In 1978, Loren Kohnfelder invented digital certificates while working on his MIT undergraduate thesis. Today, digital certificates are widely used to distribute the public keys that are the basis of the Internet’s encryption system. This is important stuff! But when I tried to find an online copy of Kohnfelder’s 1978 manuscript, I came up blank. According to the MIT Libraries’ catalog, there were just two copies in the system: a microfiche somewhere in Barker Engineering Library, and a “noncirculating” copy in the Institute Archives.

Google couldn’t find anything. Nor could CiteSeer, an online database of scholarly papers in computer science. Finally I found an e-mail address for Kohnfelder himself in MIT’s online alumni database. A few hours later, he informed me that a scanned copy of his thesis could be downloaded from the website theses.mit.edu. And as it turns out, a copy of Kohnfelder’s thesis has also been entered into DSpace, the big digital-repository project that MIT Libraries and Hewlett-Packard started back in 2002. That copy is indexed by Google Scholar, Google’s academic search engine. But I hadn’t thought to check there.

DSpace is a long-term, searchable digital archive. It creates unchanging URLs for stored materials and automatically backs up one institution’s archives to another’s. Today, DSpace is being used by 79 institutions, with more on the way. But as my little story about Kohnfelder’s thesis demonstrates, archiving data is only half the problem. In order to be useful, archives must also enable researchers to find what they are looking for. Sending e-mail to the author worked for me, but it’s not a good solution for the masses.

Long-term funding is another problem that DSpace needs to solve. “The libraries are seeking ways of stabilizing support for DSpace to make it easier to sustain as it gets bigger over time,” says MacKenzie Smith, the Libraries’ associate director for technology. Today, development on the DSpace system is funded by short-term grants. That’s great for doing research, but it’s not a good model for a facility that’s destined to be the long-term memory of the Institute’s research output. Says Smith: “We need to know how to support an operation like this in very lean times.”

(1) Submitter uses a Web-based interface to deposit files. DSpace handles any format from simple text documents to datasets and digital video.

(2) Data files are organized together into related sets. “Metadata,” technical information about the data, is kept to support preservation.

(3) An item is an “archival atom” consisting of grouped, related content and metadata, which is indexed for browsing and searching.

(4) Items are organized into “communities” corresponding to parts of the organization such as departments, labs, and schools.

(5) DSpace’s modular architecture allows for expansion across disciplinary as well as institutional boundaries.

(6) In functional preservation, files are kept accessible as technology formats, media, and paradigms evolve over time.

(7) The end-user interface supports searching and browsing the archives. Items can be opened in either a Web browser or a suitable application program.