Thursday, March 11, 2010

Stemming, Spell Check, and Wildcards

The High Plains Library District's catalog only retrieves results that exactly match your search terms. This is one reason it is very important that your search terms are spelled correctly. In addition, unlike many modern search engines, it doesn't support stemming 1.

While our catalog doesn't provide spell check or stemming, it does allow the use of wildcard characters. A wildcard character is a symbol that can be used in place of other letters. In our catalog, the wildcard characters are an asterisk, *, and a question mark, ?.

How do I use wildcard characters?
The asterisk is a multi-character wildcard. This means it matches zero or more letters. For example, if I submit a catalog search for infant*, the catalog will return results that match infant, infante, infants, infantry, and more.

The question mark is a single-character wildcard. It can be used to replace exactly one letter. If I submit a catalog search for infant?, the catalog will return results that match infants and infante but not records that only contain the words infant or infantry.

Why would I want to use wildcard characters?
Wildcard characters are useful when you wish to force the catalog to do basic stemming or are unsure of how to correctly spell a word. For example, let's say I heard someone talking about the recipient of the 2009 Newbery Medal but all I remember is that the title contains the words grave and book. Searching for grave book returns no results. However, grave* book reveals the book I'm looking for is The Graveyard Book by Neil Gaiman.

Speaking of the Newbery Medal, I can never remember if Newbery should have one "r" or two. If I can't remember while I'm doing my catalog search, I can find results for both Newbery and Newberry by searching for newber*y medal.

Obviously, it would be easier if our catalog automatically returned the results we want. However, until we upgrade to a smarter catalog, using wildcards can be very helpful.

1Stemming is the process of finding the root word of your search term and retrieving results with the same root word. For example, if you search Google for the word cook, it may also return words such as cooks and cooking.

UPDATE: Our catalog still can't do stemming but our Did You Mean feature is finally live!

No comments:

Post a Comment