Research repositories, whether they are a physical library, an electronic
journal archive, an institutional repository or a subject repository, are
collections of interconnected components. Understood in this way, as systems,
they have emergent properties. That is, properties of the collection that none
of the components (eg.: individual research articles) have, as well as
properties of the components that the components have as a result of being a
part of that collection (eg.: relevance ranking with respect to a topic within
that collection). What are some examples of emergent properties of
repositories: the subject coverage, the intended purpose of the collection, the
demographics of the readers and authors of the collection, etc.
When a researcher makes the decision to publish/provide access to their work,
the emergent properties of the repository are a relevant consideration.
Consider the following hypothetical situation: a researcher in Buddhist studies
may, for example, object to being "mandated" to the act of placing his article
on the topic of "interdependent co-arising" in the same repository that is also
home to articles from another department in his institution that specializes
in, say, promoting the philosophy of Charles Darwin in social science. That
researcher may wish to place his article in the Tibetan and Himalayan Digital
Library, but not in the IR of his university. I agree with Thomas Krichel that
researchers currently have the freedom to choose and promote the channels of
distribution for their work.
About Arthur Sale's statements such as:
Arthur Sale:
"Researchers are not free agents.
[..]
I strongly support academics being required to contribute to their discipline
and access to knowledge (and opinion). Otherwise why are they employed?"
In my opinion these statements can only succeed in creating resistance from
researchers. I don't think that "the compulsory open society" is what Karl
Popper had in mind when he wrote The Open Society and Its Enemies; "Open
access in your employer's IR, or else!" The fact that the Open Society
Institute claims to be inspired by Karl Popper's Open Society and its Enemies
does not mean that Popper ever intended to have his theories be implemented
through OSI's NGOs, or at all. The Open Access Initiative claims to define and
promote "open access", but the concepts of open society and open access reach
back to antiquity and touch on paradoxes of freedom and political theory. As
an aside, OAI-PMH is a "a low-barrier mechanism", but a barrier nevertheless -
perhaps not a paradox, but there is something counterintuitive about promoting
open access with a new barrier.
The concepts of open society and open access existed long before Budapest OAI
and OAI-PMH. And here we can go back to Jean-Claude Guédon's point about the
fact that this debate goes a long way back, and that there is an important
difference between theory and practice:
Jean-Claude Guédon:
"This is an old debate where one should carefully distinguish between two
levels of analysis.
1. In principle, is it better to have institutional, distributed, depositories,
or to have central, thematic, whatever depositories?
2. In practice, we know we will not escape the will by various institutions to
develop central, thematic, whatever depositories (e.g. Hal in France). And
these depositories will exist. The question then becomes: how do we best live
with this mixed bag of situations?
Pursuing the battle on principles is OK with me, but it does not get me
enthused.
Pursuing the battle on the pragmatic, practical level, knowing that various
tools exist that will restore the distributed nature of these depositories
anyway, appears to me far preferable."
I agree with Jean-Claude Guédon regarding the important difference between
theory/principle and practice. However, I don't agree with the last sentence,
where he expresses confidence that "various tools exist that will restore the
distributed nature of these repositories", I am not convinced of this. A while
back there was a posting on this list about A Physicist's Challenge to
Duplicate Arxiv's Functionality Over Distributed Institutional Repositories:
"If you want to convince me [that institutional self-archiving plus central
harvesting can provide all the functionality of Arxiv ], then try to do so by
conducting the following experiment with any... "harvesting" vehicles you like:
(1) Choose an area, such as Mathematical Physics, or Integrable Systems,
and find all the papers that have been deposited in any of the archives that
they cover, within the past week. (If they cover 95% of the arXiv, they must
necessarily producethis information just as well). No other barrage of junk;
just that simple list of papers.
(2) Do the same with respect to all the posted publications by a given
author for the past ten years. Again: not a barrage of google-like junk dumped
upon you, but this specific information. (If I want a ton of junk, I can also
go to Google scholar, and waste endless time trying to find what I need.)
(3) Find out, at one go, if a given article, or set of articles, from the
above list, has been published in a journal , and what the journal reference
is.
(4) Get a copy of any of these articles, at once, in any convenient format,
like .pdf, that is available.
(5) Be equally sure that all the above is simultaneously done for all such
articles deposited in individual institutional repositories.
If you can do all the above, successfully, you will have given the 'proof of
principle'."
This challenge is about recreating some of the emergent properties of arXiv
with distributed IRs. I think that even this problem is currently unsolved and
it will be very difficult to solve at best. It calls for authority control on
the researchers' names in a distributed environment that includes thousands of
repositories from all subjects. And this challenge calls for the re-recreation
of only some, and not all, of the emergent properties of arXiv.
I think that ignoring the relevance of emergent properties of collections is a
mistake. I remain skeptical of attempts at formalizing this abstract notion of
"collection" into the data model of an IR software (such as is the case with
DSpace), as well as the vision of future harvesters that recreate the emergent
properties of subject-thematic repositories with probabilistic algorithms. I
do not object to trying to create these new algorithms and technologies, in
fact the topic is of great interest to me, but I don't think it is helpful
trivialize that which is far from trivial.
I am not an opponent of IRs, in fact, I am preparing one for Concordia
University, but I see IRs as a service that the university offers to its
faculty.
Tomasz Neugebauer
Digital Projects & Systems Development Librarian
[email protected]
Concordia University Libraries
1400 de Maisonneuve West (LB 341-3)
Tel.: (514) 848-2424 ex. 7738
-----Original Message-----
From: American Scientist Open Access Forum
[mailto:[email protected]] On Behalf
Of Thomas Krichel
Sent: Thursday, February 05, 2009 9:46 PM
To: [email protected]
Subject: Re: Repositories: Institutional or Central ? [in French, from Rector's
blog, U. Liège]
Stevan Harnad writes
> (Academic freedom refers to the freedom to research (just about) whatever
> one wishes, and to report (just about) whatever one finds and concludes
> therefrom.
in the channel of one's choice. IRs should make themselves
publication channels of choice.
Cheers,
Thomas Krichel http://openlib.org/home/krichel
RePEc:per:1965-06-05:thomas_krichel
skype: thomaskrichel
