Mathematical Journals Past, Present and Future--A Personal View

Document Sample
Mathematical Journals Past, Present and Future--A Personal View Powered By Docstoc
					    Mathematical Journals:
  Past, Present and Future—
             A Personal View
                                                                                         Donald Babbitt

As most people in the mathematical research                  Until a few years ago these too were paper jour-
community are aware, there is a revolution tak-              nals which also grew exponentially in size with
ing place in the way information—in particular,              the literature being reviewed, as librarians can
mathematical research—is being produced and                  readily attest. Beginning in the early 1980s, MR
disseminated throughout the world. I am, of                  and ZBL became available online on a “per hit”
course, referring to the electronic revolution               basis. (This means a user is charged for the
taking place in scholarly publishing. In this ar-            amount of time he/she is “online” and for the
ticle I will give my personal view on how this is            number of search results printed out.) In 1989
affecting and will affect the publication of math-           MR became available on a searchable CD-ROM
ematical journals.                                           (MathSciDisc). These were just an inkling of the
                                                             electronic revolution to come.
The Past                                                        Finally about twenty years ago Donald Knuth
For the past two hundred years original math-                invented the highly flexible typesetting/word-
ematical research was disseminated primarily                 processing language TE X that was especially
through refereed paper journals. In addition,                suited for typesetting mathematics. With a sig-
during the past twenty-five to thirty years many             nificant input of resources by the AMS this
authors disseminated preprints of their papers               typesetting language became the standard for
via mail so that colleagues would be apprised of             publishing mathematics, especially for jour-
their new results a year or more before they
                                                             nals. One very important outgrowth of this
were published. The libraries collectively archived
                                                             was that mathematicians themselves were able
the journals by basically keeping them dry and
                                                             to typeset their own papers, and a surprisingly
accessible to the mathematical community, just
                                                             large number of mathematicians chose to do
as they had been doing for centuries with the
                                                             so, often with great virtuosity. (This virtuosity,
world’s scholarly literature. By utilizing interli-
                                                             however, has the down side that journal pub-
brary loans, the libraries were able to make vir-
                                                             lishers often have to significantly modify such
tually the entire research literature available to
                                                             author-prepared manuscripts before they can
the mathematical community.
   During most of the twentieth century journals             be used to produce the journal.) But a more im-
such as Mathematical Reviews (MR) and Zen-                   portant byproduct was that papers written in
tralblatt fur Mathematik (ZBL) have aided math-              TE X were now electronic files which could be
ematicians in finding and evaluating the expo-               transmitted over the Internet.
nentially growing amount of research literature.
                                                             The Present
Donald Babbitt is publisher of the American Math-            Although most journals remain paper only, many
ematical Society.                                            now are published in dual format, i.e., there is
Although the predictions made in this article are strictly   both a paper and an electronic version. The lat-
my own, the lively discussions of the AMS Committee          ter typically is available over the Internet, either
on Publications have helped me immeasurably in fo-           as a free add-on to the paper subscription or with
cusing on these issues. I am also sure that several past     a modest premium. In some cases one can sub-
and present members of the Committee will disagree,          scribe only to the electronic version for some-
probably strongly, with some of my prognostications.         what less than the paper subscription. The elec-
                                       —Donald Babbitt       tronic versions vary in their level of functionality.

JANUARY 1997                                                 NOTICES   OF THE   AMS                                  29
     Some are only an electronic version of the printed      server for all fields in January 1995. The vast ma-
     copy, viewable on screen in TEX code and down-          jority of preprints posted on these servers are
     loadable and printable in one or two formats.           still destined for refereed journals. Perhaps the
     Others have additional functionality, such as           most important preprint server for mathemat-
     being hypertext-linked HTML documents with              ics, at least as a model, is one that is primarily
     real mathematics on screen (as opposed to TEX           outside of mathematics but in the closely related
     code), which, for example, can link MathSciNet          field of theoretical physics: the Ginsparg preprint
     subscribers to the reviews of papers in the bib-        server at Los Alamos. Despite having strict for-
     liography. In many of these journals there is the       mat requirements, this server has an extraordi-
     opportunity for readers to post comments to go          nary number of preprint submissions and an
     along with the electronic version of the article.       equally extraordinary amount of traffic. It is di-
        A new form of journal publishing in math-            vided into various fields, such as quantum field
     ematics began in 1993: free electronic-only jour-       theory/string theory, condensed matter physics,
     nals on the Internet. Papers in these journals are      and now even some mathematical fields of spe-
     refereed and are reviewed in MR and ZBL just as         cial interest to theoretical physics such as alge-
     the papers in the traditional paper journals. The       braic geometry. The NSF has thought enough of
     hardware, software, and system administration           the Ginsparg preprint server that it is investing
     for the journal are provided typically by the           more than a million dollars in supporting fur-
     mathematics department of one of the principal          ther developments of the preprint server as a tool
     editors at no cost to the journal. (The Electronic      for disseminating scientific information.
     Research Announcements of the AMS [ERAAMS],                 As mentioned above, MR and ZBL had already
     created in 1995, is somewhat of an exception to         developed electronic access to their databases
     this model.) Authors submit TEX files of their ar-      by the early 90s. What the mathematical com-
     ticles to the journal in question following fairly      munity began to ask for, especially with the ar-
     flexible format restrictions; the journal editors       rival of the World Wide Web (WWW) and so-
     arrange for the refereeing of the articles and, if      phisticated browsers such as Mosaic and then
     accepted, provide the editing of articles for pub-      Netscape, was to have unlimited access to the MR
     lication. Consequently, all of this is done by free     and ZBL database at their desk and even at home.
     contributions of material and labor, which is           This possibility was realized for those math-
     why these journals can be made available for free.      ematicians affiliated with institutions that sub-
     Simplifying somewhat, the model here is au-             scribed to MR with the introduction of Math-
     thor-to-editors-to-readers with no intervention         SciNet in January 1996. This is a user-friendly
     of a publisher. It should be mentioned that the         hypertext-linked online version of MR with a
     functionality of most of these journals is rela-        powerful search engine and with the capability
     tively minimal. Almost all of them now dissem-          of browsing as one can do with the paper ver-
     inate their journals from their own Web sites.          sion of MR. A subscription is based on site li-
        The AMS has just begun to publish electronic-        cense, so that all individuals—including the
     only specialty journals with the rather modest          mathematicians, computer scientists, physicists,
     subscription prices of $125 for institutions and        etc.—at the site can have unlimited access to the
     $50 for individuals, with a lengthy free sub-           MR database. One important caveat is that this
     scription period (Representation Theory, An Elec-       mechanism for delivering the MR database de-
     tronic Journal of the AMS; and Conformal Geom-          pends on the robustness and speed of the World
     etry and Dynamics, An Electronic Journal of the         Wide Web, something that is not always present
     AMS). At the present time these journals have a         even in the U.S. Because of this, there still is
     higher level of functionality than the free e-jour-     considerable demand for both the paper and
     nals, with the exception of ERAAMS.                     CD versions of MR and ZBL.
        As already mentioned, preprints have been                Document Delivery has been around for many
     around for some time in the mathematical com-           years and, until recently, has been a comple-
     munity. In the early 90s, however, preprints            ment to the interlibrary loan service. This is a
     went electronic with a vengeance. Preprint              service provided by several organizations, in-
     servers were set up in several departments by           cluding the AMS (MathDoc), that obtains copies
     mathematicians interested in specific areas, such       of journal articles to which a library or scholar
     as mathematical physics, K-theory, and alge-            otherwise would not have access. This is done
     braic geometry. Mathematicians in these areas           for a fee depending on the journal, with a por-
     were invited to post their preprints at these           tion of the fee being paid to the journal in ques-
     sites. There were also preprint servers such as         tion. In the past two to three years, a small num-
     the one at MSRI which posted preprints from             ber of academic libraries have cancelled a
     many fields, although the preprints themselves          significant number of journal subscriptions, in-
     mostly arose out of workshops, etc., that had           cluding some to mathematics journals, and have
     taken place at MSRI. The AMS set up a preprint          replaced these subscriptions by the commit-

30                                  NOTICES   OF THE   AMS                               VOLUME 44, NUMBER 1
ment to their patrons to obtain, via Document           • In the next few years much of the important
Delivery, any article from these cancelled jour-          preelectronic research literature in journals
nals that they need. These libraries have found           will be converted into electronic files. At a
that by and large they save a considerable                minimum, articles in these journals will be
amount of money, while, of course, the pub-               electronically accessible from major libraries
lishers receive considerably less revenue from            and will have hypertext links to the review-
the Document Delivery fees for the articles de-           ing journals (expected to be purely elec-
livered than they would have had the subscrip-            tronic in five to six years). The first step in
tions not been cancelled. This is being widely dis-       realizing this gigantic task has recently been
cussed in the library community.                          undertaken by JSTOR, an independent non-
                                                          profit corporation established with the as-
The Future                                                sistance of the Mellon Foundation. They are
In my prognostications below, I will not attempt          currently creating electronic files for all of
to go beyond ten to twelve years. I also will be          the articles in the pre-1991 issues of several
assuming that there is a robust WWW or a suc-             major journals, including the AMS primary
cessor in place. Despite the increasing occurrence        journals.
of bottlenecks and time delays for many users
                                                        • Those journals that have not made credible
of the current Web, I think the huge and in-
                                                          arrangements for their electronic archiving
creasing investment in the Web by Microsoft,
                                                          in perpetuity will probably not survive. By
Netscape, and others makes this assumption
                                                          electronic archiving in perpetuity I mean
not totally unrealistic.
                                                          that as electronic formats evolve over time
  • After ten or so years, all primary research           the electronic files of the journal’s articles
     journals will have highly functional elec-           will be changed accordingly so that they
     tronic versions, and almost all of these will        will always be electronically readable and
     be electronic-only journals. Only the most           easily accessible (not necessarily for free) to
     selective general journals will survive, while       the mathematical community.
     the remaining ones will be specialty journals.
                                                        • Many if not most of the free electronic-only
     The electronic files of the articles will have
                                                          journals will either disappear or become
     a standard structure (SGML?) with varying
                                                          subscription-based journals in the next
     degrees of functionality and will be part of
                                                          decade. Some of the main reasons for this
     the “universal” database of the mathemati-
                                                          predicted demise are: most of the free jour-
     cal research literature. The files of individ-
                                                          nals receive a substantial subsidy by way of
     ual articles will be structured so that they
                                                          equipment, system support, etc., from a uni-
     are technically accessible from the review-
                                                          versity or similar institution, an inherently
     ing journals (by now solely electronic); their
                                                          unstable situation; the required enthusiasm,
     reviews will be technically accessible from
                                                          time, and energy of the founding editors
     the article; and the articles and reviews of
                                                          are very difficult to pass on to a second
     the articles in the bibliography, when they
                                                          generation of editors, especially with the
     are in a reviewing journal’s database, will be
                                                          additional work required of editors of free
     accessible from the article. By accessible I
                                                          journals; although authors seem willing to
     mean readable online in real mathematics
                                                          produce usable TEX files, especially if they
     and downloadable and printable in the stan-
                                                          are given considerable leeway in format,
     dard formats of the time. I use the word
                                                          they almost certainly will be much less in-
     “technical” because there may, and often
                                                          terested in producing the standardized and
     will, be barriers to access due to subscrip-
                                                          highly structured files required for the up-
     tion requirements.
                                                          coming electronic milieu; making arrange-
   Above I have predicted that only the most se-
                                                          ments for electronic archiving will be a spe-
lective general journals will survive. (I obviously
                                                          cial challenge for these journals.
have been very careful not to name them.) This
has less to do with the electronic milieu than with     • There is the question of the cost of elec-
Document Delivery. Because most of the articles           tronic-only journals versus paper or dual
in the other general journals have very limited           journals and the economic effect on libraries
readership, libraries will realize, as some already       when the transition from paper and dual
have, that it is much more economical to utilize          journals to primarily electronic-only sub-
Document Delivery than to subscribe to such               scription-based journals takes place. For
journals. These journals will then become eco-            both print and electronic journals the most
nomically unviable. The electronic-only specialty         significant value-added (to the authors’ con-
journals should avoid this fate, because they             tributions) is provided by the journal editors
will be quite inexpensive, with many individual           and referees. This is oftentimes free, but
subscribers.                                              not always: publishers will frequently pay for

JANUARY 1997                                          NOTICES   OF THE   AMS                                31
                                                                                                                            secretarial assistance, provide computing
              American Mathematical Society                                                                                 equipment, pay for reduced teaching time
                               Communicating Mathematics
                                                                                                                            for managing editors, etc. For a print jour-
                                                                                                                            nal the additional value-added provided by
                           “…rated among
                                                                                                                            the publisher consists of paper; printing;
                           the top 5% of all
                                                                                                                            distribution, including postage; some key-
                           sites on the
                                                                                                                            boarding; editing for language; and appro-
                                                                                                                            priate paper journal format. For an elec-
                                                                                                                            tronic journal the cost of the paper; printing;

                                                        Electronic                                                          distribution, including postage; some key-
                                                                                                                            boarding; and editing files for appropriate
                                                                                                                            format of the paper journal is replaced by
           Delivery of AMS                                                                                                  the cost of hardware, software, and system
                                                                                                                            support to produce and distribute the elec-

             Products and                                                                                                   tronic journal, and editorial work to pro-
                                                                                                                            duce an appropriately structured file from
                                                                                                                            the authors’ submissions. There could be in
            Services on the                                                                                                 some cases additional costs for creating in-
                                                                                                                            teractive files and elaborate graphics. It

                                                 About the AMS

                                              Society Information

                                                                                                      World WideWeb         seems reasonable to expect that the steady-
                                                                                                                            state cost of producing an adequately func-
                                                                                                                            tional electronic journal compared to the
     Society                                    What’s on e-MATH                    The AMS
                                                                                   Bookstore                                cost of producing a print journal of com-
                                              What’s New in Mathematics
                                              Frequently Used Services
                                              Complete List of Services
                                                                                                                            parable “size” could be significantly less
                                      S s
                                                Electronic Publishing              Books, Journals,
                                                                                                                            than the latter, provided almost all of the
                                              —Mathematical Reviews on the Web
                                                                                        and more
                                                                                                                            work needed to create the files for produc-
      Math Digest
                                              Electronic Journals
                                              AMS Bookstore
                                                                                  Sponsor                                   ing the journal is done by the authors, edi-
      Math News
                                                e -MATH by Subject
                                                                                  Happening                                 tors, and nonproprietary software. If the
                                                                                  in the

      Ethical Guidelines
                                              AMS Publications
                                              Science Policy
                                                                                                                            electronic journal has state-of-the-art func-
                                              Professional Information
                                              & Services
                                                                                                                            tionality and/or puts minimal burden on
       Secretary's Page
                                              Meetings & Conferences
                                              Mathematics on the Web
                                                                                                                            the authors, then the cost could be higher,
                                                                                                                            although probably less than that of a cor-
                                                                                                                            responding paper journal. It seems safe to
                                                                                                                            predict that the cost to libraries of the pri-
                                                                                                                            mary mathematical research literature in
                                                                                                                            the electronic milieu will be less per unit of
                                                                                                                            information than in the paper/dual journal
                                                                                                                            environment that presently exists. The above
                                                                                                                            discussion also implies that revenues and
                                                                                                                            therefore income derived from publishing
                                                                                                                            mathematics journals could decrease sig-
                                                                                 URL:                                       nificantly in the next ten to twelve years. In
                                                                                                                            particular this could impact the activities of                                                                                                 the AMS supported by the income from its
                                                                                                                            journal publications.
                                                                                                                          In the end it will be the authors who will de-
                                                                                                                       termine what the electronic publishing envi-
                                                                                                                       ronment of the future will be. It is they who will
                                                                                                                       decide how much nonmathematical effort they
                                                                                                                       want to put into publishing their mathematical
                                                                                                                       research; which journals will offer a sufficient
                                                                                                                       guarantee of having their work electronically
                                                                                                                       archived in perpetuity; how much, if anything,
                                                                                                                       they feel libraries or individuals should pay to
                                                                                                                       have access to their work; and how much func-
                                                                                                                       tionality a journal (or whatever other electronic
                                                                                                                       publishing model that may arise) in which they
                                                                                                                       publish their research should have.

32                                                                                   NOTICES            OF THE   AMS                                VOLUME 44, NUMBER 1