Re: Nature 10 September on Public Archiving

From: Mark Doyle <doyle_at_APS.ORG>
Date: Thu, 24 Sep 1998 14:16:07 -0400

On Thu, 24 Sep 1998, "Albert Henderson <NobleStation_at_compuserve.com>" wrote:

> On 22 Sep 1998 Mark Doyle <doyle_at_aps.org> wrote:

> Don't you think you should consult the King work before disagreeing with
> it?

I didn't disagree with it - I questioned its validity after 20 years with
the rise of all sorts of new technologies. But, yes, I haven't read it and
was relying on your characterizations of it.

> King says that reading -- not acquisition, copying, etc. -- represents
> over half of the total cost of journals communications. That was when
> the reader did not need the equipment, supplies, energy, space, and
> connectivity essential to your idea. I don't think you can claim that
> working in the library is equally expensive.

Well, then I think the study has even less relevance. Reading and
comprehending articles will always be a dominant cost because readers spend
time doing it. This is unrelated to the means of distribution. Anyway, this
agrees with my statement that it is researchers' time that is valuable.
Electronic access greatly reduces the time spent retrieving articles and
effective searching and linking help researchers focus in more quickly on
the articles that they will be spending their time reading. Have you ever
tried to do research using the standard paper process? It means walking back
and forth between different stacks (or even buildings -- will only get
worse if we expand paper serials collections exponentially) constantly going
back and forth between primary and secondary resources (like citation
indexes). Then when you find what you want, you have to stand there and
photocopy tens or even hundreds of pages (often of poor quality because of
the binding preventing you from laying the book flat). This is a model of
inefficiency and these time-wasting things are expensive because it is
researchers time that is expensive. Electronic efficiencies that save time
will more than compensate for the costs of electonic access.

> I have seen no evidence -- and I use electronic means whenever I can
> because I am not connected to a huge library collection -- that I am
> getting more faster and cheaper.

Are you doing active research in science? I have done active research both
using the old system and the new electronic ones, and am I fully satisfied
that the new way saved me time and made me more productive.

> Systems are not only bugridden, they
> want upgrades constantly to feed their makers.

Unwarranted generalization....Electronic journals and xxx are quite easy to
use and work virtually error-free.

> My 5 1/4 disks are not
> only obsolete, I find many have lost their data in less than 10 years!

Data used on a regular basis and in high demand will always migrate. Costs
for storage are constantly dropping. And I have repeatedly made the point
that I view one as one of the APS's most important functions is to create
preservable, archival electronic versions of the material we peer-review,
something which needs to be retained through the transition (preferably by
giving authors tools so that such preservation is incorporated right from
the beginning).

> Libraries are the main economic support for academic journals

They don't have to be.... That is what this forum is about.

> Libraries object to the price of automated databases as much as they do
> journals. You should hear them on Chemical Abstracts Service!

I think you misunderstand my use of the term "automated". I meant things
like xxx where everything that can be reliably automated, is automated --
hence there is a one-time development cost, but the feature itself isn't
charged for. Building of search indexes. linking of articles, notifications,
and format conversions are some examples of things that are relatively
straightforward to automate. Hierarchial classification (a la Yahoo) is
something that isn't easily automatable -- either you put the burden on
authors to classify their own work or you can create a value-added service
that can be sold (hopefully for reasonable costs) to interested parties.

Electronic procedures could drastically reduce the cost of many services
that I am sure librarians consider expensive. For instance, to compile the
Science Citation Indexes, ISI scans the paper journals, does OCR, manually
corrects the results, and then publishes the information. We are working
with them to send them an electronic feed so that all of this expense is
bypassed. Do you not see the virtue of such things?

> Many databases have responded by capping their coverage. How does that
> fit your model?

It doesn't -- if they can't do it efficiently, someone else can step in and
do it better.

> ah> Users would have to have their own equipment and supplies paid by
> ah> tuitions, grants, employers or out of the users' own pockets.
>
> MD> They already do this....
>
> Some do. Many don't. Libraries have always prided themselves on their
> lack of an economic threshhold for patrons. No fees and equal access to
> rich and poor. In some cases, it is mandated by law. Your vision
> doesn't do that, does it?

Sure it does. Libraries can provide centralized electronic access and
knowledgeable librarians can help patrons navigate to find the information.

> Some librarians don't get it.

Many do... And getting it doesn't mean agreeing with you.

> MD> Wrong. They will play a modified role.
>
> Yes? Please expand.

There is a whole spectrum of roles that libraries can play. I mentioned
above about providing centralized electronic access and for help in
navigating and searching to find information. Other libraries are developing
experience with archiving electronic information instead of paper. Their
role as collector of paper serials will diminsh, but they will expand their
roles as classifiers, searchers, and navigators.

[cynical diatribe deleted]

> Why not put libraries back into their central economic role?

Because it is a grossly inefficient system that cannot be sustained, even
if universities were to lavish them with money. Traditional paper libraries
can't grow exponentially in step with the exponentially growing literature.

> If universities restored the conservation of knowledge to their mission
> and provided the market essential to support publishers of research,
> A&I services, reviews, handbooks, etc., the conversion to an
> electronically-oriented system would be a lot smoother.

Hmm, this is the first time you have implied that an electronic-oriented
system is desirable. So you favor it, but only if publishers can maintain
their status?

> As it is, publishers must replace subscription revenue with document
> delivery fees, licenses, etc., etc.

Only to maintain the status quo.

> First, it is groups of researchers who usually propose journals to
> publishers -- not the other way around. Most proposals that I have seen
> represent the failures of existing publishers to recognize budding
> "twigs" on the tree of knowledge. Association politics ... well I
> shouldn't have to draw a picture.

Sometimes researchers take cost-effectiveness into account, sometimes they
don't. A concrete example: PR-STAB was started because accelerator
physicists approached the APS in part because they felt that journals in
their field were too expensive and this was adversely affecting distribution
of their work. I believe if more researchers were more in tune with costs,
they would make more decisions like this. Similarly, researchers who were
unhappy with the current costs and limited access started journals like ATMP
and JHEP. More examples would be the Optical Society's Optical Express and,
no doubt, the New Journal of Physics.

> Second, why didn't university administrators ask faculty senates if it
> would be OK to cut library spending from 6% to under 3% of their
> budget?

Because it wasn't a single step? Because it isn't clear that library
budgets must scale with the full university budgets? Because library
subscription costs increase at a much faster rate that inflation and there
has to be a braking mechanism somewhere?

> Third, the exponential growth of science activity and its work product is
> a matter of record. Suddenly, with the introduction of the Xerox
> 914, we have legislated "library photocopying" to aid dissemination.
> It backfired, giving us the "serials crisis" instead.

Hardly. No paper library can grow exponentially. Decreeing that all
libraries must grow exponentially is inefficient and unsustainable.

> I think they should have kept buying books, too. I don't think
> the curve of library growth should be different from that of
> research activity.

See above.

[more cynical diatribe deleted]

Mark Doyle
APS Research and Development
Received on Tue Aug 25 1998 - 19:17:43 BST