ieee eee Free online availability substantially increases a paper’s impact Sir —The volume of scientific literature far exceeds the ability of scientists to identify and we all relevant information, The ability to locate relevant research 4quidey will dramatically improve ‘communication and scientific progress Although availability varies greatly by discipline, more than amnion research articles are now frely available on the web. Here we investigate the impact of free ‘online availabilty by analysing citation sates, Onlin availability ofan article may not greatly improve access and impact without efficient and comprehensive search services; substantial percentage ofthe lterature needs to be indexed by these search services before scientists consider them useful In computer science, a substantial percentage ofthe literature ie ‘online and available through search engines such as Google (www google com), or specialized services suchas ResearchIndex (www.esearchindex.or8) —although the greatest impact of online availabilty is yet to come, because compre- hensive search services and more poner search methods have become available only recently. ‘We analysed 119,924 conference article in computer science and related disciplines, obained from DBLP (hip uni-trierde). In these lds, conference articles are typically formal publications and.areolten more prestigious than journal articles, with acceptance rates at some conferences below 1096, We estimated citation counts and online availability using ResearchIndex, excluding self-citations “The figure shows the probability that an articles freely available online as fanction of the number of citations tothe article, and the year of publication of the article. The results aze dramatic showing a clear correlation between the number of ‘umes an article i cited and the probability thal the article is online. More highly cited articles, and more recent articles, ate significantly more likely tobe online, in ‘computer science, The mean number of citations to offline articles is 2.74, and the ‘mean number of citations to online articles is 7.03, an increase of 157%. ‘We analysed differences within publication venues (the proceedings of Tr Sener ee eet conference fora particular year, for example), looking atthe percentage increase in citation rates for onlin articles ‘When offline articles were more highly cited, we used the negative ofthe percentage increase for offline articles hence ifthe average number of citations for olline articles is two, andthe average for online articles is four, the percentage {increase would be 100%. For the apposite situation, the percentage increase would be — 100%, Averaging the percentage increase across 1,194 venues containing at least five ofline and five online articles reeultein an average of 336% (median 158%) more citations to online computer science articles compared with offline articles published in the same venue. Ifwe assume that articles published in the same venue are of similar quality, then the analysis by venue suggests that online article are more highly cited because of their easier availability, This assumption is likely w be more valid for top-tier conferences with very high acceptance standards, Restricting our analysis tothe {op 20 publication venues by average citation rate gives an increase of 286% (amedian 2846) inthe citation rate for line articles Free online availability facilitates aceessin many ways, including provision of online archives dteet connections among scientists or research groups; hassle-free links from e-mail, discussion groups and other services indexing by web search engines; and the creation of Number of etstons Analysisof119,924conferenceartidesin computer science and elated disciplines. The actual percentage of articles available online greater, owingtolimitationsin the ‘xtracton of artcleinformation fom online Aocuments and limitations inoceting artileson the web. Only points with greater ‘han 100 articles are computed. {#80 2001 Macmillan Magazines Ltd correspondence third-party search services, Free online availablity of scientific literature offers substantial benefits to science and society. To maximize impact, minimize redundancy and speed scientific pro authorsand publishers should aim t0 rmnake research easy to acess Steve Lawrence NEC Research inti, Independence Way, Princeton, New Feri 9540, USA Authors willing to pay for instant web access Sir— Authors ofjournal articles want their efforts certified by peer review and made convenient availabe tothe widest possible readership. They do not expect or receive royalties. What they hope for is ‘impact — attention, especially from other researchers, and recognition, especially from those who decide on hiring ‘and promotions, Journal articles have greater impact they are immediately and ‘widely accessible. Maximum impact is achieved by immediate fee web access (see ‘Nature 410, 1024-1025; 2001), Journal publishers are taking advantage ofthe web's quick, convenient delivery of information by ceating electronic versions oftheir traditional print journals, accessible only to subscribers otto cients ofan institution with asitelicence. ‘Authors are not offered immediate free ‘web acces for their articles The Entomological Society of America (ESA), publisher of four leading entomo- logical journals, recently began selling immediate free access (see httpl/essrvr entnem,ufl.edu/~valker/tjwbib walker ‘htm). The results suggest thatamarket- driven transition to fee acess forall, articles inall journals is possible ESAs business plans simple: it will wide immediate fre web access, at fair price, to authors who wantit Asthe cost of offering this rises (because of subscription cancellations), the price will increase. No author will be required to purchase it, and sales of subscriptions tothe journals will continue as long a they are profitable. The endpoint ofthis plan is uncertain, but it ray lead tothe demise of paper publication and subscriptions, as authors and the institutions that support them embrace fhee access and strive to reduce costs Direct cost of the present system, include printing paper issues, imiting aceess to electronic versions, and making past and present volumes accessblein hundreds of research libraries, Inditect costs are reduced impact of articles and severely restricted access by researchers in smaller institutions and in developing countries. Nonetheless, many stakeholderscorrespondence believe that printed issues, or atleast tolls in the form of subscriptions and site licences, will continue indefinitely. ESA began selling immediate free web access in January 2000, During the fist ‘wo months of the service, authors bought itfor 13% of articles, ising steadily to 5996 during March and April 2001. The price for the servicei currently 75% ofthe price ‘of 100 paper reprints, for example $90 fora T-page article. This price provides a greater posit margin than for paper reprints, ‘which ae expensive to produce and. deliver. Immediate fee web access requires only that the PDF file ofthe article ismade freely accessible on ESAs web server. immediate fre access ita profit: making service that many authors want and will pay for, why is ESA apparently the only publisher that sells? For scientific societies, the answer is probably that their institutional inertia is great and their members have yet to lobby for it— as ESA, members did. Commercial publishers may fear that selling immediate free access to those who want it may lead toll authors ‘buying it, in which case revenues from subscriptions and site licences might cease. (On the other hand, societies have supplemented modest incomes from lower-priced lirary subscriptions with member dues and page charges. Without journal subscriptions, societies and ‘commercial publishers will collect page charges to pay for refereeing, editing and ‘composing. Publishers will pay nothing to ake the articles freely web accessible, because research ibraries and PubMed Central will post them without charge ‘Authors should encourage publishere to provide immediate free access ata fair price, Other things being equal, many will prefer to publish in journals that provide it especially as electronic literature indexes begin linking directly to the e-versions af articles, Most authors would like nothing better than for ther articles tobe available in fll ex, without tolls, via links in widely used literature indexes, ‘Thomas). Walker Department of Entomology and Nemtolog, University of Flrida, PO Box 11082, Gainervil Flori 32512-2620, USA Impact factors, and why they won't go away Sir—Brunstein in Correspondence asks ‘whether online publishing wll herald the end ofimpact factors, Unless hee forecasting the end of print publications altogether this is doubtful. Were print journals tw disappear, however, am confident tata new impact factor would beinvented. Information scientists are already computing web impact factors Tr would be more relevant to ue the actual impact (citation frequency) of individual papers in evaluating the work of individual scientist rather than using the journalimpact factor as asurtogate The later practic is fraught with dficules, as Soglen and others have pointed out A longa scientists publish articles containing list of cited references, twill be posible to calealate impact factors Tis to behoped that citation practices on the web will become sufficiently standardized to permit accurate calculations. Tevnillbenecesraryto distinguish between citations to URLS for research articles, on the one and, and, om the other, to readerships as reflected in ‘webometre’ studies measuring web activity One ordinarily assumes that there sremany mote readers than cers, but there isa widespread mythology that authors are cited mate than they aze read Eugene Garfield Instat for Seen formation, 350 Maret Sire Pilea, Pensvanis 19104 USA Evolution is what’s needed, not revolution Sir-—In the current debate on web availabilty of peer-reviewed scientific literature there seems tle attempt to retain the advantages of the present aystern of ournals. “These ate, frst the element of special- ization (the journal's field) and, second, the quality and novelty ofthe science the journal demands for publication (ts place in the pecking order). Both aspects are crucial, ae authors wish their articles tobe immediately read by the appropriate audience and to accrue the kudos appropriate to the significance oftheir work, So any ‘universal web-based system must accommodate these two ape: This could be achieved if existing journals were replaced by equivalent ‘websites that were run not by commercial publishers but by the learned societies on non-profit bass, Provided the societies did pol use their sites to generate extraincome, they could keep costs to a mininsum, Reviewing could be cartied out as nows, the costs would be covered by page charges toauthors (for both accepted and rejected submissions), and by small charges for web aceess to published articles. The subscription charges would be scale: relatvely high for an institution #86 2001 Macmillan Magazines Ltd subscribing forthe benefit of lite staf. at intermediate level for an individval laboratory, and very low fora personal subscription. (Control of access could be through the predesignated IP addresses ofthe servers sand individual terminals of subscriber. Thus an individual laboratory could have online subscription tits favourite sites for a fraction of the cot that it must now pay lo get them as printed journals fa library ‘wanted paper on its shelves, it could either pay the ste to sell it what we currently call 1 journal’ or it could have a licence ta Print the content out ae part ofits institu- tional subseription ‘A great deal ofthe effort for web publication (for example, generating PDF versions of text and figures in the house style ofthe site) canbe undertaken by the asuthors, since they are the keenest to see the article in the public domain, The current requirements for publication in the Journal of Biological (Chemisty, in which every aspects electronic, show that this isa straight forward procedure, which, while not costfree, is not prohibitively expensive ‘when printing costs donot have ro be covered. Colyn Crane-Robinson BiophyncsLaboratre,Univerty of Porton, Sr Michaels Building, White Swan Road, Portimouth POL2DT, UK The Net is many people’s only chance of access Sir—Pakistan isnot on the publishing map. 1 doubtifthe collective population of {HO raion manage to rubsrbe to mere than a few cost journals. On the ather hhand we have only about 300,000 computer users. Withholding fll txt ffom country suchas Pakistan is thus fidcaloue {would suggest that lltext thould be made avuable to everyone on thelinternet This would not affect journals financially, since mos people using this service in countries sch as Pakistan would never be able to subscribe to them, EAKhin 1AWhite owe Lane, Sunder Das Road Ire Pakistan Erratum Stevan ara Comment on te te e Mert ae $10, 1024-1025; 200", tsimae fe nem cota peereiew 38 ost th merian htt of Physics ut rom a sarmaryeta gop aeeuson by Wark Doe atte ‘eran Pyseal Soi The $500 esate sein ‘hatlscusin ned operon oss, gost acepance cats. Te URL rth estat ulocuments cer charchvelelconiclther agerdaa01 198/201 T95551 ransparercie