Electronic archiving

["Peter B. Boyce" <pboyce@aas.org>]

Pardon me if I chime in here to add some thoughts based upon our
experience in astronomy.

The electronic versions of all dual-version journals are diverging rapidly
from the paper version. In astronomy we have color plates on the Web and
B&W in print -- for cost reasons, since authors pay the color costs.  We
have video clips, we have machine readable data tables which do not even
appear in print any more, and the list is growing.  What with the
hyperlinks to references and citations which provide immediacy to the
information web, of which the journal is just a part, archiving in paper
is not adequate, even now.  The trend to electronic-only features is
accelerating. In another year, saving the paper will not be considered by
anyone to be an adequate archival strategy, even as a backup. Let's give
up talking about this now.

Don't take this wrong, but libraries can physically not maintain an
electronic archive for all their journals. An electronic journal is not a
collection of individual articles any more.  It is a whole complex system
of files, software, and protocols -- which are different for each
publisher.  One year of our Astrophysical Journal comprises about 60
GBytes in over 250,000 files, and a multitude of scripts and programs, all
of which are needed to have the journal function correctly and completely.

And when XML supplants SGML, and the HTML browsers change and diverge to
the point that the old HTML version will not work any more, who will do
the necessary maintenance. Even laying aside intellectual property issues,
I don't think any library has the expertise to undertake this on a broad
scale. Only the publisher, who presumably will have automated tools for
archival maintenance, stands a chance of keeping the electronic journal
alive and functioning in the future.

This electronic complexity is changing the information environment
radically. Many people do not understand this yet. Librarians who insist
on electronic interlibrary loan, may not understand that one article at a
time, and without links included, will not be enough in another few years.
Publishers who fight interlibrary loan, likewise do not understand that
one article at a time is no threat to ther electronic offerings. The
threat occurs if their electronic offerings are nothing more than PDF
delivery of their paper journal, in which case the publisher will not
survive in the long term anyway. Incidentally we do allow electronic
interlibrary loan of the PDF version of our jornals. The more people who
see it the more likely we are to pick up some new subscribers.

My experience with the astronomical information system, much of which is
mow seamlessly integrated so that backward references, future citations,
searchable abstracts, and massive data files on astronomical objects are
seamlessly linked into a working, distributed data system. We have had
this sytem for three years now, and while the single article still has
it's place in astronomy, no one would be willing to give up the fully
linked information access we now have.  That is why the archival problem
is one of maintaining the whole system -- and this requires cooperation
among all the information providers,. Users may be able to help,
especially the larger libraries who can perhaps, in collaboration with the
publishers, take on some of the necessary work.

We, in the non-profit publishing sector of the astronomical information
enterprise, are headed in this direction. Admittedly we have a good head
start on understand the interlinked information environment and how it
differs from what we were brought up with.  But perhaps our efforts can
help other communities address this problem.  I believe that the solution
must come from the individual disciplines, because hey have the strongest
concern for making the information accessible into the indefinite future.

___________________________________________________
Dr. Peter B. Boyce, Chercheur Associe, 
Centre Donnees de astronomique de Strasbourg, France
               --  boyce@cdsxb6.u-strasbg.fr
             
and Senior Consultant for Electronic Publishing              
American Astronomical Society
              --   pboyce@aas.org    --  http:www.aas.org/~pboyce