SPARC review of OAI-archive-creating code

From: Stevan Harnad <harnad_at_COGPRINTS.SOTON.AC.UK>
Date: Fri, 5 Oct 2001 16:15:36 +0100

SPARC E-News 08-09/2001 |

>From the Scholarly Publishing and Academic Resources Coalition responses to:

 Review: Software

 (Editor's Note: With this issue of SPARC e-news, we inaugurate a
 periodic feature that focuses on evaluating technology solutions which
 enable a more open scientific publishing marketplace. SPARC believes
 in providing active support for scholar-led publishing initiatives
 that offer transformative potential, and we offer this review of the
 Eprints software in the hope that it may encourage interested
 parties.) Software: A Review

 Ed Sponsler and Eric F. Van de Velde
 California Institute of Technology

 Stevan Harnad just made it easy to join him in his quest to free the
 scholarly literature. Download Eprints software from, and build your own repositories quickly and
 easily. Developed by Harnad collaborators Robert Tansley and Chris
 Gutteridge, Eprints provides a web interface for managing, submitting,
 discovering, and downloading documents. Eprints repositories are
 compliant with the Open Archives Initiative (OAI)
 ( Therefore, once a repository is
 registered as an OAI data provider, OAI-aware information services
 will be able to discover its content.

 This review is based on our eight-month experience with Eprints.
 During this time, we built several technical-report repositories and
 one online conference proceedings. Our repositories are available at

 Setup, Configuration, and Administration

 The cost of getting started is minimal. An experimental system,
 suitable for initial testing and even for hosting a few small
 production repositories, costs less than $1000. Obviously, one must
 move up to higher price and performance points when the number and
 size of the repositories, the number of users, and the performance
 requirements increase. Eprints requires the Linux operating system
 with a standard configuration of supporting software:

   Apache, the web server
   MySQL, the relational database
   Perl, the scripting language
   Various plug-in modules for Perl

 The operating system and all of the supporting software are Open
 Source software licensed under the GNU General Public License (GPL).
 (See Eprints developers intend
 to make Eprints officially Open Source as soon as they have
 implemented certain critical features. In the meantime, the University
 of Southampton holds the copyright, but it allows users to view,
 modify, and redistribute the source code. This is very close to being
 Open Source in the GPL sense. (See
 for full details.)

 After installing the basic software, one must configure the
 Eprints system for local use:

   Customize the look and feel of the local Eprints web site by
   adapting scripts that control the presentation. These scripts are
   well separated from the core Eprints code that deals with archiving,
   database management, and internal workflow. Therefore, we expect
   that future upgrades will leave the customized scripts largely
   unaffected. Decide what metadata fields to use for describing a

   Decide what metadata fields to present to the user during a

   Set up subject hierarchies that provide meaningful browsing
   options to users.

   Register the repository with OAI. Since OAI is a built-in
   feature, the registration is easy.

 The repository is now ready to accept documents. Authors place
 documents in a temporary storage buffer. Before moving documents from
 the buffer to the public area, Caltech librarians perform the
 following quality-control checks:

   Enforce repository policies with respect to author affiliation,
   subject area, departmental approval process, or any other criteria
   appropriate for each repository. Verify and (if necessary) improve
   the metadata. Good metadata enhance discoverability.

   Check document formats:

     Ensure online readability of all submitted documents.
     Convert documents to formats that conform to best

   Take one of the following actions:

     Return the document with comment to the author.

     Reject and delete the document.

     Accept the document.

   Create a unique document identifier. We create our own persistent
   identifiers independently of the Eprints system. This is a
   safeguard in case we switch from Eprints to another system in the
   future. As much as we like Eprints now, better systems may come
   along. Moreover, no one can guarantee the long-term survival of any
   software. (A detailed description of our identifiers and associated
   resolver will be self-archived in the caltechLIB repository at

   Create browse pages. We generate a browseable view of the repository
   by executing Eprints-supplied scripts on a regular nightly schedule.
   These scripts generate static web pages containing subject-grouped
   lists of links to documents in the repository.

 User Features

   Eprints supports any type of document format, including HTML, Adobe
   PDF, and PostScript. However, repository administrators should
   carefully consider which formats they are willing to support and

   Authors must create a repository account in order to be able to
   submit documents. The repository administrator controls what
   information is requested and what information can be used to create
   metadata for submitted documents. Readers are encouraged to create
   an account. Registered users may set up an e-mail alerting service
   for new content in their subject areas.

   The current version of Eprints assigns each user a password. This
   is somewhat of an annoyance. Because users cannot choose their
   passwords, they are likely to forget them.

 Overall Evaluation

 EPrints is a powerful and inexpensive solution for sharing scholarly
 works with the world, a concept Harnad calls "self-archiving." The
 web-based submission process is intuitive and requires minimal effort
 on the part of authors. However, long-term preservation requires an
 institutional commitment. The Caltech Library System is committed to
 preserve indefinitely those documents that are self-archived in its
 repositories. To this end, the library performs quality checks on
 submitted documents and metadata, enforces repository policies, and
 assigns persistent identifiers. Eprints gives us web-based tools to
 perform these management tasks efficiently.

 Eprints validates many of Harnad's claims. It is possible for
 researchers to make their research freely available to everyone,
 increasing the impact of their research in the process. Because of
 Ginsparg's arXiv, physicists already profit from this revolution in
 scholarly communication. With Eprints and the Open Archives
 Initiative, the fundamental building blocks are in place to spread
 this revolution to all other disciplines. The only roadblock is the
 willingness of researchers to experiment with this new
 scholarly-communication model.

 To subscribe to e-news, please email a request to

 2001, SPARC - The Scholarly Publishing and Academic Resources
 Coalition Unless otherwise noted, copyright is held by SPARC.
 Permission is granted to reproduce and distribute or post.

 Posted: October 2, 2001
Received on Fri Oct 05 2001 - 16:36:18 BST

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:46:15 GMT