On PDFs, Request-Buttons, Sense and Evidence

From: Stevan Harnad <amsciforum_at_GMAIL.COM>
Date: Sun, 17 Jan 2010 20:59:58 -0500

On Sun, Jan 17, 2010 at 5:25 PM, Klaus Graf <klausgraf_at_googlemail.com> wrote:

> KG: I cannot see any empirical evidence that the assertion below is more
> than wishful thinking.

Keep reading:

>> SM: Stevan asserts that researchers who cannot afford access to the published
>> version of articles are perfectly happy with the self-archived author's final version.
> KG: There is in contrary evidence that researchers strongly prefer
> publisher's PDF and that they prefer to deposit publisher's PDFs in
> IRs although they aren't free in eternity.

Klaus, you continue to miss the point. I will again try to spell it out for you:

(1) The assertion is not (and never has been) that users would not
prefer access to the publisher's version of record over access to the
author's final draft, if they can access the publisher's version of

(2) The assertion is that users prefer access to the author's draft
over no access at all (when they have no subscription access).

(3) And the fact is that they have no access at all to 85% of articles
(when they have no subscription access).

(Do you have contrary evidence for either of these assertions, or this fact?)

(4) The assertion is not (and never has been) that authors would not
prefer to make the publisher's version of record freely accessible to
all potential users online, if their publishers endorsed it.

(5) The fact is that most publishers do not endorse making the
author's version of record freely accessible online, but about 63% do
endorse making the author's final draft freely accessible online.

(6) So that is the version that funders and universities should (and
do) mandate making freely accessible online.

(Do you have contrary evidence for either this assertion, or these facts?)

> KG: The best way to get a citable version is to mail the author if he has
> a publisher's PDF.

The accessed version is the accessed version. The citable (and cited)
version is always the published version of record.

And PDF has nothing to do with it.

(By the way, nothing prevents authors from making their author drafts
immediately OA and also depositing the version-of-record in Closed
Access, if you have it, and letting the Button generate the eprint
requests if users want the publisher's PDF. [That would be a lot
faster, easier and surer than users having to look up the author's
email, having to send the author an email, and the author then having
to email the eprint.] But I bet the reason few authors -- out of the
already-few authors who are as yet self-archiving at all today --
bother with doing this extra step is because they feel it just isn't
worth it. (I have never bothered to deposit the publisher's PDF once
in the over 20 years I've been self-archiving. I just make sure that
my final draft is at least as good as the version of record: indeed,
sometimes it's better, if I catch some corrigenda or do updates, which
I mark clearly as corrigenda or updates.)

> KG: (i) It is a myth (and Harnadian orthodoxy) that mail buttons in IRs
> are working well - no empirical evidence.

Self-archived articles are still rare. Mandates are rarer still. And
the "email eprint request" Button is still the rarest of all.

All this will grow (though it's all growing far too slowly).

Klaus has not had good luck with his few Button requests:

My suggestion is to try a lot more, at different institutions. (Klaus
seems to have a particular fixation on University of Zurich, for some

When you have a decent sample, across institutions and disciplines,
please do come back and report your evidence.

For the time being. common sense suggests that if an author deposits
at all, chances are it's because he wants his article read; and if he
feels he cannot deposit it in Open Access, he will deposit it in
Closed Access, and fulfill individual Button requests if he receives

My guess is that new authors sometimes don't yet understand the
Button; or the Button requests are not yet implemented correctly by
the software, so they don't reach the authors. (Klaus Graf, if he is
truly interested in whether or not the Button "works," may care to
follow up on such cases, find out whether the authors got the
requests, and if so, why they did not fulfill them, if they did not.
He can then report that evidence to us too.)

> (ii) It is a myth that there is a well working central OAI harvesting
> scholars know. OAIster is down and this was the most unfortunate OA
> event for me in 2009 - but nobody cares (except Mr Krichel, maybe).

Don't grieve. There are plenty more harvesters: scirus, scopus,
citeseerx, citebase, base, pubmed, even ROAR (try it). And don't
forget google and google scholar: they are not OAI harvesters, yet
they are the source of most of the hits on IRs as author websites
today, as the webstats show.

I don't care about the sell-out of oaister because no harvester is of
much use as yet, with only 15% of the target content being made OA.
And once the mandates bite and OA content grows significantly, there
will be plenty of innovative new harvesting and search services with
far richer functionality than what we have now. Today's sparse content
is simply not worth investing the creativity and time into designing
more powerful functionality.

Stevan Harnad
Received on Mon Jan 18 2010 - 02:01:45 GMT

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:50:04 GMT