Re: Elsevier Gives Authors Green Light for Open Access Self-Archiving

From: Stevan Harnad <>
Date: Fri, 28 May 2004 13:32:57 +0100

On Fri, 28 May 2004, Armel Le Bail, Universite du Maine, France, wrote:

> Thanks to Elsevier, but...
> What about some special data included into the papers ? For instance,
> atomic coordinates of crystal structures are traditionnally copied in papers
> by monopolistic commercial databases (CSD, ICSD, CRYSTMET), and not
> put in open access at all (exception for the PDB - Protein Database -
> and AMCSD - minerals -and NDB - nucleic acids). This concerns
> the final coordinates possibly modified after the reviewing process.
> To my knowledge, the database institutions do not pay a cent
> to the journals (even, they pretend to a copyright on the data).

OA is about the full-text contents of the 2.5 million articles published
in the world's 24,000 refereed journals (Elsevier's comprising about
1700 of them).

It is not about what is *not* in those articles. Nor is it about what
database publishers harvest into their proprietary databases.

But data in OA contents can be harvested by more than one party.

(And, by the way, the postprint is the corrected, peer-reviewed final
draft, as also pointed out in the letter from Karen Hunter of

> Are the authors allowed by Elsevier to upload their data (CIF files)
> into an open access database like the COD (
> We would like to see crystallographers deposit CIFs with the COD prior
> to publication, with the understanding that this disclosure should not
> be considered "prior publication" when a paper is prepared for journal
> publication.

Journal publishers do not *own* the data reported in the articles they
publish. They own only the copyright to the full-text (if that has been
tranferred). If you publish the datum "atomic weight of Lugdunum = 18"
in your (Elsevier) article, no one owns (or has a copyright or patent
on) that fact. I may read it (if I have access), use it, state it,
and cite it in my own (Springer) article. I may also have reported in
my own (Springer) article the "atomic wight of Hubdunum = 21." Again,
you may use, state and cite my datum in your (Elsevier) article.

A database compiler can also harvest both those data and others like them
into a proprietary database that I and everyone else must then pay to
access. I may again use those facts in my research and cite them in my
articles, but "fair use" limits how many of them I may reproduce. And I
may not offer a compilation of them online, because the database compilers
have added value and borne expenses in their harvesting and compilation,
and I cannot pirate their efforts and expenses to re-offer it without
permission and compensation.

But this does not apply to the *data* reported in an OA primary article:
Both proprietary databases and nonproprietary databases are free to
harvest the data from OA articles and make compilations of them if they
wish (taking your L=18 and my H=21, etc., presumably citing the source
where appropriate).

Neither the full-text itself, however, nor substantial verbatim portions
of it, are merely "data" to be harvested: In other words, OA does not
necessarily entail the right to harvest and republish full-texts in
their entirety, nor substantial subsets of them, and to republish them
online (e,g, in compilations) without further permissions.

But one must also ask oneself why anyone would want to do anything
like that with OA articles! For the full-texts are all already OA,
already accessible to anyone online, and hence can be just as easily
*linked* as harvested, if someone wishes to do a compilation.

By way of an ironic example: Ulrichs responded to my recent call for
anyone with access to Ulrichs to supply me with the journal lists for
the 85 (now 93) publishers in the SHERPA/Romeo database. Ulrichs pointed
out that this would be a violation of Fair Use. Fair enough. We are
now instead in the process of requesting from or directly harvesting each
publisher's journal list from their own ("OA") website.
The primary publishers themselves are of course not treating their own
journal lists as proprietary databases but as advertising their

Soon authors will come to realize that their own full-texts journal
articles are really rather like advertisements too: written to be read
and used as much as possible.

Stevan Harnad

     Prior AmSci Subject Thread:
     "Free Access vs. Open Access"
Received on Fri May 28 2004 - 13:32:57 BST

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:47:28 GMT