Open Stack

On 5 December 2013 12:34, Tim Schnell <tim.schnell at rackspace.com> wrote:
> Hi Heaters,
...

So, I'm very glad someone is going to work on this; I believe there
may be use for it in Tuskar too.

> Start Heater as a Stackforge project with a different core team that is
> dedicated to actively working on Heater

This would silo the folk working on the repository from those working
on Heat. -1000 on that proposition.

> Incubate Heater within the Orchestration umbrella using the existing Heat
> Core team
> Incubate Heater with the Orchestration umbrella, but create a sub-project
> team responsible for reviewing and +2s

This would reduce the siloing effect, but you'd still have siloed core
reviewers.

> The idea behind creating a separate core team either via Stackforge or an
> Orchestration sub-project is so that the people actively working on Heater
> can review and iterate more quickly through code revisions than dumping
> Heater code through the already strained Heat review pipeline.

So, I understand the motivation, but this is really really really
really bad thinking.

If Heat is suffering velocity problems due to not enough core reviewers do you:

A) start a related project, which all the Heat core are going to have
to be up to speed on at some point

or

B) add people onto Heat with the explicit goal of providing more
review bandwidth until the problem is solved ?

If you answer A, we need to talk.

Seriously.

Where there is a bottleneck - such as not enough reviewers -
/anything/ we do that makes the bottleneck better makes everything
better. Anything else we do has pretty much no effect, other than
costing time and effort to do it, whatever it is.

-Rob


-- 
Robert Collins <rbtcollins at hp.com>
Distinguished Technologist
HP Converged Cloud