---------------------------- Text of forwarded message -----------------------
Date:         Mon, 13 Apr 1992 15:20:21 EDT
From:         Stevan Harnad <harnad@PRINCETON.EDU>
Subject:      The problem of multiple postings

> Date:         Mon, 13 Apr 1992 11:33:00 EDT
> Sender: "Publishing E-Journals : Publishing, Archiving,
>               and Access" <VPIEJ-L@VTVM1.BITNET>
> From: Bill Drew -- Serials Librarian <DREWWE%SNYMORVA.bitnet@VTVM1.CC.VT.EDU>
>
> Is there any reason why the discussion of referreeing of ejournal must
> be carried out on four different lists?  Why does it need to be on
> PACS-L as well as the lists specializing in ejournals?  How many
> people are truly unique to just one of the four?  I subscribe to all
> four because the material interests me not because I want to see the
> same item four times.  Please consider this comment.

The problem of what to do about multiple lists with overlapping but
non-identical subscriberships and subject matter is simply not solved,
and I would be interested in a solution too. Here are the factors
involved:

If one has information on a topic that is of interest to several lists,
one can either post only to one of the lists, so as not to risk sending
multiple postings to the same individual -- but then that is at the
expense of NOT reaching the non-overlapping portions of the lists -- or
one can post to them all, and then risk drawing complaints from those
who received the message more than once.

At the moment, there is no ideal solution. Not all topics are matched
exactly to one and only one list. Deleting multiple versions of a
message only costs a recipient the same number of keystrokes as messages
(and rarely does a topic ovelap more than a half dozen lists). One can
already create customized mail filters that automatically detect and
discard multiple versions of the same message in one's incoming mail --
just as the listservers themselves have such filters to block multiple
versions of the same message posted to the same list. When these mail
filters are in general use, this problem will vanish (except for the
extra traffic created by sending the multiple messages -- although even
this could in principle be handled by intelligent centralized
routers).

My own provisional policy is to favor reaching the nonoverlapping
constituency at the expense of the few extra keystrokes for the
overlapping segment, yet I don't wish to be antisocial. People who very
much MIND having to perform the extra keystrokes tend to be more vocal
than those who don't, so one tends to hear only from them. What weight
should be given to their (legitimate) complaint? It seems to me that
this depends on numbers: How many object, relative to those who do not?
What is the relative size of the overlapping and non-overlapping
portions of the lists? How relevant and important is the particular
topic to each of the lists? How frequently does this happen?

Obviously polls cannot be taken on each and every occasion, as these
would cause more keystrokes than they cured. But the problem is
certainly worthy of discussion.

Stevan Harnad