[tc][election] campaign discussion: what we can improve in TC & how?

Radosław Piliszek radoslaw.piliszek at gmail.com
Sun Apr 5 20:07:21 UTC 2020


Hello Ghanshyam,

> What you think we should and must improve in TC ? This can be
> the involvement of TC in the process from the governance point of view
> or technical help for each project.
>
> - Do we have too much restriction on project sides and not giving them
> a free hand? If yes, what we can improve and how?

>From my current point of view, OpenStack TC is very liberal.
I base my opinion on some discussions of yours I read on ML and IRC and
also the non-observability of TC influences in Kolla. :-)
I think the current level of control is just right for many projects.
But maybe not all. I guess this is a good question to ask all
OpenStackers rather than just us.

That said, I believe it is wise to consider this broad topic in the
context of the recent Ironic thread on this ML.
The not-so-well-defined goal of TC for the upcoming times would be to
redefine OpenStack as something more (or maybe even "else") than open
source platform for doing IaaS. OpenStack is, as the name gladly
suggests, a stack, a pile of open source software projects, mostly in
Python, sharing common quality standards (or at least trying to!) under
TC guidance. It should be considered laudable to be part of OpenStack
rather than seek a way to escape it. If it is not, then we might as
well disband this and go home (btw, #stayhome).

As for simpler matters, TC might assume and advertise its role as
coordinator of cross-project efforts. And I don't mean current
community goals. I am thinking: if someone sees that by using project X
and project Y one could potentially achieve great thing Z, TC should be
offering its guidance on how to best approach this, in coordination
with cores from the relevant projects, and not in a way that enforces
TC to always intervene. Note this idea aligns with possible upcoming
TC-UC merger.

> - Is there less interaction from TC with projects? I am sure few
> projects/members even do not know even what TC is for? What's your
> idea to solve this.

I think this is partly because OpenStack core projects are considered
very mature. Continuing on the thought of control, quality and prestige
associated with OpenStack, a good short-term goal would be to revisit
the OpenStack projects and possibly restructure/deprecate some that
need this - considering both integral usability as well as standalone.

I don't think TC transparency needs 'fixing'. This is actually good
thing (TM) - as long as projects deliver quality we expect, that is.

-yoctozepto



More information about the openstack-discuss mailing list