standardizing "postprints"

["David Goodman" <David.Goodman@liu.edu>]

The term "postprint" means somewhat different things to different
publishers, and is variously interpreted by authors. For the success of
"green OA" it is necessary that the postprint be in some way less than the
journal article, but contain exactly the same information. This is rather
a fine line!  Some of us have assumed that a version not containing all of
the publisher's corrections is sufficien; others, that only the full
identical published PDF will do.

The compromise that has been sometimes settled on, as in the NIH plan, of
final accepted manuscript but not the final edited manuscript is not
always possible, as in many small journals the editors do much of the
proofreading and copyediting as well.  The distinction between the two
must vary between articles, depending both upon the care taken by the
author and the publisher, and the way they interpret these terms. What
author would not try to correct a spelling mistake, even if the scientific
referees missed it?  What reader of a postprint would not wonder just how
many mistakes remain? What journal would like to have copies of its
articles posted, with their name and copyright attached, but with the
statement that they have been accepted, but that an unknown number of
errors remain, to be discovered only by collation with the official
version?

I suggest that all authors would want each of their readers to be reading
the same text; that all users want to read a version that they know is
authentic; that all librarians would like to be free of maintaining and
explaining the access to multiple versions; and that all publishers would
prefer that anyone reading an article they have published reads the same
thing.

The policy voiced by Tony McSean for Elsevier offers a more workable
distinction:  the author's version should be the same as the publishers in
all details of content, but not the same in presentation. Thus the
publishers will be able to have added value from the higher production
quality and special features of their version, and the readers of the OA
version would read the fully corrected material.

This statement still relies on the author correcting the copies to agree
with the published ones.  The only authors who will find it easy are those
who do not care about the errors, or are certain they never make any.

A solution for this too is in hand, building upon the policies of a number
of publishers: the publisher itself will prepare the OA version and the
published version, and arrange for the necessary posting of the OA
version. (I leave undetermined where it should be posted, but perhaps the
publisher's own site is best.)  Any author ought to be prefer to write for
such a publisher, who will do what the publisher, not the author, is best
able to do.

There are many ways to differentiate the forms, but I suggest that the
informal OA copy might be in html only, and the published one in PDF. They
could both be trivially produced from the same XML, and almost all
publishers already do publish both.  (Another possibility is a html
version, but without the live links. We were very happy to have such
versions only a few years ago.)  There is even direct precedent--many
publishers are willing to furnish to aggregators only a html version, but
not a PDF or page image. This is and will be rasonable, as it seems agreed
that most serious readers prefer the PDF, and most students are content
with the html.

The secondary services could always provide two links: one for journal
subscribers, and one for everybody. Libraries would only need to manage
the two possibilities in SFX or similar services, not the multiple choices
now necessary--and one would always be available off-campus or for
distance education, without proxy servers and similar complications. This
could replace the special arrangements for less developed countries, and
ensure respect for the publisher's copyright.

And it would eliminate the need for any special arrangements for financing
OA.

Dr. David Goodman
Associate Professor
Palmer School of Library and Information Science
Long Island University
dgoodman@liu.edu

I acknowledge the assistance from various sources in clarifying my thoughts. 


-----Original Message-----
From: owner-liblicense-l@lists.yale.edu on behalf of Mcsean, Tony (ELS)
Sent: Thu 8/18/2005 3:44 PM
To: 'liblicense-l@lists.yale.edu'
Subject: Elsevier posting policy (was Information and Computation)
 
I am happy reassure David and everyone on the list that there has been no
change of policy for Elsevier authors and that the wording reflects the
formulation you will find on the Authors section of
<outbind://2/www.elsevier.com> www.elsevier.com which I quote in full
here:

"You can post your version of your article on your personal web page or
the web site of your institution, provided that you include a link to the
journal's home page or the article's DOI and include a complete citation
for the article. This means that you can update your version (e.g. the
Word or Tex form) to reflect changes made during the peer review and
editing process."

Tony McSean

Director of Library Relations