[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]

RE: preprints



[So shoot me, I accepted another. -- MR]

Alan Brush wrote:

>Late last week Stan Friesen brought up the possibility of posting
>preprints of scientific papers to either home pages or discussion
>sites (e-mail).

I am orginally a physicist and now that I am writing papers in
paleontology I have been pretty shocked to discover that there is no
preprint system.

It is not clear to me why paleontology does not use preprints more.  My
best guess is that it is due to two factors - first, there is simply
less work going on.  If the volume of papers is small enough then there
is less demand to know things early.  Also, I think that the prevalence
of field work is also a factor.  A major motivation for preprints in
physics is that you might start working on a topic only to later
discover that somebody else was already doing similar work, but you
didn't find out until the the paper appears in a journal - a year or
more from the time it was submitted.   This scenario is less likely in
paleontology because a person excavating a site or describing a specimen
usually has unique access to it and thus does not have to worry so much
about another working having done so earlier.

Note that in NO way am I suggesting that physics is superior to
paleontology in this regard - I am simply pointing out that the issues
with preprints have been around for many years and the paleontological
community (of which I barely qualify) might want to learn from that
experience.

In physics, preprints are an essential part of the process, because the
publication delays in refereed journals are simply too long - a year or
so (and thus similar to most paleo journals).  This latency is
considered unacceptable in physics.   The system normally works like
this - you submit a paper to a journal, and then immediately send copies
of the preprint to a list of say 100 researchers and academic
institiutions that have active physics departments.    Many departments
will issue the preprints as a "technical report" with an official stock
number so that people can order them if necessary.  In other cases
individuals simply hand them out.

Everybody knows that preprints are unrefereed, so due caution is
exercised on beliving all of the results.  However, it is very valuable
for people to get some idea of what is going on, who is working on what
etc. Graduate students love this system because they get some warning if
their thesis topic is being worked on elsewhere.   

>    It is important that the author be very clear as to the intent of the
>posting. If it is considered a publication, then having the material in a
>print journal may be duplicate publication.

These issues have been dealt with for decades in the physics community.
Preprints are not considered publication, but they may be referenced in
the same sense that you may reference "personal communications".
Generally speaking, a person reading a scientific paper accords more
weight to a reference to a refereed journal paper or book than to a
"personal communication", and in physics at least, a reference to a
preprint falls somewhere in between.

Journals are still considered the definitive record of progress in the
field, and  there is no true sense in which the advantages of peer
review have been lost.  However, the latency in going through the peer
review process makes the journals too slow to be the sole means of
communicating work. Note that this preprint system orginated totally in
paper form - the 100 or so copies were just photocopied and mailed.  It
has existed this way for the last 30 years or more without threatening
journals or the peer review system.  

>    It seems to me that haveing a body of data or literature on 'the net'
> leads to problems of sttribution. Should the material, with no editorial
>review, be made available in an archived form. How would it be cited by
>subsequent authors. What if the material is subsequently changed
>substantially following review or even rejected when submitted to a journal.
>Would it be removed from the electronic database? Or, what if the paper
>is rejected by several journals, should it be retracted from the electronic
>archived. How would readers in the future find out about this?

All of these issues have been dealt with in the physics community.
Note that all of this comes up in paper form.  The net actually makes
many of the issues less severe.

In the last few years, physicists have started using online repositories
on the Internet to suppliment paper preprints, and there are many people
looking toward the day when journals disappear altogether.    Although
that extreme outcome remains controversial, few  if any would dispute
the value of the present preprint system.

>    Howerver, there are some down sides to this process. The posting
> of manuscripts or data sets mean that unedited material enters the
> process. The text has not been reviewed or edited, other than the
> author no one has questioned the app roperiateness of the
> statistical analysis, nor has the experimental design been
> reviewed. From the authors position, this is risky.
>
>   I would also recommend caution as to what precisely is the
> subject.  For example, in order to describe a new species, there are
> conventionsset by international bodies. Priorty set in cyberspace is
> not (at the present) the same as priority set by the current
> standards.

Issues of priority are actually another big benefit of preprints in
Physics.  First, they are typically not sent until the paper is
submitted so priority in terms of submission date is assured.  Second,
they make the submission date more widely known and generally inform the
community of who is doing what.  When push comes to shove, a widely
circulated preprint is very valuable in establishing priority for an
idea.

Nathan