[CentOS-devel] progress?

Mon Feb 21 15:19:01 UTC 2011
Karanbir Singh <mail-lists at karan.org>

On 02/20/2011 12:11 PM, Dag Wieers wrote:
> 4) CentOS is not able to release CentOS 5.6 after 2 months and nobody is
> allowed to be critical about it.
> (Despite the fact that the effort to rebuild CentOS 5.6 packages is a
> lot easier than CentOS 6.0 which is already 3 months late)

has it been 2 months since 5.6 was released ? Add another point of real 
interest: when was it that red hat released all sources for 5.6 ?

> 5) The same 3 people are responsible for CentOS 4, CentOS 5 and CentOS 6.
> What's more, the fact that there would be three update releases in 3
> months was predictable.

thats true. although upto 7 people can be involved at any given stage, 
not 3.

> So despite all the automation, QA team, past promises and whatnot, we
> are not doing a better job today and I had hoped at least some people
> would agree instead of denying there's something wrong with the process
> and blaming the non-volunteers/community for even bringing it up.

The fact that there are disfunctional setups in place is not something 
that anyone ( I for one ) are deny'ing. But the fact that a call for 
help got zero traction for weeks is also worth considering. We could go 
back, stop everything that is going on at the moment and try to process 
engineer a better setup before we again start working on CentOS-6, I'd 
say a target of 2 to 3 months would be reasonable if we did that.

On the other hand, we can just get this done out of the door and then 
look at process engineering for the future. We are better, stronger as a 
group with a much larger contributor base than ever before - I see no 
reason why we could not strengthen that even further and split the roles 

Besides, I've always argued that the idea of a QA team was wrong, the 
way it was setup was wrong and there is plenty of scope for us to change 
that. As a striking example: 70% of the QA team people never actually 
did any QA work at all. Losing them was step-1, moving as much as we can 
into an automated public stack is step-2, some work has happened on that 
already, but I doubt we can move to it completely in time for 6.

There are a *lot* of things like that going on around the edges, and are 
not visible or much touted at the moment. Because (a) we need to retain 
core focus and (b) the changes need to work through a process that has 
near zero user side impact. Whatever process we have now and have had in 
the past, clearly worked for a large number of people - and the fact 
that they trust the process. I guess another issue is that a vast 
majority of the centos userbase isn't represented here in the lists, the 
forums or irc. Which makes it harder for people who watch these media to 
sometimes comprehend the meandering.

> And despite what some people may think, I am not _against_ CentOS, in
> fact the only reason why I am bringing it up is because * I * still *
> care !

I can believe that :) I just dont believe you understand the entire 
centos ecosystem. Maybe we need another beer session sometime soon. 
Shame that you were unable to make Fosdem.

- KB