On Thu, 12 Dec 2013, Otis Gospodnetic wrote:
Hi,
24h is calling...
On Thu, Dec 12, 2013 at 1:10 PM, David Lang <[email protected]> wrote:
On Thu, 12 Dec 2013, Otis Gospodnetic wrote:
Hi,
On Thu, Dec 12, 2013 at 12:13 PM, David Lang <[email protected]> wrote:
I think it's going to depend on what you mean by "equal say" in rsyslog
development.
But I think we know this, no? It means agreeing on how something gets
implemented, designed, it means trusting others to commit, it means
sharing
responsibilities, etc. etc. You must have been on at least *some* Apache
mailing lists over the years where this sort of work happens! :)
dmany of the most successful opensource projects don't allow multiple
people to commit.
Yes, agreed, it's not exclusive and absolute. Btw. what examples do you
have in mind?
linux kernel, busybox, I'd have to go looking to find more.
rsyslog is already very open in terms of discussing how things are
implemented and designed. in the end it boils down to who is doing the
work. Everyone else can kibitz and try to convice the oens writign the code
that they are wrong, but in the end it's the ones who are writing the code
(or paying them to write the code) who make the final decision on how it's
implemented.
Agreed. But by allowing more people to co-own, you invite more people to
contribute.
granting people ownship without them earning it is a recipe for disaster. How
they earn it is a good question.
do you thing that any random person should be able to modify rsyslog without
their code being reviewed? I don't think that you do. So the question is who is
qualified to review the code and do the contribution without further discussion
or review.
Right now, there are _very_ few people whith that sort of knowledge of rsyslog
and the problem space. I don't think that any of those people are complainign
that they don't have the ability to commit their code.
Sharing responsibilities is a social issue, not a policy/ownership issue,
who is trusted. This isn't something that you want to cast in stone. But
people who are going to share in the responsibilities don't show up fully
formed and ready to step in, they hae to develop over time, becoming
familiar with the project, it's culture, and the other people.
Right. In the end can they earn the right to commit in case of Rsyslog?
Or do their contributions always have to go through somebody else's 24h
bottleneck?
right now, like a lot of projects, they always need to go through a maintainer.
But if there were enough contributers for this to be a problem, the contributers
would also be the solution.
the rsyslog codebase just isn't that large, if you were to have a bunch of
people commiting changes directly, they would quickly start stepping on each
other's toes and causing problems. So there would need to be some way to
coordinate changes that effect the work that others are doing. When you start
needing to do this coordination, whoever does the coordination becomes the
bottleneck, just like a committer, so it actually works really well for the
coordinater to be the committer and not have the people they are coordinating be
able to commit directly. Git (and github) support this model _really_ well,
which isn't surprising because Linus wrote git to direcly support this sort of
workflow.
As I see it, Adiscon has say in rsyslog only in that they are paying
people to work on it. If there were others contributing code, and effort,
those others would have say based on the effort they are contributing.
At the end of the day it's still Adiscon that lets something in or
rejects.
Or doesn't have the man power to review contributions. Or has to not
implement some features because they are not sponsored. That would go
away
if thigns were more open.
I disagree with this. Changing to Apache would not provide more manpower
to review contributions.
It actually would.
Good contributors are invited to become committers.
More committers means more contrib review manpower.
I've seen this over and over over the years.
and I've seen massive projects like linux-kernel work without granting lots of
contributers.
I've also seem projects like the *BSDs have major problems centered around who
is allowed to be a committer and blocking commiter access.
You are also assuming that Rainer is being told what to accept and not
accept by his management at Adiscon. That does not match reality in my
experience.
Oh, no no, wasn't trying to say that.
But I do *think* that it is *Rainer* who ultimately decides what goes in
and what gets dropped. If that is so, that sounds like a bottleneck to me.
Linus ultimately decides what goes into the Linux kernel, he is a potential
bottleneck, but if rsyslog is limited to no faster development than the linux
kernel, I think there's enough headroom that we don't need to worry :-)
David Lang
_______________________________________________
rsyslog mailing list
http://lists.adiscon.net/mailman/listinfo/rsyslog
http://www.rsyslog.com/professional-services/
What's up with rsyslog? Follow https://twitter.com/rgerhards
NOTE WELL: This is a PUBLIC mailing list, posts are ARCHIVED by a myriad of
sites beyond our control. PLEASE UNSUBSCRIBE and DO NOT POST if you DON'T LIKE
THAT.
