My votes (non-binding) > Release Goals: > - refactoring and cleanup of existing functionality
+1 And speeding up the build > - new functionality +1 for improvements or things that make existing Mahout features easier to use, like MAHOUT-833 (making seqdirectory MapReduce) and MAHOUT-944 (lucene2seq). -1 For introducing entirely new algorithms > - Mahout 0.7 +1 > Mahout 0.6.1 -1 On Thu, Mar 1, 2012 at 3:53 PM, Jake Mannix <[email protected]> wrote: > On Thu, Mar 1, 2012 at 3:34 AM, Grant Ingersoll <[email protected]> wrote: > >> >> On Feb 29, 2012, at 9:50 AM, Jeff Eastman wrote: >> >> > We have been discussing these two topics on this list for a while now >> and the PMC would like to invite all Mahout developers to weigh in on >> calling the questions. Please submit your votes in the range -1,...,+1 for >> the following 4 options: >> > >> > Release Goals: >> > - refactoring and cleanup of existing functionality >> > > +1 > >> - new functionality >> >> +0 - I'd say minor stuff is OK, as it probably could be called cleanup >> anyway. For example (not implying this will happen), new APIs for >> accessing the Hadoop functionality instead of being cmd line driven. Also, >> the parallel Baum Welch is replacing existing functionality with a M/R >> version and has been baking for a long time. >> > > +epsilon > > If the "new functionality" is something cleanup-y or has been sitting > around a long time, then "cleaning up JIRA" is cleanup too. > > >> > >> > Release Name: >> > - Mahout 0.7 >> >> > +1 > >> - Mahout 0.6.1 >> > > -1 Until we get to 1.0, I think patch versions like this are pointless. >> > > -0 > > I think its pointless until we can do a release a month. If/when we can do > that, then we can start doing patch releases all the time. > > -jake
