This article on the most excellent Federated search blog outlines some of the unrealistic expectations that people have related to federated search. In essence, we’ve heard many of these before, but there are some good points about what to expect. One of the points that got my attention (possibly because of the assumption that it’s only librarians that think that native interfaces have their advantages) was:
Federated search should be as good as searching the native sources. The librarians are right. If users would take the time to learn the query syntaxes and quirky behavior of every source they care about and if they are willing to perform advanced searches against each source, and manually deduplicate the results, then they will get more relevant results. But, most users are not interested in doing this.
It’s interesting that the oft-cited truth that you always get better results in the native interface is actually true; the problem is that users don’t have the time or the skills to perform these searches. To be honest, I’m not sure that this is entirely true; for one, a user who is sick and tired of getting weak results will actually attempt to learn the necessary skills (or approach a librarian). Secondly, one of the major problems with federated search is that it anonymizes resources, and assumes a familiarity with resources that users who don’t have prior experience with the databases they are searching — a good example of this is NTNU’s Metalib installation that seems purposely to hide information about resources from users.
Don’t get me wrong, federated search has it’s uses, and can prove to be extremely useful, it’s just never going to replace native interfaces in any meaningful way. The reason for this is as always the added value you get from native interfaces — you’re searching for specific information in a specific place, and the interface will reflect this. A reason why Google’s interface is so weak in many respects — too little focus on content presentation and too much focus on speed (you do need to know that this particular search took 0.6776 seconds?) — is that it tries to be all things to all people, and information just doesn’t work like that.
What I miss in federated search is some way of gathering information sensibly from not-so-diverse resources, say within a subject area like literature. Presenting an interface for and information from these resources should entail specific differences in the federated search interface, both from the perspective of what search options are given and how the results are presented. This kind of thing isn’t difficult (especially as you’ve got the criteria listed in connection files).
Tags: federated search