On 8/27/07, Tim Ellison wrote: > We've had some discussion, I think it is time to agree on whether we are > going to shoot for a M3 release soon or let things ride for a while longer. > > I hope I've made my view clear, that we should be addressing the known > test failures and exercising the additional tests to check that the > tremendous work that has taken place since M2 continues to enhance our > capabilities, scalability, and stability. > > Thoughts? > Tim >
Tim, Yes, we should have M3 for sure and we should have fixed/predictable milestone schedule. In case of 2 month schedule there are 6 milestones per year. But I'm afraid we will have in every second milestone situation when many folks won't be able participate in a milestone development/testing for 1 week or more due to different reasons. I think it is quite noticeable time in 2 month period (taking into accont 2 weeks of code freeze). IMHO this affect folks performance and releases quality. Because we need to run a lot of tests and have communication between many experts in different areas to release high-quality milestone builds. Say from my experience, I worked for a week after M2 and then went to vacations till the beginning of August. When I returned back to the office I had only few days to sort things out before code freeze. >From my POV such milestones are: * January-February: many holidays in different countries. * July-August: time for summer vacations. * November-December: not good to have milestone at December 3. I would move it to December 15. I think that 3 month schedule fits better for Harmony. Thanks, Stepan.
