years Robert Kosut and Shankar Sastry for providing ideas IEEE Control Systems Society
Panel Discussion on E-Print Servers and Traditional Publishing Future of STEM (Science, Technology, Engineering and Mathematics) publishing and related copyright issues - David Forney (IEEE Information Theory Society) - Chris Greenwell (Elsevier) - Jon Rockne (IEEE Vice President PSPB) - David Marshall (SIAM) - Simeon Warner (arxiv) - Christos Cassandras (IEEE Control Systems Society) IEEE Control Systems Society
E-Print Servers and Traditional Publishing Atlanta, Friday, December 17, 2010 Thursday, December 30, 2010
On the World Wide Web, everything wants to be free. If you put it all online for free, you will dramatically cannibalize your book sales. -Nathan Myhrvold, Chief Gastronomic Officer ZAGAT Surveys, quote in the NYT, Nov. 14, 2010 Thursday, December 30, 2010
Information wants to be free On the one hand information wants to be expensive, because it s so valuable. On the other hand, information wants to be free, because the cost of getting it out is getting lower and lower all the time. - Stewart Brand Thursday, December 30, 2010
The IEEE Information Theory Society and arxiv G. David Forney, Jr. CDC, Atlanta December 17, 2010
History 2004: Ad Hoc Committee recommendations Physics model appropriate for IT Society BoG should endorse and actively promote posting on arxiv October 2004: IT Board of Governors endorses posting on arxiv Author s option Preprints only IEEE Pubs Board: informal OK as experiment Founding of cs.it = math.it
History (cont.) 2005-2007: Active promotion Annual articles in IT Newsletter Authors prompted to post on arxiv when they submit 2004-2010: Adoption (articles posted on cs.it) 2004 2005 2006 2007 2008 2009 2010 41 331 365 614 791 962 1172 Second most active CS section (after Discrete Math) Current: 36 of 81 articles (44%) in Nov-Dec T-IT had been posted
Some tentative lessons Physics model (arxiv) is a good one for us and IEEE Benefits are modest but noticeable Rapid communication of new results Public, open access archive Time stamps No financial impact observable Cannot rely on authors for orderly, quality-controlled literature Cultural change requires leadership, promotion
Journal publishing in a changing world E-Print (Preprint) Servers and Traditional Publishing Christopher Greenwell Publisher, Control and Signal Processing Journals IEEE Panel Discussion on E-Print Servers and Traditional Publishing IEEE Conference on Decision and Control (CDC), Atlanta, Georgia December 17th, 2010
The Internet has changed STM Publishing The internet has changed the way that researchers find, gather, organise, and publish STM information. Find: Google search, Publisher databases (e.g., ScienceDirect), Abstract and Citation databases (e.g., Scopus) Gather: Downloads, Reference links, Social networks Organise: Reference managers (e.g., Mendeley) Publish: Personal web sites, Preprints (E-Prints), Open Access journals, Traditional journals Within this context Elsevier has adopted a test and learn approach to new technologies and methods of publication. 2
The Growth of Academic/Scholarly Serials There has been a marked increase in new launches in recent years 3
The Growth of Academic/Scholarly Serials The internet has resulted in a significant growth in serials Between 2008 and 2010 there were 205 Engineering serial launches (9% of the total of 2,398) Many recent launches are online-only with few articles and no strict publication schedule. But not all these launches succeed 1990s: 16,409 launched, over 2,600 of these (15.8%) have ceased 2000s: 15,353 launched, to date 800 (5%) of these have ceased Within this context Elsevier still launches journals with a more traditional model with the view that quality and preserving the scientific record remain the underlying principles Source: Ulrichsweb.com June 2010. NB: Data for recent years incomplete 4
Why do Journal Publishers exist? The first peer-reviewed journal founded in 1665 by Royal Society Journal publishing has evolved dramatically since, but its core functions remain: Registration of new research findings Quality Assurance through peer review Dissemination globally Archiving in perpetuity 5
Comparison of Preprints & Traditional Journals Preprints Registration: Fast Quality Assurance: Recommendation to be able to submit (some) Basic field screening Very low rejection rates Dissemination: Global Archiving: Permanent? Traditional Journals Registration: Slow Quality Assurance: Peer Review 7,000 editors 70,000 editorial board members 500,000 referees >600,000+ submissions/yr 1 million referee reports/yr 40%-90% of articles rejected Dissemination: Global Archiving: Permanent, Accessible 6
Peer review: Researcher perceived value 90% of researchers believe that the peer review process improves the quality of published research Source: Peer review: benefits, perceptions and alternatives, Mark Ware Consulting, Publishing Research Consortium 2008 7
The Case of High-Energy Physics In High-Energy Physics (HEP) preprints are seen as the traditional first step in publication (going back to the 1960s when they began paper distribution of preprints). Rapid publication is important in HEP Almost all HEP papers appear on arxiv.org The great majority of HEP authors submit their preprint to a peerreview journal as the final validation of their results Thus we aim to make it easy for authors to submit to our journals: A "download from arxiv" button in EES which allows submissions to be retrieved from arxiv using the arxiv reference number instead of uploading. The system retrieves the source files from the arxiv After publication, we collaborate with arxiv to ensure that preprints are labelled with the full citation reference of the published article 8
Evolution of Published Research Published research is reviewed, updated and edited to evolve from unpublished paper to conference paper and/or preprints to peer reviewed article these are not mutually exclusive 9
Elsevier s Policy on Preprints Elsevier's current policy on electronic preprints: Elsevier does not consider the existence of an electronic preprint of an article as ''prior publication, nor will it require authors to remove electronic preprints of an article from public servers should the article be accepted for publication in an Elsevier journal. If a paper posted on a preprints server is accepted for publication in an Elsevier journal, the preprint version may be replaced by an updated author's version of the paper but a link to the journal and a full citation should be included. In line with Elsevier policy the final published version may not be submitted to the preprints server. http://www.elsevier.com/wps/find/intro.cws_home/electronicpreprints 10
Thank you! Christopher Greenwell c.greenwell@elsevier.com 11
E-Print Servers and Traditional Publishing: A View from IEEE Jon Rokne IEEE Vice-President PSPB Friday, December 17, 2010 IEEE Control Systems Society 49 th Annual Meeting
Background IEEE has been active in publishing since its inception. IEEE has been a leader in electronic dissemination of information. IEEE s publications are among the top rated in its areas of interest. IEEE is now publishing in both paper and electronic form. IEEE moving from separate titles to an integrated database of articles (Xplore). 2 December 17, 2010
E-archives Moving beyond journals: the future arrives with a crash (editorial BJM, 1999, pp. 1637-9). IEEE discussion started in the 1990s Bill Hagen presented IEEE view of intellectual property issues in the electronic area at APS meeting in 1999. IEEE Principles of Scholarly Publishing 2005 does not mention medium 3 December 17, 2010
E-archives ADVANTAGES DISADVANTAGES Rapid posting Possibilities for comments Searchable Author control Retains copyright Governance issues Uncontrolled growth Lack of quality control Funding issues Persistence issues No publishing help Plagiarism issues 4 December 17, 2010
IEEE (Policy) Before submitting an article to an IEEE publication, authors frequently post their manuscripts to their own web site, their employer s site, or to another server that invites constructive comment from colleagues. Upon submission of an article to IEEE, an author is required to transfer copyright in the article to IEEE, and the author must update any previously posted version of the article with a prominently displayed IEEE copyright notice (as shown in 8.1.9.B). Upon publication of an article by the IEEE, the author must replace any previously posted electronic versions of the article with either 1. the full citation to the IEEE work with a Digital Object Identifier (DOI) or link to the article abstract in IEEE Xplore, or 2. the accepted version only (not the IEEE-published version), including the IEEE copyright notice and full citation, with a link to the final, published article in IEEE Xplore. 5 December 17, 2010
I.E., That is, IEEE allows the posting of the final accepted version of an article on an author s institutional server, but not the final published article. Instead, an author is encouraged to add a link to the final authorized version. 6 December 17, 2010
Third party servers (Policy) With the exception of Section 8.1.9.A.3 above which is only applicable to agency-funded research, no third party other than authors and employers acting in accordance with this Section 8.1.9 may post IEEE-copyrighted material without obtaining the necessary licenses or permissions from the IEEE Intellectual Property Rights Office or other authorized representatives of the IEEE, and only under terms approved by PSPB. 7 December 17, 2010
Reason for policy Non-profit organizations do make profit. The profit is used to sustain activities that benefit members and society in general. The mission of IEEE is: IEEE s core purpose is to foster technological innovation and excellence for the benefit of humanity. 8 December 17, 2010
IEEE publishing revenue sustain The cost of managing the publishing process Maintaining a secure and reliable way of delivering information electronically Standards activities Regional, sectional activities Educational activities Protection of authors rights 9 December 17, 2010
Questions? 10 December 17, 2010
E-Print Servers and Traditional Publishers
SIAM S copyright assignment agreement.in principle.
grants the author, among other rights, the right to post an electronic version of the final SIAM file of the work on the author's current institutional internet server It does not grant the right to post preprints or the redacted version to their personal website.
SIAM S copyright assignment agreement.in practice.
is more flexible. Very few authors ask permission to post preprints or articles to their personal websites or other repositories such as ArXiv. When they do, we remind them of our policy but don t enforce it.
And, of course, others don t bother to ask. This is clearly demonstrated by the 9,560 results from a search of ArXiv using the term SIAM.
Our unease with this issue stems from two areas: 1. The usual concern regarding the effect on revenue from subscriptions which supports many of SIAM s programs. 2. Of equal importance, staff and the Editors-in-Chief are concerned about plagiarism. We have had cases where articles have been plagiarized from a preprint.
Bottom line We are pragmatists.
Panel Discussion on e-print servers and traditional publishing arxiv Simeon Warner (Cornell University Library) @ 49 th IEEE Conference on Decision and Control Atlanta 2010-12-17
Cost per article arxiv IR Journal $500k 2010 budget 70,000 articles/year + rapid + open + inexpensive + personal service + motive to fund - visibility + peer review + official - cost - access barriers $7 $50 $500
Astrophysics e-print and journal use Fig2 from arxiv:cs/0609126 by Henneken, Kurtz, Warner, Ginsparg, et al.
PUBLISHING MODELS C. G. Cassandras Division of Systems Engineering Dept. of Electrical and Computer Engineering Center for Information and Systems Engineering Boston University Christos G. Cassandras CODES Lab. - Boston University
PUBLISHING MODELS Filter for QUALITY at the expense of LATENCY SUBMISSIONS REVIEW DELAY PUBLISH E-PRINT SERVER Avoid LATENCY at the expense of filtering for QUALITY
EFFECT OF INCREASING SUBMISSIONS IEEE Transactions on Automatic Control, 2000-2009 Full Paper Submissions 800 Submissions 600 400 200 0 2000 2002 2004 2006 2008 2010 Year SUBMISSIONS LATENCY, QUALITY
CONCLUSION E-PRINT server publishing model - Avoids increased LATENCY due to high submission volume - operates in parallel, non-threatening to traditional model - assuming it does not introduce copyright violation issues Christos G. Cassandras CISE - CODES Lab. - Boston University