On Sep 7, 2011, at 9:15 AM, Arun C Murthy wrote: > Using real data helps - from Apache Jira, here are the statistics for work on > trunk/hadoop-0.23 in Q3 of 2011 (i.e. last 2 months alone): > > Hadoop Common - 224 resolved* jiras > Hadoop HDFS - 153 resolved* jiras > Hadoop MapReduce - 161* resolved jiras > > Total of 538 jiras resolved. I'm sure LOC is much more impressive, but be > that as it may.
I wonder about the usefulness of those stats though. Many of those jiras are bug fixes to other jiras committed in the same time frame and also committed to other branches. Also, I've noticed an absolutely explosion of sub-tasks where one big patch is actually done at commit time. > So, I'd say it's pretty clear that there is significant interest, involvement > & investment from the wider community. That makes me believe that Apache > Hadoop is moving along the right path i.e. forward. Part of the problem that the Hadoop community has is a PMC and committer group that leans heavily towards a few organizations. Any movement that those organizations do offsets what the rest of the community may or may not want to do. The reality is that if anyone who isn't HortonWorks or Cloudera wants to do a release, it is likely doomed at PMC vote time. [I'll hold off on commenting on the effectiveness of our current PMC member roster. That's a different discussion altogether.]