August 3, 2003 - NYTIMES

In DSpace, Ideas Are Forever
By VIVIEN MARX

The libraries at the Massachusetts Institute of Technology are earnestly
bookish (2.6 million volumes and 17,000 journals) but increasingly digital
(275 databases and 3,800 electronic journals). And just as e-mail dealt a
blow to snail mail, digital archives are retooling scholarly exchange. A
number of universities, from the California Institute of Technology to
M.I.T., are creating ''institutional repositories'' designed to harness
their own intellectual output. M.I.T.'s archive, perhaps the most ambitious,
is called DSpace (www.dspace.org).

Scholarly Storage Traditionally, journals make research public after peer
review, which can take months, sometimes years. Archives like DSpace,
however, collect unpublished work – documents of any length, lecture notes,
photos, videos, computer simulations, blueprints, software – in all
disciplines and make most of it available to anyone as soon as it's
received.
.



Here to Eternity ''Loss'' is propelling the movement. When a grant fizzles,
when a professor resigns, retires or just buys a new computer, work can get
lost. University libraries hope to preserve this material forever – not
exactly a common time span in the digital fast lane, where hardware and
software sunset soon after reaching the marketplace. And unlike library
stacks or hard drives, DSpace won't run out of storage space.

''Everyone has lost something,'' says Ann Wolpert, director of M.I.T.
Libraries, which has designated two full-time librarians to DSpace's
dedicated computers. ''We have already lost NASA data, Census data. Early
digital work is gone because tapes were corrupted or not maintained
properly.''

Who Has It? Soon after DSpace was made public last November, a federation of
universities (M.I.T., Columbia, Ohio State, Rochester, Washington and
Toronto) formed to further the system's evolution and see what it was
capable of – for instance, the University of Toronto wants bilingual
searching. M.I.T. estimates that the free software has been downloaded 3,400
times and is aware of 100 research institutions that are evaluating DSpace
to archive their own faculty's work.

The Journal Backlash Institutional repositories are novel in that much of
their content sidesteps academic publishers, which have come under attack
from the so-called open-access movement. Some scholars complain that
journals delay publication of research and limit the audience because of
their soaring costs. The Association of Research Libraries says library
costs on journals rose 210 percent from 1986 to 2001 – an average year's
subscription might cost $5,000, with some as high as $15,000.

Out of frustration with journals' limitations, some scientists have started
their own archives. This fall, the new Public Library of Science will begin
making peer-reviewed articles accessible free to all online. Stevan Harnad,
a cognitive scientist at the University of Quebec, started a digital archive
for his field in 1997. He says that the subscription-based model ''holds
peer-reviewed articles hostage.'' He advocates that scholars put their work
in online archives first so it can be available immediately and free.

Jeffrey Drazen, editor in chief of The New England Journal of Medicine,
argues that his readers want information ''that is highly meritorious and
rigorously reviewed so that they don't make patient treatment decisions
based on premature findings.'' But he acknowledges the self-archiving
movement and says the journal is rethinking its rules that prevent it from
considering material that has been made public in a digital archive.

Archives like DSpace ''build on a growing grassroots faculty practice of
posting research online,'' says Rick Johnson, a director of the Scholarly
Publishing and Academic Resources Coalition. He doesn't think they are a
substitute for journals but offer ''the best of both worlds: you get the
work certified by a journal and the benefit that provides for promotion and
tenure. At the same time you get your work exposed more broadly than in a
journal alone.''

DSpace Sampler At M.I.T., each academic department exercises quality control
and determines what is to be sent to DSpace. Among current offerings:
working papers on campaign contributions (from the Sloan School of
Management), on amphibious assault ships for the 21st century (from the
Department of Ocean Engineering) and on traffic emissions (from the Center
for Technology, Policy and Industrial Development). M.I.T. expects to have
5,000 items archived by the fall, and will add 7,500 theses later this year.
And that, they promise, is just the beginning.



Vivien Marx is a freelance writer in Boston.