You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hawq.apache.org by Roman Shaposhnik <rv...@apache.org> on 2016/04/02 00:13:38 UTC
Concern about under-communication in the HAWQ community
Hi!
for the past month or so I've been noticing that there's
clearly under-communication going on in the HAWQ
community. In fact, the more activity I see around code
going in with next to 0 activity on dev@
A community needs conversation. Discussion. Consideration
of topics, roadmaps, and feature development.
Here's the most recent example of that:
https://issues.apache.org/jira/browse/HAWQ-410
where a commit recently went in without the JIRA ever
being updated or a solution negotiated/reviewed with
the original submitter. In fact, the content of that commit
is not problem free. For example, the following really
needs to be fixed:
https://github.com/apache/incubator-hawq/blob/master/depends/libhdfs3/README.md
==============================================
Libhdfs3 is developed by Pivotal and used in HAWQ, which
is a massive parallel database engine in Pivotal Hadoop
Distribution.
==============================================
A similar example is:
https://issues.apache.org/jira/browse/HAWQ-536
where the original contribution seems to have been
silently ignored and a new issue and code merge on
HAWQ-548 did the exact same thing without recognition
of the previous issue.
Another good example of this would be:
https://issues.apache.org/jira/browse/HAWQ-307
where there was a lot of collaboration on a JIRA driven
by Cos, there was a branch for that work developed in
the tree, but now it seems that branch no longer gets
any attention and it feels like we're restarting this from
scratch.
I can go on, but I think you get my point.
At this point, I'd like to:
1. highlight this in our Incubator report
2. understand how are you guys going to change the situation
Please provide your comments on this thread for both.
Thanks,
Roman.
Re: Concern about under-communication in the HAWQ community
Posted by Lei Chang <le...@apache.org>.
+1 for the first create jira go to dev mailing list and subsequent jira
notifications go to another mailing list is a good idea.
This let developers aware of the recent changes and also people who want to
know more details can subscribe another mailing.
Cheers
Lei
On Fri, Apr 8, 2016 at 12:29 AM, Shivram Mani <sh...@gmail.com>
wrote:
> Yes, that would be ideal. I don't see why subsequent JIRA
> notifications(post creation) need to be sent to the dev mailing list when
> there is specifically the option to 'Watch' jiras.
>
> On Wed, Apr 6, 2016 at 7:17 PM, Roman Shaposhnik <ro...@shaposhnik.org>
> wrote:
>
> > On Fri, Apr 1, 2016 at 7:14 PM, Shivram Mani <sh...@gmail.com>
> > wrote:
> > > I would have preferred having dev discussions in a thread/group
> separate
> > > from the one that has all the jiras.
> >
> > That's a good idea! There's a way for JIRA to send just the initial email
> > (when
> > JIRA gets created) to the dev mailing list and the rest of JIRA
> > notifications
> > can go to another list. Is this something you have in mind?
> >
> > Thanks,
> > Roman.
> >
>
>
>
> --
> shivram mani
>
Re: Concern about under-communication in the HAWQ community
Posted by Shivram Mani <sh...@gmail.com>.
Yes, that would be ideal. I don't see why subsequent JIRA
notifications(post creation) need to be sent to the dev mailing list when
there is specifically the option to 'Watch' jiras.
On Wed, Apr 6, 2016 at 7:17 PM, Roman Shaposhnik <ro...@shaposhnik.org>
wrote:
> On Fri, Apr 1, 2016 at 7:14 PM, Shivram Mani <sh...@gmail.com>
> wrote:
> > I would have preferred having dev discussions in a thread/group separate
> > from the one that has all the jiras.
>
> That's a good idea! There's a way for JIRA to send just the initial email
> (when
> JIRA gets created) to the dev mailing list and the rest of JIRA
> notifications
> can go to another list. Is this something you have in mind?
>
> Thanks,
> Roman.
>
--
shivram mani
Re: Concern about under-communication in the HAWQ community
Posted by Roman Shaposhnik <ro...@shaposhnik.org>.
On Fri, Apr 1, 2016 at 7:14 PM, Shivram Mani <sh...@gmail.com> wrote:
> I would have preferred having dev discussions in a thread/group separate
> from the one that has all the jiras.
That's a good idea! There's a way for JIRA to send just the initial email (when
JIRA gets created) to the dev mailing list and the rest of JIRA notifications
can go to another list. Is this something you have in mind?
Thanks,
Roman.
Re: Concern about under-communication in the HAWQ community
Posted by Shivram Mani <sh...@gmail.com>.
I would have preferred having dev discussions in a thread/group separate
from the one that has all the jiras.
The HAWQ UD(pxf component) team specifically has had more discussions
within jiras and not directly on dev@.
The other cause for fewer discussions on dev@ might be since teams are
still using team specific channels (slack/tracker,etc).
On Fri, Apr 1, 2016 at 6:56 PM, Lei Chang <le...@apache.org> wrote:
> +1 for Roman's suggestion. I think it might help on the communication if
> some best practices are followed.
>
> For example, do a search before creating a new JIRA to avoid duplicate
> JIRAs. Typically, there are 800+ messages on dev@ each month. it is
> challenging for all people to read all the details every day.
>
> For HAWQ-307 (Ubuntu patch), as I observe, contributors are verifying the
> patch, and have some comments on the pull request and waiting feedback from
> the original contributor.before merging the patch.
>
> Regards
> Lei
>
>
> On Sat, Apr 2, 2016 at 6:13 AM, Roman Shaposhnik <rv...@apache.org> wrote:
>
> > Hi!
> >
> > for the past month or so I've been noticing that there's
> > clearly under-communication going on in the HAWQ
> > community. In fact, the more activity I see around code
> > going in with next to 0 activity on dev@
> >
> > A community needs conversation. Discussion. Consideration
> > of topics, roadmaps, and feature development.
> >
> > Here's the most recent example of that:
> > https://issues.apache.org/jira/browse/HAWQ-410
> > where a commit recently went in without the JIRA ever
> > being updated or a solution negotiated/reviewed with
> > the original submitter. In fact, the content of that commit
> > is not problem free. For example, the following really
> > needs to be fixed:
> >
> >
> https://github.com/apache/incubator-hawq/blob/master/depends/libhdfs3/README.md
> > ==============================================
> > Libhdfs3 is developed by Pivotal and used in HAWQ, which
> > is a massive parallel database engine in Pivotal Hadoop
> > Distribution.
> > ==============================================
> >
> > A similar example is:
> > https://issues.apache.org/jira/browse/HAWQ-536
> > where the original contribution seems to have been
> > silently ignored and a new issue and code merge on
> > HAWQ-548 did the exact same thing without recognition
> > of the previous issue.
> >
> > Another good example of this would be:
> > https://issues.apache.org/jira/browse/HAWQ-307
> > where there was a lot of collaboration on a JIRA driven
> > by Cos, there was a branch for that work developed in
> > the tree, but now it seems that branch no longer gets
> > any attention and it feels like we're restarting this from
> > scratch.
> >
> > I can go on, but I think you get my point.
> >
> > At this point, I'd like to:
> > 1. highlight this in our Incubator report
> > 2. understand how are you guys going to change the situation
> > Please provide your comments on this thread for both.
> >
> > Thanks,
> > Roman.
> >
>
--
shivram mani
Re: Concern about under-communication in the HAWQ community
Posted by Lei Chang <le...@apache.org>.
+1 for Roman's suggestion. I think it might help on the communication if
some best practices are followed.
For example, do a search before creating a new JIRA to avoid duplicate
JIRAs. Typically, there are 800+ messages on dev@ each month. it is
challenging for all people to read all the details every day.
For HAWQ-307 (Ubuntu patch), as I observe, contributors are verifying the
patch, and have some comments on the pull request and waiting feedback from
the original contributor.before merging the patch.
Regards
Lei
On Sat, Apr 2, 2016 at 6:13 AM, Roman Shaposhnik <rv...@apache.org> wrote:
> Hi!
>
> for the past month or so I've been noticing that there's
> clearly under-communication going on in the HAWQ
> community. In fact, the more activity I see around code
> going in with next to 0 activity on dev@
>
> A community needs conversation. Discussion. Consideration
> of topics, roadmaps, and feature development.
>
> Here's the most recent example of that:
> https://issues.apache.org/jira/browse/HAWQ-410
> where a commit recently went in without the JIRA ever
> being updated or a solution negotiated/reviewed with
> the original submitter. In fact, the content of that commit
> is not problem free. For example, the following really
> needs to be fixed:
>
> https://github.com/apache/incubator-hawq/blob/master/depends/libhdfs3/README.md
> ==============================================
> Libhdfs3 is developed by Pivotal and used in HAWQ, which
> is a massive parallel database engine in Pivotal Hadoop
> Distribution.
> ==============================================
>
> A similar example is:
> https://issues.apache.org/jira/browse/HAWQ-536
> where the original contribution seems to have been
> silently ignored and a new issue and code merge on
> HAWQ-548 did the exact same thing without recognition
> of the previous issue.
>
> Another good example of this would be:
> https://issues.apache.org/jira/browse/HAWQ-307
> where there was a lot of collaboration on a JIRA driven
> by Cos, there was a branch for that work developed in
> the tree, but now it seems that branch no longer gets
> any attention and it feels like we're restarting this from
> scratch.
>
> I can go on, but I think you get my point.
>
> At this point, I'd like to:
> 1. highlight this in our Incubator report
> 2. understand how are you guys going to change the situation
> Please provide your comments on this thread for both.
>
> Thanks,
> Roman.
>