From: "s.ross" <cwdinfo@...>
Date: 2008-03-25T06:07:04+09:00
Subject: Re: For Comment: Gem Browser

On Mar 24, 2008, at 11:43 AM, Thomas Preymesser wrote:
> it does not find " Summary: roo can access the contents of  
> OpenOffice-,
> Excel- or Google-Spreadsheets" if i search for 'spreadsheet'.

Yes, this is a limitation of the tokenizer in ferret. If you type in  
spreadsheet* you will get the results you expect (I think). I'm  
considering making wildcard searching the default, but many developers  
feel that they are "smart enough" to tweak their own searches, which  
is why I left it with default behavior initially.

roo contains the word Spreadsheets, but it is in the context "Google- 
Spreadsheets", so it is not tokenized as Google, Spreadsheet but  
rather as "Google-Spreadsheet". Similarly, for spreadsheet-excel, the  
tokenizer kills it.

If you are looking for the most results, use a wildcard at either end  
(*spreadsheet*). If you think you're close, but not sure, use fuzzy  
matching (spreadsheet~), and refine your searches like this:  
spreadsheet~ AND open*.

--steve