A few weeks ago now we relaunched the new search. Here are the new features that we incorporated:
- New Design
We took some of the anecdotal use information and combined it with some user-testing to come up with a new tabbed design to replace the three-column design we’d used for the prior two years. Among other things, we found that the right-most column seemed to be “invisible” to people (we think because the Google-trained expect that column to be ads) and we wanted to add additional sources and the columnar design did not allow for that.
- New Term Parser
Before we actually hand your search terms off to the part of the search engine (called sphinx) for the actual search, we take some time to massage your terms to maximize sphinx’s ability to return the best results. Nick rewrote the search parser from the ground-up for this release, and it is much more robust and changeable now.
This new engine now offers filtering. What is filtering? On the People tab for instance, you can now sub-segment your term search to only show faculty, staff or students, or college-, graduate- or law-associated individuals. There are time-based filters for news and events. Give them a try and suggest others.
- New Did You Mean?
Our prior Did You Mean? resource was only invoked when your search returned no results, and utilized Yahoo’s spelling API to give a good suggestion. It worked quite well, but Yahoo has since killed their spelling API and we wanted to provide this service even when the we were returning some results. The new Did You Mean? now runs on every search and if there are relevant results, shows them. (We also switched to using Google as a source.) Our next iteration of this feature will pre-check the suggestions against our database before showing them.
- More Prominent Feedback
We want your feedback on this engine to continue to improve it’s performance and allow us to help it evolve with our community over time. Use it.
- Google Returns
We have built a Google tab right into the new search engine so you can try it when you aren’t finding results via the search engine itself. In general, our engine will outperform Google when dealing with people and recommended results — that is, for searches where we can apply our inherent knowledge of Lewis & Clark. Google will outperform our engine when the search is more amorphous or ambiguous.
But as ever, this is just the beginning. In the weeks since relaunch, Nick and I have fixed issues raised by you, and continued to add new features. Just in the last week, we’ve:
- moved the recommended results (we heard that people weren’t seeing them);
- added the ability to add a person to your address book from the search engine; and
- added the ability to add an event to your calendar.
Have an idea for us? Tell us! Email email@example.com or better, use that search engine feedback!