SUBJECT: workplan Dear sir , Thanks for your time yesterday . I wrote a script that finds the number of documents relevant to the topic code and ran it on the following &NUM ) Broad level codes like &NAME ( no of docs in the corpus : &NUM , mem reqd : ~ 800MB ) and &NAME ( no of docs : &NUM , mem reqd : ~ &NUM MB ) &NUM ) Fine grained codes like C17 & its subclasses ( no of docs : &NUM , mem reqd : ~ 860MB ) and C18 & its subclasses ( no of docs : &NUM , mem reqd : ~860MB ) I was thinking I could do something similar to &NAME &NAME - classifying text into broad classes like &NAME , &NAME and further into finegrained categories like C181 , C182 , C183 and then compare the results . The memory requirments for this would be in the order of upto &NAME as the upper limit . Look Fwd to your confirmation , many thanks , &NAME &NAME . It 's fast , it 's easy and it 's free . Get &NAME &NAME today ! &WEBSITE