Integrated search & news


I gave a presentation of what was termed “integrated search” and “news” at NTNU Library’s internal seminar last Thursday (2009-05-28); what I presented can be characterized in the following:

  • context-based services
  • RSS/Atom
  • Search

The last two aren’t really that interesting, but the first item really is.

I made the point that our users are dependent on profile-services (i.e. services where they have a username and password that links them to an account that contains some details about them). These profile services often contain information that is relevant to their research/study interests (remember that we’re talking about a university library here).

I created two mock-ups, one of the university intranett, where a member of staff or student will have their department listed as part of their profile data; and a second one based on the learning management system in use at NTNU, Its:Learning. In the latter, it is possible to specify search sources including recommended reading, and resources linked directly to a programme of study.

Connecting the profile data on their interests with a set of resources/news feeds can hardly be described as difficult, so why is this not yet being done? To be honest, it is difficult for the library to get access to the necessary APIs to enable us to link our data to the data in these restricted, profile-based resources. We had extensive communication with Its:Learning, only to be told that they weren’t interested in working with us (they had other priorities). A shame really in this latter case, because we pretty much had everything ready from our side.

The idea of creating relevance by linking known research/study interests together with the library’s resources is not new, and in fact I have just taken receipt of the result of a customer-driven project by some students from NTNU’s computer science department. The students delivered a nice example of a feed aggregator that can also classify feeds according to “Norsk inndeling av vitenskapsdisipliner”, a Norwegian classification system for scientific disciplines. This kind of aggregator was designed to slip easily into a model where the data si fed into other systems, preferably using the APIs mentioned above. In fact, that was the whole point (the students did a good job, by the way :)).

It was pointed out to me that Google (Scholar) rules the roost when it comes to search, but I still reckon that providing quality information to academics based on relevance criteria will provide time-savings compared to the Google alternative. Google Scholar is a really good tool, but it just can’t beat the relevance of information in the small, commercial subject databases that the library provides its users. The problem is getting the users to look at these databases first — and it is here that the integration of search is a big issue.

The news approach provides links to the latest results from the databases and journals and is really a supplement to searching, it saves users time by providing users with the newest research in their field. Google can’t really do this acceptably well yet (sort by latest is hit and miss at best), so we have a good reason to provide this kind of service.

%d bloggers like this: