I recently had a conversation about search with my sister, who’s a college librarian. It was interesting to (for once) think about search outside of its marketing potential. She told me about about students who type in natural language search queries (sometimes simply typing in the name of their class or assignment title) and dreamed of a search engine that will understand exactly what searchers are looking for.
Discuss how search engines could make search easier.
I mentioned Eurekster – suggested that professors share their personal search data bases with students. In that scenario the student could type in the vaguest of terms and likely receive relevant results.
But, she countered, how will these searchers learn how to search in the real world?
I suggested a Search help scenario where searchers who needed extra help drilling down could receive specific suggestions, similar to that (excruciating and annoying) clippy that helps beginners on their MS Word docs. You could toggle this on or off as needed, and perhaps it would even pop automatically if it detects obviously natural language searches. (There would have to be a permanent kill button for this function though.)
She mentioned the reference librarian’s standard questions and we talked for a while about automating this process, and I decided it was time to talk to Gary Price.
Gary’s a librarian, information research consultant and author of the ResourceShelf blog and he offered some insight into which engines are beginning to offer the sorts of innovations that will make searching easier for beginning searchers.
“For what your sister and I would call ‘ready reference’ questions,” he said, “all of the engines are starting to offer help. I think this is important. This idea began with AltaVista several years ago (Shortcuts) and are now in place at Yahoo, Google, and Jeeves.”
“For example, type ‘baseball facts’ into Yahoo and get info at the top of the results list directly from the Columbia Encyclopedia.”
He’s also a big fan of Jeeves and their Famous People Search as well as Smart Answers.
He mentioned Vivisimo and their dynamic clustering as a possible means of making search easier. Vivisimo is an enterprise search solution – they also offer a search on their site that pipes in results from various indices such as DMOZ, LookSmart, Lycos, and Inktomi, among others.
In addition to displaying results, along the left of the page Vivisimo clusters sites into categories it detects through snippet text, url, title, etc. This clustering mechanism could be especially useful to beginning searchers who use broad terms such as “car.”
For “Car” Vivisimo offers clusters like “Quotes, Car Prices,” “Classic,” “Motor,” and more, which presents a potentially overwhelming number of sites into a manageable array of categories. This type of sorting could give beginning searchers a means of viewing a broader selection of results at one time and more immediate access to the information they seek.
Yahoo Shopping’s smart sort “sliders” came up in our talk too – Gary suggested a beginner search engine function could allow users to toggle the type of result they’re looking for. Imagine a combination between Vivisimo’s clustering model and sliders – users could continue their search from the results page without typing in any further search terms. These “sliders,” as Gary pointed out, are an intuitive way to interact with results. (I have trouble calling them sliders – that’s what we call greasy White Castle hamburgers where I’m from 🙂
Google has similar toggles in their new personalized search offering, though its category approach did little to make searching any more personal.
Currently Yahoo and AskJeeves both offer suggestions for broad, general searches like “cars,” – used cars, cheap cars, etcetera. Yahoo’s are positioned up near the search box while Jeeve’s run along the right side of the page. The suggestion function seems to work well in both search engines for the term “holocaust,” which a student might use as he begins his research.
Google, while spotted testing a suggestion function, does not currently offer it.
“As web databases get larger and larger and the battle for the top spots on a serp becomes even more fierce,” Gary said, “it’s going to be important for the major web engines to provide easy and intuitive ways to help the searcher refine and focus their searches.”
Index bloat may also lead to “more specialized and niche search tools,” he said. “Often, a more focused database can return very precise results and can also offer greater usability to the data. Just look at all of the ways to search the IMDB. Amazing!”
Another trend Gary noted is federated search. “As federated search technology improves it will also be possible to search databases (free, specialized, fee-based) from a common interface that the user or organization can design for their specific needs.”
Search tools are obviously becoming more prevalent in the library, but can they ever replace the librarian? I know my sister’s not worried. And as Gary says, search innovations “offer plenty of opportunities for the librarian to not only help the searcher FIND the answer but also teach what tools are available and how to best use them.”
Garrett French is the editor of Murdok’s eBusiness channel. You can talk to him directly at WebProWorld, the eBusiness Community Forum.
