From: Doug B. <do...@we...> - 2008-09-01 12:28:58
|
What is the current status and effect of using a stopwords list configured in the <stopwords> element? I'd appreciate any general or specific insight to the questions below. I have found discussion archived in the exit-open list of various problems with using a stopwords list, but no clear resolutions, and can find no documentation on it. The default conf.xml in the 1.2.4 download has a <stopwords file="stopword"/> element but there is no default stopword file in the download. 1. What should the format of a stopwords file be, e.g. one word per line? 2. How about character encoding? 3. What indexes does it only apply to? (The element is outside of the fulltext element.) 4. What happens with &= queries that include both stopwords and non-stopwords? 5. With no stopwords list, I always get an error log whenever I start up the database. If I don't want a stoplist, can I simply comment out the element, or should I provide an empty file, I saw a email on the list where I understood Wolfgang to recommend an empty list. Thanks, Doug Douglas Avison Black West Rock Visions 755 Prospect ST. #205 New Haven, CT 06511 Office: 203-764-9401; 203-764-9494 ext. 134 Mobile: 203-676-5228 do...@we... |