Deep Web Tech Blog

Why A Federated Search?

Editor’s Note: This is a guest article by Anita Wilcox, E-resources Librarian at Boole Library, University College Cork in Ireland. The UCC library is an Explorit Everywhere! customer, and Anita, ever our advocate, graciously accepted the invitation to write a blog post for Deep Web Technologies’ “Customer Corner” amidst her busy schedule.  Anita’s Enterprise Search Europe presentation may be viewed here.


When we opted for a federated search system in University College Cork Library, we were mindful of UCC’s strategic mission of becoming one of the best research universities in Ireland and globally. We realised that only a Federated Search System (FSS) can complement the student experience in the University.

However, with the advancement of Discovery tools, now called, Resource Discovery tools, we have come under tremendous pressure. Using Google as a benchmark, we are constantly being told that an FSS does not bring back properly ranked results, forgetting that it is Google that doesn’t necessarily bring back properly ranked results!

Recently, I went to the Enterprise Search Europe conference in London Olympia. It was a parallel conference to the Internet Librarian International conference; only this one is aimed at business enterprises. And the resources are mostly corporate knowledge base kept in-house; so search criteria is different. In Libraries we search almost all third party resources and our own Institutional repositories plus our catalogue. What amused me was the anecdotes like the CEO of an enterprise who thinks Google catches everything that is necessary for their work, and therefore “search doesn’t work!”

And it is here that I learnt about “slow search”! In an article called “Slow Search: Information Retrieval without Time Constraints” (http://dx.doi.org/10.1145/2528394.2528395), authors Teevan et al tried to show “how people experience and value speed during search, and explores the viability of slow search as an alternative to current speed-focused approaches.” They used two user surveys (to) reveal how people trade off quality and speed… The overall aim of (their) paper was to inspire additional research on how search experiences can be optimized when less constrained by time. (p.2)

Keane, O’Brien & Smith from University College Dublin conducted a study on the use of Google search engine in 2008 (http://dx.doi.org/10.1145/1314215.1314224). In my presentation I mentioned their findings that users are often “misled by the presented order of items”. They called this behaviours “satisficing”; the tendency to choose the most convenient and easiest route that leads to “good enough” information rather than the “best” information.

" A federated search system is like an archaeological dig; the more you narrow the field and dig deeper, the system delves deeper into those entrenched databases containing pertinent information."

” A federated search system is like an archaeological dig; the more you narrow the field and dig deeper, the system delves deeper into those entrenched databases containing pertinent information.”

UCC Library opted for a federated search system because we believe one size does not fit all. Students are more mobile now with differing information needs. Our aim is to provide maximum visibility to all our resources that would include ways of clever searching. Explorit provides an intelligent search system through their Latent Semantic Text searching that is not possible in a pre-indexed service like a discovery tool which depends on the metadata provided by the publisher. To me, a discovery tool is like a key to a store, all products neatly labelled and shelved. A federated search system is like an archaeological dig; the more you narrow the field and dig deeper, the system delves deeper into those entrenched databases containing pertinent information.

Another reason is purely based on economics. A higher ranking Research University attracts more grants; so it makes a business sense to enhance our researchers’ experience by providing the necessary tools.

There are a number of challenges facing the Universities and other Academic Institutions at present – public-private partnership, change in user demography bringing in a change in user behaviour. At the same time the Discovery/Search environment is changing rapidly; we need to look for the MVP in whatever product we use; a product that will not restrict growth, but is scalable and enhance an organic development of the Institutional knowledge repositories.

In our case, the use of the federated search system became our MVP, not restricting our growth, but enhancing it, letting us grow organically. I did one training session with undergraduates before I went to the conference, and here is one comment: “Enjoyed the fact that we can create our own search engine – very helpful!”

The system allows them to develop their own critical thinking skills. Our users are growing up!!!