On Tue, 15 Jul 2003, J.W.T.Smith wrote:

> On Tue, 15 Jul 2003, Philip Hunter wrote:
> >
>ph> I think perhaps the JISC community and the digital library community
>ph> might consider preservation of the primary corpus (as you put it) as a
>ph> quite separate issue. I hadn't quite taken on board that you accord
>ph> self-archived items a secondary and purely short-term functionality.
>
>js> I don't think Stevan would say self-archiving is "purely short-term" but
>js> that the main purpose is to make research results available to those who
>js> can make use of them as quickly as possible.

Yes, but primarily in the form of peer-reviewed, published articles
("postprints"), and only secondarily (and optionally) also in the form
of their pre-peer-review precursors ("preprints)." A misunderstanding of
this distinction (and these priorities) is at the heart of many a
confusion that is holding back self-archiving and open access. Read on:

>ph> It could be that it scarcely exists because the real issues for the
>ph> research community are being ignored, which is that eprints need to have
>ph> as much of the value and properties associated with paper documents as
>ph> possible, if researchers are to feel the effort of 'deposit' is
>ph> worthwhile.
>
>js> The effort of "deposit" is almost zero compared to effort of getting an
>js> article into a refereed journal. It is paid for many times over by the
>js> average increase in citations to papers available online.

Correct. But, as usual, the equivocation here is on *what* is being
deposited. The approximateness of natural language is such that we can
all *seem* to be talking about the same thing when are in reality not:
What I *always* mean by the deposited eprint is the peer-reviewed,
published article (postprint), preceded, optionally, by its
pre-peer-review preprint. But what JWT Smith means is something else: it
is an "article" in a (hypothetical) "deconstructed journal," a speculative
proposal discussed extensively in the American Scientist Forum across
the years: http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/0216.html

>js> It is clear you [Philip Hunter] are confusing eprints with articles
>js> in refereed journals.

No confusion. Eprints *are*, in the first instance, published articles
in refereed journals, self-archived in order to maximize access and
impact: http://www.eprints.org/self-faq/#What-is-Eprint

>js>  Currently [eprints and articles] do not play the same role and
>js>  neither are they intended to play the same role. Eprints are about
>js>  communicating information to other researchers and refereed articles
>js>  are about gaining formal recognition for the work done. Refereed
>js>  articles are also likely to be 'archived' in the formal sense.

Is it becoming clear how the unending confusion keeps on appearing? All
one has to do is use "eprint" to mean only "unrefereed preprint" (and
relegate the "archiving" of refereed articles to the publisher), and we
are no longer talking about maximizing access to the existing,
published journal literature by self-archiving it, but about some
hypothetical alternative way of (self?)-publishing.

To repeat: Self-archiving is intended to provide open access to the
2,000,000 annual articles in the planet's 20,000 refereed journals in
order to maximize their impact. The publisher (and library) does indeed
archive the articles too, but that archive is toll-access, and that
is what self-archiving is intended to supplement, by providing an
open-access version.

And research communication is not about disseminating raw drafts (though
that can perform a valuable supplementary function too). It is about
publishing material that the would-be user can be confident has passed
through an answerable process of quality-control and certification by
qualified experts, in a journal with a reliable track-record attesting
to the quality that can be expected of its articles. That (simple)
process is called peer review, and until the annual 2,000,000 preprints
have gone through that process, the rule is "caveat emptor!" Moreover,
no one has a clue of a clue of what "preprints" would be like if they
were not all knowingly destined (as they are) to be answerable to peer
review. None of the hypothetical alternative systems much-debated in
these pages has been tested and shown to yield a literature of comparable
quality and usability.

Yet this speculation about new forms of communication and publishing,
and new kinds of journals, and new -- or no -- forms of quality-control
have meanwhile been distracting us from something that is not at all
hypothetical, but tried and true, namely, that the impact of the
existing research literature -- *such as it is* -- can be vastly
increased if it is made openly accessible to all would-be users by
self-archiving it.

>ph> It seems that your solution is to persuade the researchers
>ph> that 'deposit' equals 'access', and that 'access' is enough.
>
>js> In terms of communication, access is enough.

Again, the fundamental equivocation: Access to *what*? The 2,000,000
annual peer-reviewed journal articles we have now (plus, optionally,
their pre-peer-review precursors)? or access to some hypothetical
alternative entity?

>ph> Where does long term citation come into this?
>
>js> Long term citation is irrelevant to current communication.

If this sort of thing is being said by friends of open access, does
open access really need enemies? Is this the way to encourage sceptical
scientists and scholars to self-archive?

To set the record straight: In its various embryological stages,
every paper has a pre-peer-review preprint phase (lasting about 4-12+
months while the peer-review and revision goes on) and a post-peer-review
phase (starting with the refereed, accepted, published final draft, but
potentially followed by a series of updates). During the preprint phase,
caveat emptor. But if you do read and use preprints, cite the preprint
until the postprint appears. Then read and cite that thenceforward
as the official "archival" version (but track and cite corrections and
updates too, if/when appropriate).

You may safely cite the published, official version forever -- *exactly*
as you could before the online era! Nothing has changed but accessibility.

>js>  No doubt Stevan will one day put [Temp papers] somewhere more
>js>  permanent if he feels they are worthwhile and there they will be
>js>  indexed by the search engines and so anyone sufficiently interested
>js>  will be able to find them. Does it matter where they are as long
>js>  as they can be found?

To maximize the likelihood they will be found, storing them on a website
is not enough. All postprints (and those of your preprints you also wish
to make openly accessible) should be self-archived in your own
institution's OAI-compliant Eprint Archives. All of mine (including
those in my Temp website) are, within a few days or weeks of being
written.

>js>  Again you have missed the point - the purpose of self-archiving
>js>  is communication not preservation. You are thinking with the paper
>js>  publication model where the refereed article played all the roles
>js>  required by researchers, it communicated, it quality controlled,
>js>  it was archived, etc. Today these roles are falling apart,
>js>  the eprint communicates, the refereed version quality controls,
>js>  the online search services (including the RDN gateways/portals)
>js>  enable finding, etc.

The refereed version does not quality-control: The journal's referees
and editors quality-control. And the eprint (postprint) *is* the
quality-controlled version! (Smith is right that the functions are
becoming modularized, but he has the modules partitioned wrong!)

> My own suggested model, the Deconstructed Journal
>
>     http://www.kent.ac.uk/library/papers/jwts/d-journal.htm
>
> shows one way in which all the roles of the refereed journal could be
> distributed over many cooperating actors.

For critique, see these threads:

http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/0216.html
http://www.ecs.soton.ac.uk/~harnad/Hypermail/Amsci/0943.html

> Incidentally - this article shows the power of making papers available
> online - it has been accessed several thousand times - at least 10 if not
> a 100 times as many as have probably read the paper version.

Open access certainly enhances reader-quantity (for both preprints
and postprints), but it is peer-review that enhances (and ensures)
article-quality.

Stevan Harnad

NOTE: A complete archive of the ongoing discussion of providing open
access to the peer-reviewed research literature online is available at
the American Scientist September Forum (98 & 99 & 00 & 01 & 02 & 03):

    http://amsci-forum.amsci.org/archives/september98-forum.html
                            or
    http://www.cogsci.soton.ac.uk/~harnad/Hypermail/Amsci/index.html

Discussion can be posted to: [log in to unmask]