Re: Serials Review Interview

From: Mark Doyle <doyle_at_APS.ORG>
Date: Thu, 15 Oct 1998 12:23:53 -0400

Hi,

On Thu, 15 Oct 1998, Thomas Krichel wrote:

> Mark Doyle notes

> It is still operates, and it is still the one archive that will
> give your papers the largest exposure. But you can repeat that
> to authors as many times as you want, they still don't want to upload
> there.

I don't know what the status of the software and interface are for the econ
archive. xxx has come a very long way since the last time the econ archive
was updated - for all I know they are still running software from 1994 or
before (or maybe they rewrote it all). I am sure it is missing some of the
key advantages that come from having the archive on xxx itself. Anyway,
perhaps an economist should do a study of how individual economists behaving
irrationally prevents optimal distribution of their own work...

> > I had thought that one reason that things weren't catching
> > on in economics was that publishers had more restrictive policies about
> > preprint circulation, but maybe I am misinformed.
>
> One thing that has happened is that an individual author wants
> to open a personal archive only for her reprints, because a clause in
> most copyright transfer agreements allows for reprint in "collection
> that only contain the authors work.

Hmmm, seems having a systematic way of integrating these personal
collections under one umbrella, even if it is distributed, violates the
spirit of the clause, no? Why should authors stop there and not push for
even more permissive copyright agreements that permit centralized archving
as well?

> > Are all servers high-availibility servers?
>
> I am not sure what you mean by that, but presumably the answer is no,
> i.e. not all are.

I mean connected to high-bandwidth networks (xxx is on 3 T-3's) and
mirrored around the globe in over a dozen countries. And that the server
itself is up well over 99% of the time.

> > Are they run by a grad student who will move on at some point leaving
> > it to languish?
>
> Some are. Others are run by central banks or economics think tanks.
> These are larger archives that we expect to be quite stable.

I could see making a case for the latter, but I still think they too would
be better served if they also put their work in a centralized place. There
is no advantage to having papers "archived" on a departmental server that
may or may not be maintained.

> > Do they keep abreast with the latest technical developments and migrate
> > to new formats as needed?
>
> Yes.

Including legacy conversion of old papers and not just accepting new
formats that come along? All servers?

> > I don't think you should write off the advantages of having a scalable
> > centralized (but mirrored) repository.
>
> I am not doing that at all. All I am writing is that a centralised
> approach may not be suitable for all commuities.

Sorry, it didn't come across that way (mostly because of the misinformation
about NCSTRL and its relationship with xxx). Most of the reasons that it
may not be suitable seem to be to be political and that absent those, most
would agree that a centralized, mirrored system has many advantages over a
distributed system. In any case, it is clear that xxx can play a role in a
distributed system as exemplified by NCSTRL.

> Certainly when we think
> of extending free electronic documents from preprint
> disciplines to non-preprint disciplines, from TeX based collections to
> "wordprocessor babylon" collections,

preprint vs. non-preprint is really a non-issue I think. xxx is just as
effective a model for circulating reprints as preprints. It is just a
question of what rights an author retains when signing a copyright transfer
(or not signing as the case may be).

TeX vs. Wordprocessor babylon also doesn't matter as much, though it is
clear that some things that xxx does are only available to the TeX world.
But if all you are doing is circulating PDF's or Word files, the
centralized, mirrored model still has many advantages over the distributed
approach because the advantages apply to basically any set of documents
served on the web.

> from uncontested to contested knowledge

Not sure what you mean here... Surely you don't think there aren't vehement
disagreements in physics? In any case, the question of when a paper should
become available on a centralized archive to all is, to a large extent, a
community dependent one.

> it may not be optimal to consider xxx as the only possible model.

Granted that the way the physics community uses xxx is not the only
possible model, but it is clear to me that any community would benefit by
promoting unencumbered, free circulation of authors work through
centralized, globally mirrored archives. The only other approach that would
have many of the same benefits (but far less efficiency and new problems)
would be to have a distributed network of servers that all mirror each other
and are on equal footing. Otherwise, if any member of the distributed
network is ephemeral, unmaintained, or otherwise lacking, the community
loses.

Regards,
Mark
APS Research and Development
Received on Tue Aug 25 1998 - 19:17:43 BST