Building a Large-Scale Open Source Repository at OhioLINK

Thomas Dowling, OhioLINK

Or: “A cautionary tale in three acts”

At OhioLink they decided that the repository should be “A place for our stuff”, including:

  • art collections high quality/high resolution – multiple tens of thousands of images
  • art and architecture slides from U Cincinnati
  • paintings and drawings from U Cinn collection
  • items from Works Progress from Cleveland State
  • Akron art museum
  • history : historical photos e.g. Wright Brothers at Kittyhawk
  • digs of Mayan architecture
  • historical photographs from lake erie
  • national underground railroad freedom center items
  • OSU geology collection
  • forestry research ctr items
  • dolphin embryos images
  • digital video – documentaries, etc.
  • Kent State oral histories of shootings
  • 100 level undergrad physics experiments
  • OSU lab that has phenomenal archive of bird calls with thorough metadata
  • Text site: 9.1M articles of dissertations center
  • 7500 electronic books


They want a unified repository architecture with open source tools. They chose open source software: Fedora – because it has proven to be bullet proof.

Open source: now cost models seem more feasible and lots of freedom to try things out

· in the plan: ingest in xml

· goes from machine to machine xxsl protocol

· get it all into a fedora repository

· and pump it out via various softwares ext or UM digital library extension service

That was his LITA conferrence proposal, and their project is still very much in progress – real life pressures have slowed down the program –they have tried since 2005 and weren’t making progress. Recently someone created a deadline and they got to work; after two years of development: no progress other than determining that they are going to develop what they can using DSpace. Several factors were revealed, including expiring licenses and a need to have something up and working.

Much of the talk was very interesting details regarding their decision making processes and the growth of their understanding regarding what they needed to do in order to get a large-scale repository actually out of development and into production.

One comment

  1. Thomas Dowling

    To clarify our locally hosted text operations, we have 9.1M locally loaded scholarly journal articles, and about 13k electronic theses and dissertations.

    Slides for this presentation (such as they are), in PowerPoint 2007 format, are available at <http://sepia.ohiolink.edu/~tdowling/presentations/lita2007/>

Comments are closed.