incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Valentin Kulichenko <>
Subject Re: [DISCUSSION] Graduate Ignite from the Apache Incubator
Date Thu, 23 Jul 2015 23:25:49 GMT
Hi Ted,

I think the main question here, is whether Ignite community fully following
the Apache Way, or in other words has an active meritocratic community
which is open and makes it easy for people to join.

In my opinion, absolutely Yes.

The project guidelines are fully documented on the website and Wiki,
guidelines on how to contribute and become a committer are documented as
well. All emails about joining the community and becoming a contributor are
answered in timely fashion [1]. In fact, Ignite has one of the easiest
barriers to become a committer - all you need to do is contribute at least
one patch to qualify [2].


Is the community open or has some secret Jira or other types of dicussions?
I have been on the PPMC from the very beginning, and believe we now have
come a long way. I believe that everyone in the community, including old
and new members, have absolutely embraced the Apache Way. There are no
secret Jiras or secret discussions taking place of any sort. Quite frankly,
I personally find it hard for myself to code anything without discussing it
on the dev list first, and I believe that everyone in the Ignite community
feels the same way. I think that most people looking at our project would
agree that we have a very active dev list and everyone in the community is

Do all issues get discussed on the dev list? I would say we are very close
to it. Yes, of course, there is some Skype communication given that the
project is very complex and the team is distributed. But if you look on the
dev list you will see that  all (or close to all) of the discussions do
find their way to the dev list and the community does participate very

I do agree that our Jira handling could be better and believe that
community has already responded to these discussions and addressed some of
the raised concerns. The truth is that so far many Jira discussions have
happened on the dev list, including community members sending notifications
about starting and ending work on Jiras and discussing Jira issues on the
dev list as well. This was a preferred way selected by the community that
we followed. I do agree that Jiras should be updated better and will
encourage everyone to do so going forward.

Hope this clarifies some points raised here.


On Wed, Jul 22, 2015 at 8:31 PM, Ted Dunning <> wrote:

> On Wed, Jul 22, 2015 at 6:54 PM, Valentin Kulichenko <
>> wrote:
> > Sorry for some late responses. It so happens that most of the community
> > members are on european time zones.
> >
> Thanks so much for jumping in!
> > > I recognize that the activity on the ML is diverse FINE !
> > > Concerns have been raised about a off-list issue system, that seems to
> be
> > > left open ?
> > >
> >
> > As Dmitry has already responded, there is absolutely no separate Jira
> > maintained for Ignite. Denis accidently sent his employer's Jira ticket
> > status to the dev list. In fact he is so used to send the most of his
> > emails to the dev list, that he sent this one to the dev list by accident
> > as well.
> >
> Valentin,
> As an interesting test of point of view, could you approach the project as
> if a stranger?
> Look at the mailing list.  Look at JIRA.  Try to understand what was
> resolved from the evidence you see.  Look for evidence of how the design
> for the solution was arrived at.
> Take for instance IGNITE-1134.  This is a hang after some stimulating
> event.  This sort of problem is often caused by subtle consistency issues
> in distributed systems.  This JIRA was resolved 21 hours after it was filed
> with no discussion or review as far as I can tell.  It was closed 20
> seconds after resolution. How can an outsider be part of this process?
> After take a look from that perspective, please tell us what you think
> about whether design decisions are being made on the list. My view is that
> they appear not to be, but you can probably say more from the inside of the
> project.

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message