Re: 0.6 and 1.0

[Sean Owen <sr...@gmail.com>]

(moving to dev@ per Jake)

I support moving towards 0.6. I think there are 1-2 issues assigned to me
that I'll close in the next few weeks and I do not plan any more work in
the foreseeable future. January sounds possible. As ever, it'll take as
long as you let it take -- it's more a question of being disciplined in
closing stuff that won't be addressed (or moving out those that you
honestly think will), and just fixing more issues than you open.

I don't think things are near enough 1.0 yet to worry about
backwards-compatibility. There are bigger questions first, like what might
be deleted, refactored, added or not added. Then it's possible to think
seriously about that.

I personally think there is too much unfinished/experimental stuff in the
project already. I'm less interested therefore in mechanisms for getting
more stuff in as "experimental". It feels like just shuffling papers to
mark these things rather than make plans and decisions about them.


On Mon, Dec 5, 2011 at 7:47 PM, Jeff Eastman <jd...@windwardsolutions.com>wrote:

> Opening this up as a PMC discussion for now. Seems like we ought to be
> driving a stake in the ground for 0.6 and starting the JIRA triage process
> so we can start to converge on that date. We have a lot of open issues and
> the list is growing so some triage will be necessary. I don't think Dec is
> possible, but could we get 0.6 to an RC1 state by January end?
>
> On the longer horizon, Isabelle posted a thread a while back discussing
> what sorts of backwards-compatibility guarantees a 1.0 release would imply.
> I was in the middle of house sale/packing/moving and did not have
> bandwidth. Hadoop is still 0.x and they are much farther into the
> production lifecycle than Mahout. What is the market force that is driving
> us for such a release?
>
> We've talked in the past about ways to partition Mahout into
> production-ready and experimental components as a way to set expectations
> about the amount of volatility in our APIs, CLIs, data formats, etc.
>
> With my new PMC hat on, I'm starting to think about Mahout as a product
> manager might think about it. I don't have any great insights yet but I'm
> looking at all our JIRAs in an effort to develop an appropriate overview.
>
> Cheers,
> Jeff
>
>