linaro-project-management team mailing list archive

[Loïc Minier <loic.minier@xxxxxxxxxx>]

On Fri, Mar 02, 2012, Deepak Saxena wrote:
>                                  I propose we track our deliverables
> against kernel merge windows and break roadmap items into smaller
> chunks that across multiple windows.

 It's a great idea; interestingly, in discussions around a new roadmap
 process Kiko is also proposing tracking the upstreaming of kernel
 patchsets as a core part of the process.

 You're completely right that we should plan upstream deliverables
 against the upstream release trains.

 For certain features, members care more about this or that feature
 being implemented in an upstream kernel -- "the right way" -- even if
 it takes longer; this would fit these cases nicely.  For other features
 they need to know at which date they will be able to consume some
 production-ready patchset, even if it's not fully upstreamed yet; this
 is because they need to plan their own product trains.

 Kiko pointed out that we currently lack good history around
 patchsets/features we're developing; we have cards tracking the high
 level idea, sometimes verbose workitems with blueprints, and of course
 there are mailing-list discussions, but it's hard to answer questions
 such as:
 * when/where were current and previous versions/iterations of the
   patchset published?
 * roughly, how far are we from being in mainline?
 * when/where will the next version/iteration of the patchset be
   published?

 Let's take some time later today in our call to discuss this further
 :-)

-- 
Loïc Minier