Center for Digital Education & Converge: research in education technology for K-12 and higher education

Villanova Liberates Library Search

on August 16, 2010
Through the open-source software VuFind, libraries around the world improve user search experiences. Front: Demian Katz, the lead VuFind developer at Villanova University. Back: Library director Joe Lucia. | Photo courtesy of Demian Katz

As search engines around the Web change with the times, library search engines have stayed the same.

If you go to a university library page, you can't search for keywords by just hitting the "enter" key. You can't narrow your search results easily. And you can't see similar items. 

But with open-source software developed by Villanova University, you can. The Pennsylvania university has helped bring library search up to date with the rest of the computer science world, without the hefty price tags of commercial search engines.

After going through beta for two years, the university released its 1.0 version of VuFind in July. And libraries around the globe have adopted VuFind, including Georgia Tech, London School of Economics, National Library of Ireland, University of Southern Queensland and the University of Michigan. Others have entered public beta, including Yale and Auburn universities.


Free library search

When you search with VuFind, you'll see an interface that looks like Google's search engine. You can search by keyword, see suggested searches, narrow the results with a variety of filters and look at related items. And you'll find the same technology that Netflix uses, as Villanova's lead VuFind developer Demian Katz likes to say.

“I think the biggest thing here is that it’s using more contemporary technologies found elsewhere on the Web.”

If you're looking for a book, you can see whether it's available in real-time and what it looks like. If you need to cite it in APA or MLA style, the engine will give you the citation with a warning that it may not be 100 percent accurate.

And with the click of a mouse, you can favorite a search result, save it, send it to your phone, e-mail it, tag it or comment on it.

These technologies allow users to step outside of the traditional containers that hold library catalogs hostage, Lucia said. “We wanted to basically liberate library search." 

Typical catalogs center around pre-defined search headings and types, such as author, title and subject. VuFind still includes those features — they're just inside a more flexible search experience.

In Yale University's case, its Orbis online public access catalog often returned zero results, said Katie Bauer, director of usability and assessment for the library. Many people selected "search title" as the search field, but were actually trying to do a basic keyword search. That didn't work.

Unless you knew exactly what you wanted, you couldn't find it.

But with VuFind, you can. 

Yale modified VuFind, branded it as YuFind and started using it in its beta form in 2008. This year, YuFind became widely available at the university and shown more prominently on the library homepage alongside Orbis.


Narrow search results

In early testing, Yale students liked the simple search, the big search box that looks like Google's, and the way the faceting breaks down the search results, Bauer said. The faceting tool on the right side of the search results allows you to narrow them down by format, call number, language, author, genre, era and region.

The faceted search brings a lot to the table, said Greg Pendlebury, who helped bring the software to the University of South Queensland in 2009.

Through studies of users, the Australian university found that they didn't pick up on faceted searching straightaway because it was too new. But once they started using it, those users gravitated to the feature.


Collaborate worldwide

Pendlebury has worked with other open-source projects, and usually he sees a painful transition period when it moves into the broader community. But Villanova has minimized the pain. A core group of people at Villanova and around the world, including Pendlebury, do the gritty work that must be done to help the software succeed.

"You can dig into the tool that’s doing the job and tweak it a little bit,”  said Pendlebury, who is now a senior systems developer and integrator with the Australian Digital Futures Institute at the university.

By making the software open source, libraries around the world can bring their search engines up to date. And unlike commercial solutions, they don't have to wait for a vendor to give them a new feature, Lucia said.

“If there’s something we want to do or something we don’t like, we can adapt the code and make it change and have the change in place sometimes in a day or less.”


Plan for the future

When Yale started looking for a next generation online public access catalog a few years ago, VuFind was one of the more advanced services — and one of the few that was open source, said Jeffrey Barnett, senior research analyst at Yale University Library.

Back then, faceted search seemed out of reach to libraries, said Daniel Lovins, metadata and emerging technologies librarian at Yale University Library. In 2006, a commercial search engine from Endeca made a big splash in the library world when North Carolina State University started using it.

But the search engine was expensive.

That's when Villanova started developing a prototype engine using Solr, an open-source enterprise search platform from the Apache Lucene project. Yale began the alpha and beta process early on.

Now the Yale library staff members face a dilemma. They've provided good functionality with VuFind, but they also have two interfaces that users see, Bauer said.

If you need an exact search, Orbis still functions better. And the library decided to keep the ordering of books in Orbit, which means users have to switch back and forth if they use VuFind.

“I don’t like having this kind of added layer of complexity that we’ve given our users,” Bauer said.

At some point, Yale will solve the dilemma. But for now, the library staff has a vision for what a next generation catalog should look like.

“What we envision the ending system to be is one that meets the requirements, continues to grow, is easy to maintain and maximizes the impact that we have at a reasonable cost,” Barnett said.

Improving access to non-western language materials and scripts is important to Lovins. (VuFind currently has translations available for nine languages, including Chinese and Japanese, and the ability to create more.)

He wants to see a mature, next-generation catalog "that’s equally powerful and does equally good artificial intelligence on any kind of language that you might use.”

Villanova is already planning ahead for VuFind's future. The university will host a VuFind summit in September with 40 people who develop at some level at their institutions. They'll talk about what version 2.0 should look like and collectively figure out a plan to get there, Lucia said. 

And this fall, they're working on a programming interface that incorporates results from Summon, a giant index of full-text articles, into the local VuFind, Katz said. In the same search, users will be able to find library books and articles at once instead of doing a separate search.

“There’s always the challenge that first time library users don’t necessarily know what they’re searching when they go to the library catalog, and you’ll get a lot of people trying to find articles in the catalog when it’s really just books,” Katz said. "But now with this new setup, they really can do the search in one place and get pointed in the right direction from there.”

You may use or reference this story with attribution and a link to

If you enjoyed this story, subscribe for updates.

View Sample


Add a Comment

Add a Comment