Long term storage of data sets

[JOSEPH H. KIRKBRIDE]

On the 8th of December we published a major article in _Flora_Online_,
and for sometime I have been considering the "half-life" of electronic
publications.  Richard Zander in his e-mail message of 3 December
1992, took the practical point of view, i.e., it is the responsibility
of the libraries and the librarians to maintain electronic
publications.  In the long run, I hope that the libraries and
librarians face up to the responsibility.  At the present time, my
impression is that they are not
 doing so because the volume is so low
that they do not want to expend limited resources on it.

As the era of electronic publishing begins, the entities putting out
electronic journals are going to have to shoulder the burden of
guaranteeing that the contents of their electronic journals are
effectively archived in SEVERAL places in a form that will not degrade
before the libraries start to assume the responsibility.  I would
recommend that CD-ROM copies of all electronic journals be deposi
ted
in at least two separate institutions.  The write once-read many times
technology has arrived at the point that this is possible for a
reasonable cost.

So far, computer technologies have had a limited life span (the
National Archives in Washington, DC, has tons of computer material
that it can not read because the appropriate equipment is no longer
available).  Something is needed to span the gap of a few years until
the libraries assume their responsibilities towards electronic
jo
urnals, and hopefully the CD-ROM technology will still be readable
at that time.

Joseph H. Kirkbride, Jr.
USDA, Agricultural Research Service
Systematic Botany and Mycology Laboratory
Building 265, BARC-East
Beltsville, Maryland 20705-2350 USA
Voice telephone: 301-504-9447
Internet: JKIRKBRIDE at ASRR.ARSUSDA.GOV