Eprint versions and removals

From: Stevan Harnad <harnad_at_ecs.soton.ac.uk>
Date: Fri, 6 Jun 2003 20:13:12 +0100

Posted with permission.

    [This is a reply to query about why Eprint Archives puts any
    constraints at all on author-removal of archived papers]

Date: Wed, 28 May 2003 18:01:00 +0100 (BST)
From: Stevan Harnad <harnad_at_ecs.soton.ac.uk>
To: Franck Ramus <f.ramus_at_ucl.ac.uk>

"Let authors add and remove at will!" That was what I too first thought,
when proposing self-archiving. But then the library and historical
community educated me that an archive is an archive, and a historical
record is a historical record. If someone makes a document public,
and publicly accessible (I will not say "publishes it" because that
should be reserved for the official, journal-certified, peer-reviewed
published draft), then others can read it, use it, cite it. If it then
vanishes without a trace, or gets arbitrarily and irretrievably changed,
all you do is confuse scholarship.

It is for that reason that we have made the following compromise:

Yes, an author can have a text removed from an Eprint Archive if they
elect to do so, but (1) this is discouraged rather than encouraged,
for the reasons mentioned, and (2) an empty file for the original draft
remains as a place-holder if it is removed, so that if someone uses or
cites *that draft in particular*, and then they or someone else want
to go back and look at it again, and it is gone, then there is a marker
for it, and it is distinguished from other drafts that might have come
before or since.

Eprint archives are merely supplements to formal publication, not
substitutes for it. They are intended to provide immediate open access
to both unrefereed drafts and refereed published drafts (and also
post-publication updates and revisions) for those who cannot afford
access to the toll-access version. But because the open-access drafts
are public, there must be a way for scholarship to continue without
turning drafts into moving targets, phantoms or apocrypha. Much better
to keep and track versions, making sure that the latest version is
high-lighted, than to make prior versions vanish after they have been
publicly accessible.

I hope this now makes the constraints we have put on removing drafts in
Eprints and CogPrints a little more understandable.

Stevan Harnad

> On Mon, 26 May 2003, Franck Ramus wrote:
> As both a contributor and a user of Cogprints, I am not satisfied with
> the current feature of the archive where all the earlier versions of
> a paper remain fully visible. As an author, I want people to read the
> latest version of my papers, and only that one; while in Cogprints
> there is a real risk that they download an older version. I know that
> all the versions are interconnected and point at the most recent one,
> but I find that this link is not visible enough and I suspect that many
> users fail to retrieve the most recent version when several are available.
> One easy solution would simply be to exclude old versions from searches
> and lists: this would make search results and lists clearer, since not
> filled with irrelevant duplications; and this would ensure that users are
> led straight to the most recent version. Old versions would be reached
> only via the page of the most recent version, just for the record (in
> case this has any interest, which in most cases I doubt). And it should
> be made less difficult for authors to altogether remove a record if they
> want; it's theirs after all, they should be allowed to remove their own
> material without justification if they want to. In my case some papers in
> the archive have been revised significantly and subsequently published;
> I don't want the old versions to remain accessible, this could only
> cause confusion.
> On Mon, May 26, 2003 at 03:03:29 +0100, Stevan Harnad wrote:
> Thanks for the suggestion. I will ask Chris Gutteridge
> to make the accessing of the top version more prominent and failsafe.
> I agree that the top version should be the one users access, but I
> disagree that older versions should be suppressed, and are of no
> interest. Scholars are interested in successive versions, and an archive
> is meant to be permanent. (There was a lot of debate about versioning,
> and most felt the benefits of retaining it all out-weighed the risks.)
> But you are right that the author should have the right to remove
> a text if he wishes, and he has the right. All he has to do is request
> that it should be removed (and say why) and it is removed.
> On Tue, 27 May 2003, Christopher Gutteridge wrote:
> OK. How does this look?
> http://cogprints.ecs.soton.ac.uk/archive/00000870/
> On Wed, 28 May 2003, Franck Ramus wrote:
> Yes, this is definitely more visible.
> I still think that older versions should be invisible in search results
> and topic listings. Those duplications serve no purpose but visually
> confusing users. And I don't see why authors should give a reason to
> have their paper removed. Who are you to judge how good the reason is,
> what criteria would you use and why, and whatever the reason given,
> how could you refuse to do what the author wants?
Received on Fri Jun 06 2003 - 20:13:12 BST

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:46:58 GMT