I appreciate Peter's cogent and reasonable comments, and would like to add
just one more of my own (hopefully as cogent and reasonable as his):

> One aspect of this discussion, it seems to me, which has been omitted or
> ignored though it has been implied under the rubric of 'stability of
> e-formats' is the long-term access to backfiles.

The advent of online journals has made the whole issue of archiving much
more complex than it used to be.  Ten years ago it was a given that
libraries would provide the long-term archives of journal content; today, in
the online realm, that role logically shifts to publishers, and there is no
reason to expect that they're going to do a good job of it.  After all,
archiving is completely foreign to the publisher's natural role, which is to
get stuff out into the marketplace and let others archive it as they see
fit.  But if publishers are going to sell us online access instead of print
copies, they seem to have stuck themselves with the responsibility of
providing a robust archive.

But at the risk of alienating everyone I haven't already alienated by
dissing print, I'd like to suggest that long-term archiving is neither an
appropriate role for publishers nor an appropriate central mission for most
libraries.  I think it's time, frankly, for librarians to focus much less on
the permanence of our collections and much more on developing the ability to
provide current content with greater speed, ease and transparency to our
patrons.  This can't be equally true for all libraries of all types, of
course; a deep and permanent archive is an integral part of some libraries'
practical function.  But contrary to our automatic professional assumptions,
I don't think such is the case for most libraries, and not even for most
research libraries, which are now in a situation where the ability to give
people what they tell us they need is hindered by a focus on ordering,
processing and storing things that we think they may need in the future.
(For an example of this tendency, think about your institution's
interlibrary loan policy.  If it's like ours, it says to patrons "Oh, you
want Book X?  And we don't have it?  Well, we can't bother the Acquisitions
staff to rush-order a copy -- they're busy processing the approval plan
shipment, which consists of books that we think may be of use to you later.
So we'll see if another library has Book X and then go through the tortuous
and extremely inefficient process of borrowing it from that library on your
behalf.  Hang on for a few weeks, and we'll get back to you.")

I know this sounds crazy, but think about it: what if we could rely on a
handful of institutions (either a few selected libraries or some other,
as-yet-uninvented third party) to house a permanent archive of most journal
content, and the rest of us were freed up to focus on improving our patrons'
access to current information?  We could establish a policy of rush-ordering
copies of the old articles piece-by-piece as requested by patrons, and could
then redirect tons of staff time to actually improving our real-time patron
services instead of investing it in the massive just-in-case gamble that is
our current collecting and cataloging practice.  It has long seemed to me
that the great unspoken scandal of our profession is not (_pax_ Albert and
the BMJ) the fact that we're supporting poor research, but the fact that we
spend such a huge amount of time and money on information that NO ONE EVER
USES.  I think we have a chance to fix that, and it seems to me that _that_
project should be of prime importance to us.

(Hang on, please, while I don my flame-retardant goggles.)

OK, fire at will.

-------------
Rick Anderson
Director of Resource Acquisition
The University Libraries
University of Nevada, Reno      "I'm not against the modern
1664 No. Virginia St.            world.  I just don't think
Reno, NV  89557                  everything's for sale."
PH  (775) 784-6500 x273             -- Elvis Costello
FX  (775) 784-1328
rickand@unr.edu