Re: Nature 10 September on Public Archiving

From: Mark Doyle <doyle_at_APS.ORG>
Date: Tue, 22 Sep 1998 13:15:18 -0400

On Fri, 18 Sep 1998, Albert Henderson wrote:

> King's work on the economics of scientific journals was an enormous
> undertaking, financed largely by the National Science Foundation at
> a time when "dissemination" was recognized by science policy as
> important. It gave us benchmarks like the distribution of costs
> described above. I feel this is a key question.

But the market conditions have changed considerably over the last 20 years
since the study. Time for a new study. You conveniently ignored my point
that a lot of the reader costs you claimed would go away with an electronic
environment and that the size of the costs to be divided will also be
significantly reduced.

> My understanding of the allocation envisioned by Harnad and the
> Association of American Universities crowd would probably be:
>
> 30% authors
> 70% users
>
> No publishers. No libraries. Do you agree with it or do you dispute it?

I dispute it. First of all, you can't just quote percentages without
putting them in context. I don't know what the exact breakdown of costs will
be. I haven't looked at the King study, but I presume most of the cost to
readers was in their time spent looking for and retrieving information.
Since a researcher's time is quite expensive and there are many, many
researchers, this both increases the total costs and inflates the percentage
of costs attributed to readers in the analysis. Clearly the current paper
distribution system feeds this waste of time. Going electronic has the
potential to eliminate most of the costs attributed to readers and thus
greatly reduce the percentage of costs they have to bear and the total size
of the cost burden.

As for publishers, they can try to continue to do things in a costly
manner, but I think authors, readers, and libraries will reject such
publishing. Libraries will have to bear much reduced costs for their serials
collections (less expensive subscriptions, less shelf space, less climate
control, etc.) and can focus more on helping researchers use the electronic
medium to their benefit and build better collections of books which are
better suited to print distribution. You make it sound like libraries exist
solely to distribute serial journals.

> As I read Harnad, electronic journals will be financially
> supported by authors' page-charges and will be otherwise
> free.

The basic distribution fits that model. If there are other services (better
searching, better linking) that can't be easily automated, then there is
room for a publisher to step in and create value-added services that they
can sell to libraries and users.

> Users would have to have their own equipment and supplies
> paid by tuitions, grants, employers or out of the users'
> own pockets.

They already do this....

> The elimination of libraries' spending is a huge
> carrot to university managers who want that money
> for their own reasons (I suspect real estate
> development, ever more impressive back office
> empires, interest-free loans, travel, and so on).

It is funny the most librarians don't seem to agree with you. They want to
spend money on better staffing and non-serial collections because they think
this is a better way to spend their resources in aiding research.

> Yes, as I interpret Harnad, libraries and librarians will be gone.

Wrong. They will play a modified role.

> Authors will pay for distribution, presumably with page charges
> financed by their grants or employers.

Authors will pay reasonable fees that cover peer-review and a much cheaper
form of distribution than in the print model.

> Users will pay for the
> technology required for 'free' access.

They are already paying for that as part of their research environment. And
they pay similar costs today in photocopying charges (not to mention
expensive journals). They may pay for value-added services that go beyond
simple distribution and automated indexing.

> Universities will spend the
> "former" library money to grow their bureaucracy a la Max Weber's
> widely recognized analysis of administrative behavior.

Doubtful. More likely they will be able to invest the money in better book
collections and better electonic facilities for researchers and libraries.

> ah> - It would eliminate the dissemination services - primarily
> ah> organization, presentation, and procurement -- provided by publishers
> ah> and libraries. Why would researchers favor this?
>
> md> None would of course and no one is calling for this. Another strawman.
> md> Publishers and other entities could still provide value-added overlays
> that md> serve as guides to the literature. They just wouldn't be tied to
> the md> delivery of the documents itself (though if someone wanted to buy
> paper then md> a publisher could easily print and bind it and sell it
> without affecting md> the cost to those who like free electronic access to
> the literature).
>
> You are not calling for this, but it is inevitable if the
> AAU cabal has its way. They have abandoned 'dissemination'
> as a part of science policy (as I describe extensively in
> the current issue of SOCIETY).

Don't know what you wrote. My view of the AAU proposals (which do have some
problems) is that they desire to increase availability of research
information by removing constraints imposed by the current print model of
distribution and copyright while reducing the amount of money being funneled
into corporate profits.

> They have attempted to set commercial and nonprofits, print
> and electronic, against each other. Divide and conquer.

Hmmm, I think those dichotomies have existed for quite a while with or
without the AAU proposal.

> If they succeed, they will stand alone with a monopoly on
> knowledge (if there is any knowledge left to monopolize).

You are fond of gross exaggeration, aren't you?

> I think your ideas are different from those of Harnad when it
> comes to the money. The money is important. Please think it out.

A little divide and conquer of your own, eh? Well, you are wrong. I think
you purposefully mischaracterize what Harnad is proposing. You can go back
and read my other contributions to this forum. There are real costs involved
in today's publishing environment that are present mostly because we aren't
taking full advantage of the electronic medium from author all the way to
reader. If we were to switch immediately to the proposed economic model, it
is likely something valuable would be lost because publishers like the APS
with large publications can't turn on a dime and because we do other things
like create long-term archival versions of articles which don't necessarily
play a direct role in disseminating an article today. But both Harnad and I
agree that such a transition is inevitable and possible, will benefit the
research community in the long run, and needs to proceed orderly so that
publishers who are willing can adapt.

> md> They were still maintaining a dual system -- hardly optimized. Having a
> md> paper deliverable doesn't preclude free electronic distribution. It
> should md> also be noted that a sizeable portion of the Astrophysical
> Journal (and the md> rest of the astrophysics literature) appears on xxx in
> the astro-ph archive md> which has grown about 400 manuscripts per month).
>
> Based on the needs of their community, I think
> they could reach no other conclusion.

It is the dilemma of current publishers that it is rather difficult to take
a mature publication that is tied to print and recast it into an
electronic-oriented publication. We are saddled with having to maintain dual
systems for the transition. However, this doesn't address new enterprises
(The New Journal of Physics is going to be a quite interesting test of all
of this and we are already doing Phys. Rev. Spec. Topics in Accelerators and
Beams, not to mention journals like JHEP and ATMP which have some maturing
to do, but are on track to compete effectively). Nor does it say that such a
transition is not worth doing. We view the costs of the transition period
as an investment in the future.

> I don't know that xxx has any significance here
> other than to provide a useful mode of informal
> exchange. It is complementary I think. We cannot
> call it a substitute for Astrophysical Journal,
> can we?

It certainly substitutes for many roles that were formerly exclusive to
ApJ. xxx is the primary means for which new results are circulated. Its
search engine searches across many journals so it can be a more useful place
to search. Researchers feel quite comfortable retrieving papers from there
(and if they have the Journal-ref stamp of ApJ, so much the better). There
is nothing that precludes xxx from being more than an informal channel of
communication. The point is that the peer-review aspects of ApJ can be
disentangled from the dissemination aspects. This wasn't true in the pure
print world.

> Here is what discourages librarians:
>
> Twenty ARL libraries received less money in FY1997 than they
> did the prior year. Similar numbers the year before. This has
> gone on for 30 years, long before the internet and email.
> I call THAT discouraging, particularly as research work product
> continues its inexorable increase.

Surely they lament having to spend larger and larger portions of their
budget (whether shrinking or not) on serials collections which increase in
price much faster than inflation? Anyway, I'll let the librarians speak for
themselves.

> How is xxx in the market economy? It is free to everyone but
> the government which, in its wisdom, could abandon it just like
> it dropped so much solar energy.

Perhaps, but xxx is a heavily used service and the amount of money involved
is fantastically small compared to the rest of the research spending. The
outcry would be quite vocal if xxx were to lose its funding (which now comes
from multiple gov't sources so it is more secure). It is completely
different than solar energy. Anyway, xxx competes in the market of
disseminating information. True, it isn't commmercial, but so what? It is
still a market force which publishers have to contend with.

> This is absolutely incorrect. British economist David J Brown
> (Electronic Publishing and Libraries. Bowker-Saur 1996: 41-42)
> points out the 'market' problem arises because libraries'
> finances are not controlled by anyone interested in the
> dissemination of knowedge. Agencies that finance research
> do not finance libraries (although they supposedly
> contribute via research 'overhead' payments; these are
> under the complete control of university managers).

That is one take on it. I disagree with your implied conclusion. Look, if
researchers (who have the highest interest in seeing their work
disseminated) had to oversee library budgets, they would have long ago
realized that the system is broken and would have taken measures to avoid
feeding it (submitting to less expensive journals, starting new journals,
etc.). I think we both agree that there it is problem that those with the
largest interest don't control the money. But I am afraid you are wrong if
you think this would lead to ever increasing library budgets to buy ever
increasingly expensive serials.

> Weberian dynamics, moving the bureaucracy to cannibalize
> library money for its own ends is described by Crawford and
> Gorman (Future Libraries. American Library Assn. 1995).

Perhaps it is true that university bureaucracies squeeze libraries. But I
think you are wrong if you think that unfettered librarians would just
continue to spend and spend and spend on serials...

> As we envision possible futures, take the Weberian ideas a
> step further. If one gives the university an interest in
> authors' exclusive rights, what will stop them from attempting
> to exploit them for profit -- as they try to do with patents?
> They monopolize research grants already in the US. Basic
> research yields only journal articles, rarely a patent. I am
> convinced they want it all.

You mean they want publisher profits? There will be much less profit (if
any) involved. And I think their motivations are different than your
maximally cynical take on it.

> Why has [Phys. Rev] remained in print and not gone totally electronic?

Because the transition takes time and since we compete in a marketplace we
can't act always ideally. Furthermore, the transition is a global one that
involves more than just publishers. We are starting new journals that are
totally electronic and I predict our journals will be primarily electronic
(print would be a secondary product) within 5 years. PRD is already close to
that.

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