Open Stack

On Tue, 2015-05-05 at 10:39 -0700, Joshua Harlow wrote:
> > James Bottomley <mailto:James.Bottomley at HansenPartnership.com>
> > May 5, 2015 at 9:53 AM
> > On Tue, 2015-05-05 at 10:45 +0200, Thierry Carrez wrote:
> >> Joe Gordon wrote:
> >>> [...]
> >>> To tackle this I would like to propose the idea of a periodic developer
> >>> oriented newsletter, and if we agree to go forward with this, hopefully
> >>> the foundation can help us find someone to write newsletter.
> >> I've been discussing the idea of a "LWN" for OpenStack for some time,
> >> originally with Mark McLoughlin. For those who don't know it, LWN
> >> (lwn.net) is a source of quality tech reporting on Linux in general (and
> >> the kernel in particular). It's written by developers and tech reporters
> >> and funded by subscribers.
> >>
> >> An LWN-like OpenStack development newsletter would provide general
> >> status, dive into specific features, report on specific
> >> talks/conferences, summarize threads etc. It would be tremendously
> >> useful to the development community.
> >>
> >> The issue is, who can write such content ? It is a full-time job to
> >> produce authored content, you can't just copy (or link to) content
> >> produced elsewhere. It takes a very special kind of individual to write
> >> such content: the person has to be highly technical, able to tackle any
> >> topic, and totally connected with the OpenStack development community.
> >> That person has to be cross-project and ideally have already-built
> >> legitimacy.
> >
> > Here, you're being overly restrictive.  Lwn.net isn't staffed by top
> > level kernel maintainers (although it does solicit the occasional
> > article from them).  It's staffed by people who gained credibility via
> > their insightful reporting rather than by their contributions.  I see no
> > reason why the same model wouldn't work for OpenStack.
> >
> > There is one technical difference: in the kernel, you can get all the
> > information from the linux-kernel (and other mailing list) firehose if
> > you're skilled enough to extract it.  With OpenStack, openstack-dev
> > isn't enough so you have to do other stuff as well, but that's more or
> > less equivalent to additional research.
> >
> >>   It's basically the kind of profile every OpenStack company
> >> is struggling and fighting to hire. And that rare person should not
> >> really want to spend that much time developing (or become CTO of a
> >> startup) but prefer to write technical articles about what happens in
> >> OpenStack development. I'm not sure such a person exists. And a
> >> newsletter actually takes more than one such person, because it's a lot
> >> of work (even if not weekly).
> >
> > That's a bit pessimistic: followed to it's logical conclusion it would
> > say that lwn.net can't exist either ... which is a bit of a
> > contradiction.
> >
> >> So as much as I'd like this to happen, I'm not convinced it's worth
> >> getting excited unless we have clear indication that we would have
> >> people willing and able to pull it off. The matter of who pays the bill
> >> is secondary -- I just don't think the profile exists.
> >>
> >> For the matter, I tried to push such an idea in the past and couldn't
> >> find anyone to fit the rare profile I think is needed to succeed. All
> >> the people I could think of had other more interesting things to do. I
> >> don't think things changed -- but I'd love to be proven wrong.
> >
> > Um, I assume you've thought of this already, but have you tried asking
> > lwn.net?  As you say above, they already fit the profile.  Whether they
> > have the bandwidth is another matter, but I believe their Chief Editor
> > (Jon Corbet) may welcome a broadening of the funding base, particularly
> > if the OpenStack foundation were offering seed funding for the
> > endeavour.
> 
> +1 to that, although lwn.net is partially subscriber only (yes I'm a 
> subscriber); so if say we had a 'openstack section' there (just like 
> there is a kernel section, or a security section, or a distributions 
> section...) how would that work? It'd be neat to have what we do on 
> lwn.net vs having a openstack clone/similar thing to lwn.net (because 
> IMHO we already make ourselves 'special' enough...).

These are just details that would need to be ironed out.  Lwn is a
business as well as a service, so it has to get paid somehow.
Subscriptions are how it currently does this.  However, the Linux
Foundation does sponsor lwn.net to produce some content (the Linux
Weather Forecast: http://www.linuxfoundation.org/news-media/lwf) which
is then freely posted.

Assuming you agree with the general principle that lwn.net could do the
job, OpenStack could choose to follow the LF model (pay for some of the
most relevant general stuff, but leave LWN to pursue subscribers who
want the details) or it could choose to develop its own model for
working with them.

James

> -Josh
> 
> > James
> >
> >
> >
> > __________________________________________________________________________
> > OpenStack Development Mailing List (not for usage questions)
> > Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:unsubscribe
> > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
> > Thierry Carrez <mailto:thierry at openstack.org>
> > May 5, 2015 at 1:45 AM
> >
> > I've been discussing the idea of a "LWN" for OpenStack for some time,
> > originally with Mark McLoughlin. For those who don't know it, LWN
> > (lwn.net) is a source of quality tech reporting on Linux in general (and
> > the kernel in particular). It's written by developers and tech reporters
> > and funded by subscribers.
> >
> > An LWN-like OpenStack development newsletter would provide general
> > status, dive into specific features, report on specific
> > talks/conferences, summarize threads etc. It would be tremendously
> > useful to the development community.
> >
> > The issue is, who can write such content ? It is a full-time job to
> > produce authored content, you can't just copy (or link to) content
> > produced elsewhere. It takes a very special kind of individual to write
> > such content: the person has to be highly technical, able to tackle any
> > topic, and totally connected with the OpenStack development community.
> > That person has to be cross-project and ideally have already-built
> > legitimacy. It's basically the kind of profile every OpenStack company
> > is struggling and fighting to hire. And that rare person should not
> > really want to spend that much time developing (or become CTO of a
> > startup) but prefer to write technical articles about what happens in
> > OpenStack development. I'm not sure such a person exists. And a
> > newsletter actually takes more than one such person, because it's a lot
> > of work (even if not weekly).
> >
> > So as much as I'd like this to happen, I'm not convinced it's worth
> > getting excited unless we have clear indication that we would have
> > people willing and able to pull it off. The matter of who pays the bill
> > is secondary -- I just don't think the profile exists.
> >
> > For the matter, I tried to push such an idea in the past and couldn't
> > find anyone to fit the rare profile I think is needed to succeed. All
> > the people I could think of had other more interesting things to do. I
> > don't think things changed -- but I'd love to be proven wrong.
> >
> > Joe Gordon <mailto:joe.gordon0 at gmail.com>
> > May 4, 2015 at 12:03 PM
> > Before going any further, I am proposing something to make it easier 
> > for the developer community to keep track of what other projects are 
> > working on. I am not proposing anything to directly help operators or 
> > users, that is a separate problem space.
> >
> >
> >
> > In Mark McClain's TC candidacy email he brought up the issue of cross 
> > project communication[0]:
> >
> > Better cross project communication will make it easier to share 
> > technical solutions and promote a more unified experience across 
> > projects.  It seems like just about every time I talk to people from 
> > different projects I learn about something interesting and relevant 
> > that they are working on.
> >
> > While I usually track discussions on the mailing list, it is a poor 
> > way of keeping track of what the big issues each project is working 
> > on. Stefano's 'OpenStack Community Weekly Newsletter' does a good job 
> > of highlighting many things including important mailing list 
> > conversations, but it doesn't really answer the question of What is X 
> > (Ironic, Nova, Neutron, Cinder, Keystone, Heat etc.) up to?
> >
> > To tackle this I would like to propose the idea of a periodic 
> > developer oriented newsletter, and if we agree to go forward with 
> > this, hopefully the foundation can help us find someone to write 
> > newsletter.
> >
> > Now on to the details.
> >
> > I am not sure what the right cadence for this newsletter would be, but 
> > I think weekly is too
> > frequent and once a 6 month cycle would be too infrequent.
> >
> > The  big questions I would like to see answered are:
> >
> > * What are the big challenges each project is currently working on?
> > * What can we learn from each other?
> > * Where are individual projects trying to solve the same problem 
> > independently?
> >
> > To answer these questions one needs to look at a lot of sources, 
> > including:
> >
> > * Weekly meeting logs, or hopefully just the notes assuming we get 
> > better at taking detailed notes
> > * approved specs
> > * periodically talk to the PTL of each project to see if any big 
> > discussions were discussed else where
> > * Topics selected for discussion at summits
> >
> > Off the top of my head here are a few topics that would make good 
> > candidates for this newsletter:
> >
> > * What are different projects doing with microversioned APIs, I know 
> > that at least two projects are tackling this
> > * How has the specs process evolved in each project, we all started 
> > out from a common point but seem to have all gone in slightly 
> > different directions
> > * What will each projects priorities be in Liberty? Do any of them 
> > overlap?
> > * Any process changes that projects have tried that worked or didn't work
> > * How is functional testing evolving in each project
> >
> >
> > Would this help with cross project communication? Is this feasible? 
> > Other thoughts?
> >
> > best,
> > Joe
> >
> >
> >
> >
> > [0] 
> > http://lists.openstack.org/pipermail/openstack-dev/2015-April/062361.html
> > __________________________________________________________________________
> > OpenStack Development Mailing List (not for usage questions)
> > Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:unsubscribe
> > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
> __________________________________________________________________________
> OpenStack Development Mailing List (not for usage questions)
> Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:unsubscribe
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev