The naming scheme sounds good. Since we want to start out sooner, I am assuming we are not limiting ourselves to two alphas as the email might indicate.
Also, as the release manager, can you elaborate on your definitions of alpha and beta? Specifically, when do we expect downstream projects to try and integrate and when we expect Hadoop users to try out the bits? For Hadoop 2, we started a more formal notion of compatibility, but that document has a lot of gaps where we don't have a policy. Should we be filling those gaps for Hadoop 3 and if yes, should that include wording around something being beta? On Mon, May 2, 2016 at 2:50 PM, Zheng, Kai <kai.zh...@intel.com> wrote: > Ok, got it. Thanks for the explanation. > > Regards, > Kai > > From: Andrew Wang [mailto:andrew.w...@cloudera.com] > Sent: Tuesday, May 03, 2016 5:41 AM > To: Zheng, Kai <kai.zh...@intel.com> > Cc: common-dev@hadoop.apache.org > Subject: Re: Release numbering for 3.x leading up to GA > > >> but I'm going to spend time on our first RC this week. > Sorry what does this mean? Did you mean the first RC version or > 3.0.0-alpha1 will be cut out this week? > Anyway will try to get some tasks done sooner. > First RC for whatever we name the first 3.0 alpha release. > > There's no need to rush things to make this first alpha, since there are > more alphas planned. > > That said, if you have changes that affect compatibility, the sooner the > better :) >
