Open Stack

On 3/7/2018 6:12 AM, Chris Dent wrote:
> # Talking about the PTG at the PTG
> 
> At the [board
> meeting](http://lists.openstack.org/pipermail/foundation/2018-March/002570.html), 
> 
> the future of the PTG was a big topic. As currently constituted it
> presents some challenges:
> 
> * It is difficult for some people to attend because of visa and other
>    travel related issues.
> * It is expensive to run and not everyone is convinced of the return
>    on investment.
> * Some people don't like it (they either miss the old way of doing the
>    design summit, or midcycles, or $OTHER).
> * Plenty of other reasons that I'm probably not aware of.

All of this is true of the summit too isn't it?

When talking about the PTG, I always hear someone say essentially 
something like, "you know, things would be better if we did <now 
describe exactly what the old design summit format was>". It's funny how 
we seem to only remember the last 6 months of anything.

> 
> This same topic was reviewed at [yesterday's office
> hours](http://eavesdrop.openstack.org/irclogs/%23openstack-tc/%23openstack-tc.2018-03-06.log.html#t2018-03-06T09:19:32). 
> 
> 
> For now, the next 2018 PTG is going to happen (destination unknown) but
> plans for 2019 are still being discussed.
> 
> If you have opinions about the PTG, there will be an opportunity to
> express them in a forthcoming survey. Beyond that, however, it is
> important [that management at contributing
> companies](http://eavesdrop.openstack.org/irclogs/%23openstack-tc/%23openstack-tc.2018-03-06.log.html#t2018-03-06T22:29:24) 
> 
> hear from more people (notably their employees) than the foundation
> about the value of the PTG.
> 
> My own position is that of the three different styles of in-person
> events for technical contributors to OpenStack that I've experienced
> (design summit, mid-cycles, PTG), the PTG is the best yet. It minimizes
> distractions from other obligations (customer meetings, presentations,
> marketing requirements) while maximizing cross-project interaction.

Agree.

> 
> One idea, discussed
> [yesterday](http://eavesdrop.openstack.org/irclogs/%23openstack-tc/%23openstack-tc.2018-03-06.log.html#t2018-03-06T22:02:24) 
> 
> and [earlier
> today](http://eavesdrop.openstack.org/irclogs/%23openstack-tc/%23openstack-tc.2018-03-07.log.html#t2018-03-07T05:07:20) 
> 
> was to have the PTG be open to technical participants of any sort, not
> just so-called "OpenStack developers". Make it more of a place for
> people who hack on and with OpenStack to hack and talk. Leave the
> summit (without a forum) for presentations, marketing, pre-sales, etc.

I don't understand why some people/organizations/groups think that they 
shouldn't attend the PTG - maybe it's something in the 'who should 
attend' docs on the website? But I hear time and again that operators 
think they shouldn't attend the PTG, but we know a few do and they are 
extremely valuable in the developer discussions for their perspective on 
how they, and other operators, run their clouds and what they want/need 
to see happen on the dev side. The silo effect between dev and ops 
communities is very weird and counter-productive IMO. And the Forum 
doesn't solve that problem really because not everyone can get funding 
to travel to the summit (Sydney, hello).

Case in point: the public cloud WG session held at the PTG on Monday 
morning where we went through the spreadsheet of missing features; I 
think I was the only full time core project developer in the room which 
was otherwise operators (CERN, OVH, City Network and Vexxhost were 
there) and it was much more productive actually having us sitting 
together going through the list and checking things off which had either 
been completed already, or were bugs instead of features, or that I 
could just say, "this depends on that and Jane Doe is working on it, so 
follow up with her" or "this is a known thing, it's been discussed, but 
it needs a driver (project manager) - so that's your next step". That 
wouldn't have been possible if the public cloud WG operators weren't 
attending the PTG.

> 
> An issue raised with conflating the PTG and the Forum is that it would
> remove the
> [inward/outward 
> focus](http://eavesdrop.openstack.org/irclogs/%23openstack-tc/%23openstack-tc.2018-03-07.log.html#t2018-03-07T08:20:17) 
> 
> concept that is supposed to distinguish the two events.
> 
> I guess it depends on how we define "we" but I've always assumed that
> both events were for outward focus and that for any inward focussing
> effort we ought to be able use asynchronous tools more.
> 

I don't get the inward/outward thing. First two days of the old design 
summit (ops summit?) format was all cross-project stuff (docs, upgrades, 
testing, ops feedback, etc). That's the same as what happens at the PTG 
now too. The last three days of the old design summit (and now PTG) are 
vertical project discussion for the most part, but Thursday has also 
become a de-facto cross-project day for a lot of teams (nova/cinder, 
nova/neutron, nova/ironic all happened on Thursday). I'm not sure what 
is happening at the Forum events that is so wildly different, or more 
productive, than what we can do at the PTG - and arguably do it better 
at the PTG because of fewer distractions to be giving talks, talking to 
customers, and having time-boxed 40 minute slots.

> 
> * Rather than having Long Term Support, which implies too much, a
>    better thing to do is enable [extended
>    maintenance](https://review.openstack.org/#/c/548916/) for those
>    parties who want to do it.
> 

Good lord I'm already regretting even thinking it would be a good idea 
(or fun) to just throw something up as a resolution based on previous 
discussions about all of this. Tony, save me.

-- 

Thanks,

Matt