On 02/08/11 19:05, Peter Linnell` wrote:
On 08/02/2011 01:57 PM, Andrew Bayer wrote:
I'm personally neutral on this. I've had more experience with
commit-then-review, but as Andrei pointed out, review-then-commit is more
common in the Hadoop world.
I do think that at this point, when there are still fairly trivial
changes
to be made in terms of license headers, name changes, etc, that a hard
and
fast review-then-commit approach may be a little over the top. Some
flexibility definitely makes sense for now.
CTR is far more agile and used in most other ASF projects. RTC slows
down hadoop and creates more veto conflict -it may be appropriate when
the value of HDFS is the IPO value of Facebook and others, but not for
bigtop.
