Being a subproject would be good; we're pretty loosely coupled with
Hadoop at this point.

+1.

--Ari

-- 
Ari Rabkin [email protected]
UC Berkeley Computer Science Department

Reply via email to