Configuring RRC Folder Structures

What are the best practices that should be followed when configuring RRC folder structures?

I have read the following article and have recommended a folder structure that consisted of both functional areas and artifact types.  There would be high-level functional area folders such as UI, DB, etc.  And then under each of those folders there would be a consistent group of folders that are based on artifact type.  That way the UI team, for instance, could easily find all of their requirements.

• http://www.ibm.com/developerworks/rational/library/10/using-stakeholder-collaboration-strategy-with-requirements-composer-part2/index.html
  

Are there any gotchas associated with this type of folder structure?