Re: Role of arXiv

[Stevan Harnad <harnad@ecs.soton.ac.uk>]

On 2010-10-13, at 7:02 PM, Okerson, Ann wrote:

> HMMMM, We see huge value in robust subject repositories that 
> present readers with tools and content appropriate to their 
> disciplines.  My not-so-secret hope has been that subject 
> repositories will thrive and endure, and will not depend on the 
> existence of thousands of institutional repositories.

For this secret hope to be realized, there has to be a way to get 
all of OA's target content -- the c. 2.5 million articles per 
year published in the planet's peer reviewed scholarly and 
scientific journals, across all disciplines -- deposited in one 
or the other (i.e., in at least one central or institutional 
repository).

Now, institutions worldwide (universities and research 
institutions) are the universal providers of *all* that target OA 
content, whether funded or funded, across all disciplines. 
Virtually every author of every one of those annual 2.5 million 
articles has an institutional affiliation. Each author shares 
with his institution a joint interest in showcasing and 
maximizing the visibility, access, uptake, usage, applications 
and citations of their joint research output. The researcher's 
institution (and employer) is also in the position to mandate 
that that research output be deposited in the institutional 
repository, in order to provide those benefits to the author and 
the institution, just as the researcher's institution already 
mandates publishing (or perishing) in the first place.

In contrast, a "discipline" has no interests (except perhaps a 
maze internally competitive ones), and no power to mandate 
anything -- neither publishing nor OA. Moreover, how many 
disciplines are there? and how many subjects? and does an 
author/article falling into several need to be deposited in 
several central repositories? How many? And if not, which? And 
why? If central collections are the place to *search* (and they 
are), why on earth should they be the place to *deposit*? Do we 
deposit in google, and other search engines? No, they harvest. 
>From not thousands of sites, but hundreds of millions of them. So 
should central OA repositories. (The planet only has about 10,000 
universities and research institutions...)

Whereas the author's own institution should host and archive the 
deposits; the central harvesters need only harvest the metadata. 
And many different central harvesters can harvest, mix and match 
the same metadata. But the author need deposit only once: In his 
own IR.

Funders can help. But not by foolishly mandating 
institution-external deposit for that small fragment of an 
institution's research output that they happen to fund. 
Institutions are the universal providers of *all* research output 
(funded by different -- sometimes multiple -- funders, and some 
of it unfunded). Funders should encourage and reinforce 
institutional OA mandates by specifying that the (preferential) 
locus of mandatory deposit for their fundees is in their own 
institutional repositories.

One locus of deposit, convergent rather than divergent, 
collaborative rather than competitive, mandated by both 
institutions and funders: in the author's own IR. Then harvest 
and search centrally as, "robustly" as you please.

No point hoping for a preferred kind of repository if you haven't 
a hope (or a viable plan) for getting it filled the target 
content.

A word to the wise.

Stevan Harnad