Re: Nature 10 September on Public Archiving

From: Stevan Harnad <harnad_at_COGSCI.SOTON.AC.UK>
Date: Thu, 17 Sep 1998 11:49:54 -0400

Albert Henderson <> wrote:

> sh> To compare citations of e-journals to p-journals is
> sh> simply to compare new-journals to old-journals. What on earth can
> sh> one learn from such a comparison? Look instead at xxx statistics,
> sh> including deposits, hits and citations.
> If you wish to claim value for ejournals, you must use
> recognized measures and submit to reasonable comparisons.

And one must use comparable baselines. Right now, all inferences
about e-journals are really just inferences about new-journals
(and initial reactions to new-media). That's why xxx-usage statistics
are much more representative and informative than e-journal citation
data in 1998 about actual practises and preferences. (Having said that,
e-journal citation data are not doing badly either.)

> Does a 'hit' measure value? Or does it signify a 'browse." These are
> preprints, right?

It signifies USE in the new medium, everything from reading a full
article to checking a reference. But note that I also said citations.
The CITATION rate for papers in xxx is doing very well -- and why
shouldn't it? That's the literature!

No, it's not just preprints any more! Those preprints have since gone
on to become reprints, published in refereed journals, but it's still
xxx that is being used to access them. The depositing authors (100
papers deposits per weekday) of course go on to swap the updated final
draft when there have been nontrivial changes (wouldn't you?). So it's
just contemporary USE of the research literature -- the actual usage,
not the conjectures that are offered below, about reader preferences for
review articles, etc. -- that those "hits" reflect.

One must at and then
contemplate what it means; one cannot just repeat papyrocentric slogans
that are better descriptions of awkward, inefficient, and waning old
ways of accessing and using the literature, than of the powerful,
evolving ways that are on the rise today.

> sh> And hybrid paper/online editions, sustained by S/SL/PPV,
> sh> are just Trojan Horses in this regard, designed to hold us hostage to
> sh> S/SL/PPV forever, if possible.
> The paper editions are sustained by the marketplace, a compelling
> argument for their survival.

Just a little more patience, while the xxx usage pattern continues to
grow and generalise, and the de facto preferences of the user
community become more apparent, as they vote with their eyeballs and
fingers, and the ineluctable library cancellations begin to reflect this.

The marketplace cuts both ways, especially where perestroika is
involved, and what used to be a good (journal sold to subscriber)
becomes a service (quality control sold to author).

The market and indicator changes too: No use reckoning the advantages
of motor cars in terms of savings on hay. After the shift to
page-charges, the S/SL/PPV "market" vanishes like the Cheshire Cat's

> I would say most of the Third World and FSU scientists, for starters,
> are not 'state-of-the-art.' I still see typewriters in use in many
> places in the US.

And you won't find the library shelves very well stocked in the Third
World either. It would be risky to build one's case on which medium
will provide greater access to the serial literature in the Third
World; that might lose ground even more quickly and decisively than if
one ignores the Third World in the reckoning (as the real expenses of
paper journal publishing unfortunately left no choice but to do
until now).

And no one following this debate on their screens will be much
impressed by the typewriter argument; nor are many of the unfortunate
scholars who are still constrained to that technology for economic or
practical reasons going to want to make common cause with forces
conspiring to keep them in that state for any longer than necessary,
let alone holding back others because of their disadvantages!

> Studies like Red Sage and TULIP suggest that the infrastructure is too
> expensive.

This is the late 1990's and we are discussing state-of-the-art projects
and proposals; we mustn't tar it all with the blinkered brush of
ill-conceived misadventures like Red Sage and TULIP.

> Not everyone can effort a workstation that costs $5000+, an internet
> connection, energy, and supplies of paper.

So let them eat paper (if their institutions can afford the

But this seems silly; PCs, public workstations, connectivity and the
rest of the scholar's (and student's) contemporary vademecum are surely
not an issue here. How are you and the rest of us accessing this email
or web-site? Well that's all you need to access the entire serial
corpus too!

> The library which provides 'free' access to all qualified users is far
> more democratic and in the long run it is less expensive using paper if
> it is properly supported.

Are we all living in the same decade? Is it more democratic to have
access to the serial literature depend on (1) whether your institution
can afford a subscription, (2) whether you can afford the time to go to
the library to seek and find, and (3) whether the one "democratic"
paper copy of the journal, if it is subscribed to at all, is not being
used by someone else at the time? Would those 65K daily accesses to the
papers in xxx have been freer and more democratic if they had instead
been attempted in that paleolithic manner?

It is not only hard to take such arguments seriously; it is hard to
imagine that they could be expected to convince any but those who have
not experienced the power and potential of the new medium, and have a
priori reasons for not wanting to do so!

> Not long ago, the publisher of SCIENCE, AAAS, made a major investment
> with the ONLINE JOURNAL OF CLINICAL TRIALS. They had a star-quality
> editorial board and the attention of the scientific world. After taking
> a "bath" they sold it, you probably recall.

Like Red Sage and TULIP, the OJCT was a nonstarter for a very simple
reason (indeed the same reason in all three cases): It was predicated on
S/SL/PPV. Hence a lot of money was put into making a paper-like
"product," and then trying to sell it the old way, through S/SL/PPV.

Well, although the world may still be ready to pay for a tried-and-true,
brand-name journal in hybrid paper/online form via S/SL/PPV (even
though it is a Trojan Horse), it is certainly as little ready to pay
for a NEW online-only journal via S/SL/PPV as it is to pay author page
charges (for paper-journal papers, obviously, as that would just add
insult to injury, but not for online-only journal papers either, yet).

The difference, however, is that a free, author-page-charge-based,
on-line-only literature has a future, indeed it is optimal for scholars
and scholarship alike, whereas S/SL/PPV, astride its Trojan Horse, is
rife with conflicts of interest, and could only disserve scholarship,
keeping it hostage to toll-gates and fire-walls.

So OJCT is ailing/failing because of S/SL/PPV, not because it is
online. Only tide-over subsidy till culture and practise are ready for
author page-charges will keep OJCT afloat.

> Odlyzko, A.M. (1998) The economics of electronic journals.
> [Odlyzko's] question, whether ejournals can operate at lower costs and
> still provide all the services that scholars required, must be answered
> with a resounding no. As long as all the scholars do not have
> electronic access, he will not near a "yes" this these ideas.

It's odd that as an advocate of a medium that is much less accessible
(for both physical and economic reasons) you should be the spokesman
for access for ALL. In any case, the 65K xxx hits daily certainly don't
sound like a resounding "no" to me...

> [Odlyzko] also suffers from his association with Bell Labs which has been
> promoting the electronic alternative since the late 1960s with no
> success. They have an apparent agenda. ATT, after all, is a major
> provider of related technology.

This is nonsense, I'm afraid, and symptomatic of the kind of desperate
conspiratorial scattershot that is needed to defend an indefensible
paradigm. Odlyzko is a world-renowned mathematician in the old Bell
Labs tradition that couldn't give two figs for selling phones or
services. He moonlights in electronic publishing for the same reason
others of us -- e.g., Paul Ginsparg, Elliott Lieb and myself -- do:
precisely because we do NOT have any financial interests or conflicts,
and hence can fight for what we think is optimal for scholars and

> - Users already have been absorbing more costs of equipment, printing
> supplies, photocopycosts, document delivery fees, and travel to use a
> decent library. They complain about this. Why would authors and other
> researchers favor more cost shifting in their general direction?

Nonsense. The 65K xxx users are just using the equipment they already
have for other reasons (word-processing, email, network, research)
for this new purpose that does not add a penny to their expenses, but
is an incomparably more powerful and efficient way of accessing the
serial literature.

> - It would eliminate the dissemination services - primarily
> organization, presentation, and procurement -- provided by publishers
> and libraries. Why would researchers favor this?

When the articles of the whole serial corpus are in xxx, freely
available to all, what publisher/library "dissemination services"
are lacking?

> - It would eliminate the special features -- bibliographies, books
> received, directories, reference works and other information compiled
> by publishers and libraries. Why would researchers favor this?

On books and magazines I have no opinion, and mixing them in merely
clouds what is otherwise a very clear issue: For the refereed journal
corpus, the online papers, suitably tagged by journal, are all that
one needs. Modern electronic search and browse tools can do the


> An electronic version of Astrophysical Journal was
> unable to achieve savings suggested by Harnad without dropping
> editorial standards. Costs eliminated by author keyboarding were
> replaced by costs of online editing. The authors also note that the
> strong demand to continue publication in paper form will double many
> make-ready costs of production and distribution.

I would like to hear the facts from the editors/publishers of the
journal in question; this sounds a bit too partisan to me...

> The preference of most researchers is for review articles and reference
> materials, not preprints or formal primary reports. All forms are
> important, but the preference suggested by every measure is clear.

On what empirical data is this odd and rather incoherent principle
based (for of course to have a review literature, we need a primary
literature to review)? You probably mean that for certain purposes
people prefer consulting reviews rather than sifting through the
primary literature. So what is supposed to follow from that?

And what is this (false) opposition between reviews and unrefereed
preprints? Most of the serial literature (though at one stage all
preprints) is neither, and it is the future of that literature that we
are discussing.

> Library impoverishment is discouraging to publishers. Poor library
> collections are an obstacle to writing good reviews and syntheses.

I have no idea what complicated causal nexus is being imagined here.

First, we are talking about (refereed) serial collections only. Not
being able to afford serials is surely an obstacle to a lot more than
"writing good reviews and syntheses"! And it's access to that serial
corpus that we are talking about, access that is blocked by the
expense and inefficiency of paper, and the S/SL/PPV that supports it,
not by the necessary finiteness of all budgets, including serial
acquisition budgets. Deepening those pockets (assuming more money
could be diverted from less worthy causes: which?) would only
perpetuate the status quo, which consists of a bad medium (compared
to online), high, needless costs, and a bad cost-recovery mechanism
(S/SL/PPV, with its attendant access blockage).

> Productivity in research is controlled by the information used as an
> ingredient. It dictates method, materials, hypotheses, and the other
> resources applied. When researchers and reviewers have not acquired and
> fully digested the findings of others, the result is error and
> duplication.

One wonders what inspired this homily, and what earthly relevance it
could have to the matters at hand!

> Library dissemination has always "free" to the user.

As long as one forgets (1) - (3) in the "democracy" discussion above:
The cost of the institutional subscription, hoofing it to the
library, and hoping it's not under use by someone else. Compare this
to xxx.

> Until recently it
> has not required the user to connect an expensive work station to an
> expensive wire.

Neither did mail; just licking a stamp and hoofing it to a mailbox,
and waiting. And now we're all the better off for email. So what is the

> ah> I would add, nor is library growth able to disseminate what
> ah> it cannot absorb. In my eyes this is due to universities diverting
> ah> library finances to support administrative growth.
> >
> sh> Ah me...
> >
> ah> A solution to the problems of dissemination would be easily derived by
> ah> financial support of library growth that would keep up with research.
> >
> sh> Yes, let's just make paper journals fatter and serials acquisitions
> sh> budgets fatter; that should solve all of our problems...
> Then you are against a market-oriented solution to problems of
> dissemination and the bottleneck in research communications?

What can be more market-oriented than letting researchers vote with
their fingers and eyeballs, as they are doing with xxx? The
trickle-down to serials cancelations, as usage moves West, is just a
matter of time.

> Your are against better financial support for libraries?

Ah me. And the nuclear family?

I'm for spending the resources available for access to the serial
literature in the optimal way. Pouring more money into libraries
through S/SL/PPV is not that optimal way.

Stevan Harnad
Received on Tue Aug 25 1998 - 19:17:43 BST

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:45:27 GMT