RE: Errors in author's versions

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

Dear Cliff,

Yes, it is sometimes necessary to rewrite an author's manuscript, 
and, yes, authors make amusing typographical mistakes. Librarians 
for their part tend to be amused by publisher claims of 
perfection, for we have been dealing with the wrong citations in 
copy-edited, proof-read published papers since the beginning of 
our profession. But I agree, the authors left to themselves would 
do even worse.

But we are talking about differences in scientific content 
between posted and published versions. Among the few substantial 
differences I have seen between versions, is where a publisher 
dropped a line of a table. Should we therefore insist the the 
author manuscript always be permanently available, so the reader 
can check whether the publisher made any mistakes?

Cliff, if you know any instance where there have been significant 
scientific errors in the posted version, but not the published, 
please cite the examples, because not one has been shown, neither 
by you, by Anthony, by Lisa, or by Peter. Peter even emphasized 
that he did not know of any.

Find one, any of you. It will be good to have something concrete 
to discus. I posted in the first place to see if I was mistaken, 
for I had not expected to find them so similar. They were 
proposed to collect counter-examples. I expected some and am 
disappointed in one sense that they have not yet been found, for 
I had planned to analyze them.

To say that versions could differ profoundly, when no cases have 
been found, is misleading, because any type of any single version 
could have an error. To claim that authors' versions are 
inadequate in content, you need to show that they do have an 
increased occurrence of errors beyond those occurring in all 
published material. It could be proven that a type of material 
does contain errors by finding one, and one could then proceed to 
determine the frequency. What I have proposed are propositions 
that can be falsified.  For example, if my propsition had been 
that published peer-reviewed articles never contain fraud, it 
could easily have been falsified.

=======================

All of this said, I agree on the true practical point:

There should be no need for authors' versions, at least after the 
date of publication. The version available to all should be the 
version as published--and then, if necessary, publicly corrected.

However, some on this list will now disagree, for, as I see it, 
the need for authors versions is a make-shift. It is only 
necessary because of publisher unwillingness to allow the best 
version to be posted. It may be a more reliable make-shift than 
any of us thought, but it will always be an unnecessary 
complication.  May proper OAJournals soon replace it.

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

-----Original Message-----
From: owner-liblicense-l@lists.yale.edu on behalf of Cliff Morgan
Sent: Tue 7/18/2006 7:07 PM
To: liblicense-l@lists.yale.edu
Subject: RE: Errors in author's versions

David Goodman wrote: "perhaps all the controversy is obsolete 
about what version to deposit, and all the discussion about 
exactly what name to use for what version."

The published version may well differ considerably from the 
author's accepted manuscript (known in SHERPA/RoMEO terminology 
somewhat counterintuitively as the "post-print"). Sometimes the 
variations are fairly trivial (e.g. conformance to journal house 
style) but they can certainly go as far as substantive editing, 
especially if the author is a non-native speaker. Both Lisa 
Dittrich and Peter Banks have made this point - that you can make 
no general pronouncements about whether subsequent versions 
differ significantly or not: sometimes they do, sometimes they 
don't, depending on the state of the author's manuscript. There's 
no point in saying that the published version *never* differs 
significantly nor that it *always* differs significantly: the 
point is that it *probably* differs in some way, and *may* differ 
profoundly....

Identifying versions does matter because versions do vary, and 
using unambiguous terms to identify the different versions is 
useful because it is not always clear what is meant when people 
use terms such as "final edited version", for example. ...

The PMC distinctions remain pertinent.

Cliff Morgan
Chair NISO/ALPSP WG on JAV
____________________________