SIFT Implementation

Collect new documents daily from source

Match each article against stored queries

Generate matching files - daily and non-daily

(can use indexes for matching)

Sort files by (user, profile) and mail document to users