[OAI-eprints] Re: The RePEc (Economics) Model

[Thomas Krichel]

  Stevan Harnad writes

> The Repec model is one in which many distributed institutions,
> each having archives of multiple economics papers of
> their own, have their metadata gathered together and
> enriched to provide OAI-like interoperability: http://repec.org/

  The interoperability is more complicated then in a conventional
  OAI setting, because the structure of the data exchanged goes will
  beyond what can be done with oai_dc.

> Instead of using the OAI protocol, Repec uses the "Guildford"
> protocol -- ftp://netec.mcc.ac.uk/pub/NetEc/RePEc/all/root/docu/guilp.html --
> but it has been announced that Repec plans to become OAI-compliant
> eventually.

  I already operate a gateway at http://oai.repec.openlib.org. It's
  oai_dc data may be a bit thin, but there is plenty of AMF metadata.

> (Repec does *not*, as I had wrongly assumed, cover individual
> websites too, as ResearchIndex/citeseer
> http://citeseer.nj.nec.com/cs does, only multi-paper institutional
> archives.)

  Departmental archives, as distinguished from institutional archives.
  Some archives serve special purposes, they hold no docuemnt
  data at all. 

> Repec is accordingly a form of institutional self-archiving,
> pre-dating the OAI, but (1) focused on one discipline only
> (economics), and (2) not requiring the individual archives to be
> OAI-compliant (but Guildford-compliant).

  Correct, which is basically just a way to dump files on a disk,
  nothing more. 

> It is a very activist project, "a collaborative
> effort of over 100 volunteers in 30 countries to enhance the dissemination
> of research in economics."

  Correct, and almost all are economics faculty. Some folks do 
  little, but the construction of the whole enterprise means that
  even if they do little, since there are many 

> It should be noted at once that if every discipline had its own
> institutional Guildford-compliant archives and volunteers, as Economics
> has, then I and many others would today be promoting Institutional
> Guilford-compliant repositories rather than Institutional OAI-compliant
> repositories (and the free software that Southampton designed for creating
> OAI-compliant institutional repositories for self-archiving
> http://www.dlib.org/dlib/october00/10inbrief.html would have
> been Guildford-compliant software).

  The technical protocol for the transport matters little. This
  really (!) is a technical matter. We continue with what we
  got because we can not rearrange 250+ archives that otherwise
  do just fine. 

> What distinguished Repec is hence not its interoperability protocol
> (since it plans to become OAI-compliant anyway) but (a) its activism
> and (b) its discipline-specificity.

  and (c) its metadata model. This is by far the most important, but least
  well understood distinction. 

> If there were a way to spread Repec's activism from economics to the
> other disciplines, it would certainly be very welcome, just as it
> would be very welcome if there were a way to spread ArXiv's
> central-archiving tendency to the other disciplines.

  Could not agree more.

> Unfortunately, no such generalization of either Repec or Arxiv to the
> other disciplines has taken place (Repec began in 1997, Arxiv in 1991).

  RePEc has its origin in a project called WoPEc that I started on
  February 1, 1993. In 1997, RePEc was born essentially out of WoPEc
  and some other partners, but WoPEc had the lion's share (I am
  simplifying here a bit.)

> http://www.earlham.edu/~peters/fos/timeline.htm It is for this
> reason that it is OAI-compliant institutional self-archiving that I
> happen to be promoting. And this is at last showing signs of
> generalizing http://www.ecs.soton.ac.uk/~harnad/Temp/tim-arch.htm
> though still not fast enough. It is for that reason that various
> forms of activism need to be promoted too, especially institutional
> activism:

  There is no contradicition between institutional and departmental 
  archives, and aggregator strutures. It is by no means an either
  or choice. And let me emphasise again: having discipline-based
  aggregators will be the best way to stimulate institutional 
  and departmental archiving. The problem is, of course, that
  there are not many aggregators around. Therefore I have been
  argueing for a while thet the institutional self-archiving
  community should stick together to elect one area of discplinary
  priority. That is rather that to fight a war on all fronts,
  concentrate the effort and build systems that are interoperable
  beyond the unqualified DC data model. The DC data model is too simple
  for academic self-documentation.

> At first, FTP sites and Web sites seemed the simplest, fastest and
> most direct way for researchers to self-archive, on a distributed,
> institutional basis;

  They still are, just look at the amount of stuff that is on the
  web. There are so many grass-roots initiatives. The larger
  public is not aware of them because they serve specific communities. 
  This is where I get so angry with Clifford and his---implicit---call
  to shut them down, to fit all publishing activities into a central
  straightjacket. 

> but then the slow progress in this, and the success of the
> physicists' centralized disciplinary model suggested that
> centralized, discipline-based self-archiving might be faster, with
> the Physics Arxiv itself perhaps subsuming it all
> http://cogprints.soton.ac.uk/documents/disk0/00/00/16/99/ (Thomas
> Krichel argued against central archiving,

  Nope. I simply argued that the centralized model would not
  carry through to many disciplines. Where it worked it 
  was certainly an extremely good model. But you insisted
  that because the Physcists had done it everyone could
  and would, it was the optimal way (your flavour of the day).
  But I am still right. arXiv has a very unequal distribution
  of papers even in sub-areas of Physics, I am told. Ebs will
  know better. arXiv is still growing and that is a good thing.

> But central archiving did not catch on (Cogprints has only reached
> 1500 papers in 2003) or generalize to other disciplines,

  Exactly as I had forecasted! And that, depite the fact that
  it was a project subsidized by public funds. When WoPEc became
  a funded project, by the same funders, it had around 5,000
  papers accumulated as a labor of love, only. Much of that
  work was done by José Manuel Barrueco Cruz. 

> and Arxiv itself kept growing at only an unchanged linear rate from
> year to year: http://arxiv.org/show_monthly_submissions
  
  Sure, but it is still is the finest self-archiving project on the planet.
  But it really is self-archiving. Self-archiving is only a part 
  of what I call self-documentation. 
  
> And then came the OAI protocol in 1999, making distributed
> self-archiving equivalent to central (because of interoperability)
> http://www.openarchives.org/documents/index.html
 
  They are not quite, but that is a matter for another email...

> which immediately prompted me to ask Rob Tansley to redesign the
> Cogprints software to make it OAI-compliant and then turn it into
> free generic OAI archive-creating software for institutions
> http://www.dlib.org/dlib/october00/10inbrief.html

  And I think your team are doing a very good job with this.

> I think I now understand this. See above. Both Repec's
> aggregation of institutional multi-paper archives in economics and
> Citeseer/ResearchIndex's harvesting of arbitrary individual websites
> in computer science

  Citeseer are a truely fab project. The material that is there
  should become part of new, RePEc-like data structure called
  rclis and pronounced "reckless". Watch out for it over the
  next few years. 

  With greetings from Minsk, Belarus,


  Thomas Krichel                     http://openlib.org/home/krichel
                                 RePEc:per:1965-06-05:thomas_krichel