** Apologies for Cross-Posting **

In case there is even the slightest of doubts about which software I
recommend for creating and maintaining an Open Access Institutional
Repository (OA IR), it is, of course, GNU EPrints:

    http://www.eprints.org/

EPrints is free: the world's first, most widely used, and *by far* the
most functional of all the available OA IR softwares. It is created
for and specifically focussed on OA functionality, with the most
advanced features, being designed and added continuously by the EPrints
developmental team, as it keeps up with (and indeed often leads) the
accelerating development and evolving needs of the worldwide OA movement.

Enough said. See the testimonials (and add your own!) at:

    http://www.eprints.org/news/features/worlds_best_practice.php

To find out exactly what the focus is on and for, see:

    http://www.eprints.org/openaccess/

Now, assuming you already have an EPrints (or other) OA IR, you can go
back to lobbying for OA self-archiving mandates for your institution
and research funder, the current number-one priority!

    Generic Rationale and Model
    for University Open Access Self-Archiving Mandate
    http://openaccess.eprints.org/index.php?/archives/71-guid.html

    A National Open Access Policy
    http://www.ecs.soton.ac.uk/~harnad/Temp/weaker-OApolicy.htm

But don't forget to register your IR in ROAR (Registry of Open Access
Repositories):

    http://archives.eprints.org/

and to register your OA IR policy in ROARMAP:

    http://www.eprints.org/openaccess/policysignup/

Your arch archivangelist,

Stevan Harnad
http://amsci-forum.amsci.org/archives/American-Scientist-Open-Access-Forum.html

PS Ceterum Censeo: *All* OA texts and their metadata should be directly
deposited locally, in the author's own OA IR. That is the primary
content-provider. If it is desired to also include them in one or more
central repositories such as PubMed Central or Arxiv, they can and
should be *harvested* from the local OA IR. EPrints is implementing
automatizable import/export features for doing just that: Exporting
to central repositories, as well as importing from them (e.g., papers
already deposited centrally prior to the creation of the local IR).