IUBio Biosequences .. Software .. Molbio soft .. Network News .. FTP

How to Word Institutional Self-Archiving Policy

Stevan Harnad harnad at ecs.soton.ac.uk
Mon May 16 17:26:16 EST 2005

Here is an anonymized exchange with the librarian of a US university
about the wording of their institutional self-archiving policy. The gist
is that it must be made clear that:  

    (1) only online access is concerned, not paper or DVD distribution,

    (2) there is no need whatsoever for the author to retain or
    renegotiate copyright with the 92% of journals that have already
    given their green light to self-archiving,

    (3) for the 8% of journals that have not, authors should deposit all
    their metadata and full-texts in their Institutional Repositories
    (IRs) anyway, and can choose for now whether to set the full-texts
    (OA) Open Access or (IA) Institution-Internal Access: The metadata
    are OA in any case and the authors can fulfill eprint requests for
    them by email as long as they wish.

    (4) authors are strongly encouraged to self-archive pre-refereeing
    versions (preprints) too, and encouraged to leave them in, for
    the scholarly record, depositing corrected and updated versions in
    addition (the software will always display and give them priority),
    but withdrawal should only be discouraged, not forbidden

> I must admit that the ramifications of the qualification "in any
> format" escaped me until reading this. In fact, print was probably
> not the primary thing our lawyer thought about. 

OA self-archiving of published articles is a brand-new phenomenon.
Lawyers (bless their hearts) have to be *told*, not *asked*, what it's
all about! Otherwise they respond mechanically, based on what they are
familiar with. The result is then Procrustean, and the (needless)
loser is OA and its new possibilities and purposes, not grasped by
the lawyers.

> More like: what if we want to make a DVD of all [university-name 
> deleted] works available in the future? 

To remind you: The OA/IR cupboards are *bare* today, and the goal
is to fill them today. Not to worry about what to do with their (full)
contents after after-tomorrow!

Moreover, self-archiving of already-published articles (which is the *only
thing* I am talking about) is *access-provision* to already published
work; it is not publishing. What (green) publishers have agreed to,
and what authors want, and research needs, is that self-archiving should
maximize the *online* access to their already-published work. It is the
publishers who have all the rights to sell and distribute the on-paper
*and on-disk* versions, and that is just fine, and has nothing to do
with OA or institutional self-archiving of the institution's published
work. Hence there is no need to rock that boat, or tilt at the windmill
of copyright retention/re-negotiation as yet another needless prerequisite
and obstacle to self-archiving.

OA is about providing free *online* access webwide. It is a monumental
step forward, and a lot comes implicitly with the territory (as virtually
every user today has access to the web): No need for IRs to distribute
paper copies or DVDs. Just distribute the URL! The publisher has all the
paper and DVD distribution rights, and that's fine. The author and the
author's institution does not *need* them in order to provide OA to their
own research output. So let's not demand those rights, needlessly,
instead of doing what is really needed, which is providing OA!

> Also, print copies do matter if you want to hand them out in
> class. Your argument that it requires renegotiating the copyright
> with the green publishers is well taken, and we'll have to mull
> it over.

Instead of mulling it over, why not just tell your bright-eyed and
bushy-tailed instructors that this is the 21st century, the online age,
and we no longer hand out print copies in class: We hand out URLs!

> As to our non-withdrawal requirement, this was a conscious and
> deliberate requirement we put in. 

It is (almost) harmless for the published version, which is OA's main
target, but it is disastrous for the unrefereed preprint, which we should
also be encouraging authors to self-archive (for many reasons); yet this
desirable practice will (I promise) be utterly inhibited, and will remain
still-born, if authors are told they may never withdraw their unrefereed
preprints from their IRs, no matter what!

(Authors can be *encouraged* to leave them in the IR, for the
historical record, and reassured that there will be version control,
as in GNU Eprints, with the most recent, corrected version being always
prominently displayed. But authors should not be told that uncorrected
and embarrassing errors will be etched in stone forever if they are
foolish enough to self-archive their unrefereed preprints in their IR!)

> When the library takes on the responsibility to archive (in the
> long-term sense) material, it does so only for material that meets
> minimal quality criteria.

I am afraid that here it is partly the library's incomplete grasp of
what OA (as opposed to perhaps other potential uses of IRs) is about:
OA self-archiving is not (primarily, or even secondarily) for the sake
of *preserving* the published article. The article is published, and its
preservation is the concern of the publisher and the libraries that buy
the official version, as always. The author's self-archived version is
a *supplement* to the official, published version, provided in order
to maximize access, usage and impact, as well as the visibility and
accessibility of the institution's published article output.

It is a *huge* mistake (and a category error, to boot) to conflate
this objective of *supplementary access provision to already-published
articles* with the completely different (library) problem of: "How do
we preserve our [bought-in] digital collections?"

The publisher's digital edition of journals is *bought in* by the
institution. They want to make sure it is preserved and accessible
forever. That is fine. But that has (almost) nothing to do with the
supplementary, self-archived version of the tiny proportion of any given
journal's articles that happen to have been published by authors at that
institution and that have now been made OA in that institution's IR!

I am not saying that IR's shouldn't try to ensure the longevity of their
OA contents. They can and do. I am just stressing that the longevity of
the individual OA supplements is not at all the same matter as the
preservation of the official publisher's version of the published journal
contents themselves, the version of record!

I said the two are conflated and have "almost" nothing to do with one
another because there is one possible eventual connection, but that is
not only entirely hypothetical and counterfactual now, but it is even
counterproductive to speak about it yet, as it puts the cart before the
horse and risks halting the entire parade:

*Conceivably*, if and when 100% OA self-archiving prevails, journals
*might* eventually opt for converting to 100% OA publishing --
author/institution-pays -- at which time, if and when it ever comes,
one of the ways to save the costs of *all* access-provision and
archiving, would be to offload and distribute that function among
the contributing institutions themselves instead of burdening the
publisher with continuing to have to do it. If/when that ever happens,
journal publishers will only be providing (and paid for) the service
of peer-review, copy-editing (perhaps), and certification, whereas the
paying institutions themselves will be doing all the access-provision
and archiving, permanently. *Then* today's two independent preservation
tasks -- (1) the big one, for the publishers' and subscribing libraries'
(and deposit libraries') official version of record and (2) the small
one, for IRs providing today's self-archived OA access-supplements for
their own published output -- will functionally coalesce into one.

But mixing them up today is premature (and counterproductive) conflation,
not functional coalescence.)

> However, we definitely do not want to be in a position to judge
> faculty's work. So, our minimal criterion is that the faculty member
> stands behind the work to the extent that he/she is willing to
> commit this to the archive permanently.

For published articles, there is clearly no need for the university
to "judge": The referees of the journal have judged and certified them
already, and the university is merely maximizing access, in self-archiving
them in its IR.

Unrefereed preprints are another matter; the author needs to judge whether
or not they want to make them publicly accessible *and* must retain the
right to remove them from the public eye if they find that they show egg
on their face! Yet preprint self-archiving is a useful, desirable practice
in most cases. So we should *encourage* it, and encourage non-withdrawal too,
but not *forbid* it a-priori, thereby simply locking out a lot of preprints
that would have come, and stayed, if no sense of "no turning back" were 
gratuitously attached to them. 

In other words, just as we should not conflate 

    (1) the preservation of the official published non-OA article
    itself with the preservation of the self-archived OA supplementary
    version of it in the author's IR, 

    nor conflate

    (2) providing free *online* access to the self-archived OA
    supplementary version of non-OA articles in the author's IR with
    providing free on-paper or on-disk access to the published non-OA

    we should also not conflate

    (3) the self-archiving of supplementary OA versions of (irrevocably)
    *published* articles with the self-archiving of unrefereed,
    unpublished preprints (by making them irrevocable too!).

> Allowing withdrawals would introduce significant practical
> problems. Currently, our archives run relatively autonomously. With
> withdrawals, we would have much more human intervention. Also,
> withdrawals require an extra level of security and authentication
> (making sure others do not delete things from our archive). Withdrawals
> would also introduce dead links into the system, which would frustrate
> users.

I think this might merit a bit more reflection: Do you really want to discourage
preprint self-archiving by making it irrevocable, just because it may entail
a bit more work?

As far as I know, the GNU Eprints software that your university uses has
been specifically designed to handle withdrawals: The eprint can be
removed, and leaves a null record in its place, stating that it has
been withdrawn. The withdrawal, like the deposit, can be implemented so
that either (i) the author confirms the withdrawal (or deposit) or (ii)
a moderator appointed for the department or archive sector concerned
confirms the withdrawal (or deposit).

(The degree of moderation of deposits is up to the institution. While
the cupboards are bare, I favor authenticating authors and passwords,
and then letting authors moderate their own deposits rather than adding a
needless extra layer of control.)

> Aside from this, the withdrawal of a web-published document is
> somewhat of a ridiculous notion anyway. The moment a paper is posted
> it can be harvested by any number of services, so it never really
> disappears anyway. So, the no-withdrawal is only a confirmation of
> a reality.

I agree that once something is openly accessible online, the Pandora's box
is in principle agape. But that is not a reason for discouraging preprint
self-archiving by stipulating that the self-archived version may not be
withdrawn from the IR even if the author wishes. Nothing is gained by
that, and a lot of preprints are thereby merely pre-emptively lost,
never getting deposited because you made the deposit irreversible!

Stevan Harnad

A complete Hypermail archive of the ongoing discussion of providing
open access to the peer-reviewed research literature online (1998-2005)
is available at:
        To join or leave the Forum or change your subscription address:
        Post discussion to:
        american-scientist-open-access-forum at amsci.org

UNIVERSITIES: If you have adopted or plan to adopt an institutional
policy of providing Open Access to your own research article output,
please describe your policy at:

    BOAI-1 ("green"): Publish your article in a suitable toll-access journal
    BOAI-2 ("gold"): Publish your article in a open-access journal if/when 
            a suitable one exists.
    in BOTH cases self-archive a supplementary version of your article
            in your institutional repository.

More information about the Jrnlnote mailing list

Send comments to us at biosci-help [At] net.bio.net