Authentication of versions

["Joseph J. Esposito" <espositoj@worldnet.att.net>]

>How different is this final version and is that the authentic version and,
>if the author has (as urged) deposited the postprint should he or she 
>then replace it?

JE:  I participated in a meeting this week in which this question came up
in different forms, along with the intriguing idea of developing an
authentication service for scientific papers.  (This was not a commercial
suggestion, by the way.)  What isn't clear to me is how important
authentication and versioning is outside the context of establishing
credentials for individual researchers.

To put this another way, if the whole endeavor of scholarly communications
were in some way to be decoupled from the world of tenure decisions,
professional advancement, etc. (not that this is possible or desirable,
but what if?), perhaps by establishing a policy that all publications be
anonymous, how important would it be to know what is the first, second, or
last version of something?  Presumably (and this may be wrong) the "right"
version would be the one whose ideas and information would be absorbed
into subsequent research and publication, and those subsequent
publications would obviate the need to go back to the "authentic"
publication.  If scientific research is about ideas and not the people who
create, discover, or publish them, what is the value of knowing what is
and is not the authorized version--assuming always that the *process* of
ongoing communications brings the truth to light.

Joe Esposito