Open Stack

Mark,

On May 7, 2013, at 4:12 AM, "Mark McLoughlin" <markmc at redhat.com> wrote:

> On Mon, 2013-05-06 at 22:38 +0000, Adrian Otto wrote:
>> In reference to the Oslo Wiki[1] , in the Incubation section,
>> principle number 2:
>> 
>> "The API is already in use by a number of OpenStack projects"
>> 
>> And in the Incubation section:
>> 
>> "The process of developing a new Oslo API begins by taking code which
>> is common to some OpenStack projects and moving it into the
>> oslo-incubator repository."
>> 
>> Where should cross-project code be incubated in the phase *before* it
>> is implemented by multiple projects? The Heat and Nova teams are
>> working together on StructuredStateManagement[2] and want to start
>> something new, but are struggling to determine where to put it. This
>> is probably evolve into an Oslo library, but it's not obvious where it
>> should live in the mean time. Should it be a splinter off of Heat? In
>> Nova? Neither option seems quite right.
> 
> Ideally, when a new API is proposed to Oslo, I'd like to see it already
> successfully in use by one project and that there is an immediate plan
> (preferably a WIP patch) for a second project to use it.
> 
> If there were WIP patches for two projects using the new API, that'd be
> cool tool.
> 
> So, it really depends - do you want to tackle this first in the context
> of Heat and then adapt it for Nova? Or the other way around? Or tackle
> both at once?
> 
> Pragmatically speaking, I'd recommend you first make sure you understand
> Nova's requirements, then start coding it in Heat but keep its
> implementation generic and once it has gotten merged into Heat look to
> having Nova adopt it.
> 
> Baby steps is what gets stuff done in OpenStack. The concern I'd have
> with tackling both projects at once is that you might put a tonne of
> work into it and neither project adopts it.

We have contributors from four companies in both Nova and Heat projects working together on this one with a spirit of collaboration that I have never before experienced. It would be extremely inspiring if we could find a way to make this work. The amount of red tape is getting to the point where it is now hindering innovation. There are plenty of smart developers on this effort who simply can't see a clear path to proceed.

Can we expect either Heat or Nova to accept patches that require code from a StackForge project? If not, then we have a rather serious paradox here.

Respectfully,

Adrian Otto