I fear my esteemed colleagues are having trouble keeping their eyes on
reality, which is what I was attempting in my remarks. It is all very
well to prognosticate about 10 years or 30 years down the road, but
what is actually out there right now, and what will be out in the next
year or two? For example, people have long talked about "overlays" on
the xxx archives, and finally there are some examples - let's see what
they do:
1. JHEP (
http://jhep.cern.ch/) referees and copy-edits xxx e-prints,
and provides print and CD-ROM distribution for the papers it selects.
It also provides automatic searching and indexing, thus satisfying the
roles 2-5 I discussed with considerably more automation than most print
journals. This isn't a bad start, but the copy-editing is clearly slow
and labor-intensive (some papers accepted in June have yet to be
"published"). And the funding model has yet to be worked out - aside
from the paper/CD subscriptions the journal is subsidized by grants for
now.
2. ATMP (
http://www.intlpress.com/journals/ATMP/) referees xxx e-prints
with little or no copy-editing (as long as the manuscript is in
Latex2e). The e-journal is currently available free, and consists
simply of a table of contents with links to xxx. The journal does
charge for a print copy. It's not clear to me what the finances are,
but eventually it looks like they'll charge for access to the
e-journal.
JHEP seems to be on course to publish about 300 papers per year, ATMP
about 60. JHEP charges libraries $360 and ATMP $300 for the print
edition, or $1.20 and $5 per article, respectively. By way of
comparison Physical Review D (one of our journals), publishes almost
1700 articles in a year in roughly the same subject areas (and links to
the xxx archive in the process), and charges libraries $2280, or about
$1.34 per article, not much different from the JHEP price, and
considerably lower than ATMP. Of course, PRD has quite a large
subscription base so those numbers do not compare total cost per
article - on the other hand, we don't have cost numbers from the other
journals either. At the least, these "electronic-only" journals are
not giving libraries a huge price cut now.
Furthermore, both "overlay" journals have print versions (which they
rely on for revenue), so it is hard to truly call them
"electronic-only". Since July 1 1998, PRD could also be called an
electronic journal with an available print distribution mechanism:
libraries are charged one price for access to both print and electronic
versions, the electronic versions of published articles appear long
before they are printed, and are citeable by a volume and article ID
(replacing the page number), something that is not true of JHEP and
ATMP articles before final printing (both assign a final page number
that is not known until printing).
And PRD is far from alone - many established journals are making
significant innovations in online publication that puts them well ahead
of what can be done "on the cheap". Britain's IOP, the Astronomical
Society, the ACS, Highwire Press, even our friends at AIP
(
http://ojps.aip.org/) have done many interesting things with article
inter-linking, searching and presentation, etc. Many are bringing older
material online as we have done with PROLA (
http://prola.aps.org/). All
these things provide significant value to the reader. BUT - a reader is
unlikely to pay for them alone. How many people will subscribe to ATMP
when they know every article in there is available (in identical form)
on the xxx site? Especially if they can just search for "ATMP" or some
such in the article abstract to find ATMP papers. Providing
distribution for the articles is, as Mark Doyle mentioned, a minimal
additional expense when you are already doing all those other things.
And as far as providing a single location to look things up, that is
the primary role of abstracting/indexing services, and they will also
evolve to continue fill that role (again there are electronic and
automation efficiencies that will help).
The ideal of page charges is fine - until you run into the reality
mentioned elsewhere by Professor Hurtado - many authors do not have
funds to pay the $500 or whatever it costs for each article to be
published (or even considered for publication with submission charges)
and some funding agencies even forbid it (most US science funding
agencies did not allow paying page charges out of research funds a few
decades ago).
What am I trying to say here? Harnad, and to a lesser extent Doyle and
Walker, advocate revolution, but market-based evolution is already upon
us. The efficiencies made possible through automation and electronic
distribution are available to all the players, and while legacy systems
take time to respond, in the end the bigger players are the ones most
likely to have the momentum and capital resources to take full
advantage. And, like in other areas of the economy, the new efficiences
will eventually play themselves out in lower prices and costs for all.
Arthur Smith (apsmith_at_aps.org)
PS I attempted to change the subject heading because "30% or 70%" had
little to do with what I had to say, and seems an irrelevant question
anyway when different non-electronic publishers already seem to have
factors of 10 differences in costs.
---
[Header retained: 30/70 is the topic thread and its incorrectness
or irrelevance is the comment. - Ed.]
Received on Tue Aug 25 1998 - 19:17:43 BST