Re: OAI and the rational publisher

From: Mark Doyle <doyle_at_APS.ORG>
Date: Thu, 4 Apr 2002 15:19:54 -0500

Greetings David,

On Tuesday, April 2, 2002, at 05:27 PM, David Goodman wrote:

> 1. Of all the possible organizations, your's is the strongest and the
> most
> focused and therefore the best able to take the step. I think the
> result
> will be positive for your journals, as the best material from others may
> soon be looking for another home.

The problem is that we are already "drinking from the fire hose" as
Marty Blume
puts it. It will be hard for us to expand rapidly (on the other hand, an
economic
model which scales with the number of submissions can make it possible
for us to expand - the current model which is hinged on a dwindling
subscription
base can't do this).

> 2. The readership of APS journals is not limited to the membership of
> APS.
> They are widely read in other fields--you do not serve only the physics
> community. (This is true is all academic fields.)

For sure. Physics is relatively lucky that our core readers and
submitters largely
overlap. This isn't true say for a medical journal or for many chemistry
journals.
This means it is easier for us to move away from a subscription model.

> 3. What exactly do you think is necessary for proper archiving beyond
> what Stevan's proposal calls for?

A proper system of archives would be based on widely-supported
standards and would be in a format that is unconstrained by the file
format
used by authors or end users. The canonical example would be an XML
based archive
in which the formatting of the article is left out and which is based on
standards
like Unicode 3.2, MathML, etc. and in which any reusable information is
marked up to enable confident reuse (i.e., marked up bibliographic
citations
make it easy to reuse this information for creating links between papers
and
tracking citations). There also needs to be an infrastructure for
reliably producing
the file formats (today PDF and HTML) for end users. The archive should
be
refreshable (delete all PDFs, and regenerate them perhaps with a
different formatting). arXiv.org can do a lot of this, but it is messy
and can be
unreliable.

A proper set of archives would be administered by a diverse group of
organizations
and would robust against failures. One vision put forth by Ginsparg is
to have institutions provide archives of their research and then these
would be fed
to centralized subject based archives. This would provide some
redundancy - if
a centralized subject-based archive were to fail, it could be rebuilt
from the institutional ones, and vice versa. But there is a fair amount
of work in building such systems
and figuring out the appropriate infrastructure. There is of course some
work
being done on these issues (Mellon grants (particularly Dale Flecker's
project
on SGML/XML), the OAIS model for archives, etc.). One challenge is to
make
this technology mature enough so that many institutions can deploy and
use
it. A second challenge is to find cost effective ways to populate the
archives.

Anyway, a straight author-prepared PDF repository fails on almost all of
these
grounds. They may work great for basic dissemination, but I don't think
this
is the optimal way to store scholarly output.

> 4. Can you envision any workable form of peer review that does _not_
> involve journals?

Good question. Depends on what "workable" means and what "journal" means.
The loosest definition of a journal might be a peer review label on a
collection
of articles for the purposes of citation. The journal provides a brand
name
and a persistent name (conventionally journal, volume, page) so that a
citation
conveys some information about the article's quality and how to locate
it.
If you look at most alternatives to "traditional blind peer review
moderated by an
editor", you usually find that they are more intensive in labor than the
current
system. So this can be more expensive on a per article basis for
articles that
are actually peer reviewed (this includes counting "free" referee time
as a cost).
The way out is an alternative where fewer papers are formally reviewed
so that
perhaps the overall cost will go down. However, authors would probably
be persistent until they find someone to peer review their articles.
There are
other systems ("eopinions", movie reviews, etc.) which don't result in a
persistent
label and I guess these are peer review without a journal. But they are
more
like opinions than classical peer review and while these may be
guideposts
to the literature, I am not sure they would provide a suitable way to
raise the
quality of the literature. They do have some advantages in that the
signal
isn't a binary yes or no - a "Phys. Rev." one star rating instead of a
four star rating
  would convey a lot of information actually. But the decision process
would be
much more complicated.

Cheers,
Mark

Mark Doyle
Manager, Product Development
The American Physical Society
doyle_at_aps.org
Received on Thu Apr 04 2002 - 21:39:49 BST