Open Stack

On Wed, 2019-01-16 at 12:33 +0000, Chris Dent wrote:
> On Tue, 15 Jan 2019, Fox, Kevin M wrote:
> 
> > If its optional, then that might not happen properly. Delegating that authority might be a good solution as you
> > suggest. But then delegation of prioritizing of reviews and other such things might need to go along with it at the
> > same time to be effective? The idea is for there to be a flow of leadership from the TC to the project for some of
> > the cross project work, but if that flow cant happen, then nothing really changed. Not so sure a delegate can have
> > enough power to really be effective that way? Thoughts?
> 
> Delegation also simply spreads things around over a wider base,
> rather than addressing one of the core issues: there are too many
> things going on at once for there to be anything could be labeled
> a "unified direction".
> 
> If we want (do we?) a unified direction, there have to be fewer
> directions. Positioning things as a "a delegate will do it" is a way
> of choosing to not limit the number of things being done.
well the delegate suggestion was more. if a PTL feels they are too
busy with there other duties but in the comunity and in there company
they could ask another core member to attend the TC meeting on there behalf
so the number of voice would not change vs the PTL attending.

i had assumed it woudl be the role of the TC liason be that the PTL or a delage
to bring back a update to the project team e.g. nova in the nova team meeting.
collectivly the core team would then discuss the direction form the tc and work
with there contiubtors to determin if that direction is viable and either implement
it or provide feedback to the tc as to why it is not viable for the project.

i think a lot of the valosity of openstack has come form the fact that
it has tradtionally been grass roots driven. it is true that it has also slowed
progress on some cross project issue and in some cases internal project topic where
different compaines or contiutors have differnet prioties but i also dont how
the TC would ever truly mitgate company priorits conflicting with TC goals.
> 
> There are two competing voices around these issues, both have merit,
> but they remain in competition.
> 
> One says: the role of the TC is to enable healthy contribution in
> all its many and diverse forms in the context of a community called
> OpenStack.
> 
> Another says: the role of the TC is to help shape a product [1]
> called OpenStack.
> 
> The first sounds rather positive: We're enabling lots of stuff, look at
> all the stuff!
> 
> The second is often perceived as rather negative: That stuff could
> be so much better!
> 
> None of these perceptions are complete or fully accurate, there are
> positives and negatives in all views, but what seems to come out of
> these conversations is that there are some very small and vocal
> minorities that care about these issues and have strong opinions,
> and everyone else doesn't care. Do they (you, silent readers!) not
> care, are they not interested, or is it a matter of "we've been over
> this before and nothing ever changes?"

personally i have always seen the role of the tc in providing guidence on
policies that govern openstack, its process and long term direction but
not in providing direction to indiviual projects directly. e.g. engaging
with all project to develop and implement a timely transtion to python3 first
and python 3 only is  tc role in my view, both in terms of makeing sure a timeline
is cerated and a process to get there is formulated but i dont nessicarly
think its the tc role to guide a specific project say swift or nova in how they should
migrate to python3 specifically. 

e.g. if swift had a lib depency that only it used and it
needed to change it to move to python 3 then the swift core team should be free to make that
decision without tc guidence. as a counter point if the libary was used by several project
say eventlet and supose eventlets was not going to support python 3 ever then i think
the TC would have cross project role in determining a path forward so we can have a similar
technology stack across our projects.

so if i was to sumerise my personal view point i think the TCs role is largely in the form
of facilitating cross project goals (enabling openstack to "power" the edge) and governace/policy
like the PTI, but i think indiviual project should still have the freedom to priories there work
and define how they achive that goal and if it is relevent to them. i dont think devstack for example
need a dedicated effort to make it ready for the edge, nova and neutron might need just a tad more effort
if they wanted to achive that goal.
> 
> [1] I bristle at the term "product" because it sounds like we're
> trying to sell something. I'm thinking more in terms of "a thing
> that is being produced and has a purpose".
i have a similar reaction to the term product.