[CentOS] CentOS Project Infrastructure

Fri Aug 7 13:17:14 UTC 2009
Ned Slider <ned at unixmail.co.uk>

Marcus Moeller wrote:
> Dear Russ,
> 
>>>> Don't misunderstand.  I think you have done and are doing a great job
>>>> but some things are out of any single person's control.  All I'm
>>>> suggesting is that it would be nice if there were an easy answer to the
>>>> question of "what if" those things happen to a few of you.  I think it
>>>> is a good thing that the question is being asked, though.
>>> As an outsider (as far as CentOS development goes), I think this would
>>> probably be a good time to just back off a bit, chill out, and see
>>> what comes out of the current reorganization.
>> * chuckle * Actually I was appreciated Les' comments, in the
>> first instance today and later.  If I cannot respond to
>> thoughtful comments, I've probably not thought the matter
>> through enough.  I may choose to ignore matter of course where
>> comment is not yet ripe
>>
>> Akemi, Ned and Marcus [and others who have contacted me and
>> some of the others on the core group off-list] are obviously
>> concerned, want to help, and want to participate more as well,
>> and I'll probably do yet another run at describing some ways
>> to increasingly grow as a sysadmin, a developer, and as a
>> 'person worth watching' as posts of each and others in recent
>> days have set me to thinking.
>>
>> I've done such coaching on the ML, in the wiki, and in private
>> email, so why not yet again?
> 
> That"s a great offer and what I titled as mentorship. 

I think the issue here, at least as perceived by those outside of the 
project core, is that little is done to actively encourage contributors 
(ie, mentorship). It's all very well noting and observing the talent 
develop and calling upon said talent down the line so long as said 
talent hasn't lost interest in your project in the meantime. What 
concerns me is that I see absolutely no effort on behalf of the project 
to nurture/develop/mentor the next generation of CentOS developers. Who 
will step up to the plate and commit to being lead dev on EL6 with a 7 
year lifecycle, a full update set every 6 months, security updates to 
rebuild at no notice. It's a huge undertaking.

 From my own experiences when trying to contribute, I have repeatedly 
been told not to bother, not to do it and to go away. So in the end 
that's what I did out of frustration - I went away and founded the 
elrepo project with a few others who also wanted to contribute but found 
themselves unable to do so. Initially I viewed this as a failure - I 
would much rather have seen the elrepo driver project be done as the 
CentOS Dasha project (and likewise, for fasttrack). But now I see it as 
an advantage not being part of a CentOS project - by not being part of 
CentOS we are able to support and work with the whole Enterprise Linux 
community (incl. RHEL and SL), not just CentOS. Red Hat have recognised 
our value and we are already engaged with Red Hat developers in 
discussions regarding the direction of the driver update programme in 
RHEL6. It would be nice if the CentOS Project wanted to engage too :-)

IMHO I think it's a shame CentOS doesn't presently offer rebuilds of the 
FasTrack channel. I know there is a need within the community (our own 
logs from our fasttrack offering show us that). Let me say this isn't 
particularly about fasttrack or about me, it's about highlighting how 
the process doesn't work - I merely use my own experience as an example 
to highlight this. I have expressed a willingness to contribute. I have 
shown a commitment over a reasonable length of time, so I'm not the here 
today, gone tomorrow type. I have been rejected, gone off and done it 
anyway, so I have demonstrated resilience and determination - I've 
demonstrated I'm a "do'er" not a "talker". My "product" is out there for 
others to view and judge my level of competence (I don't and never have 
claimed to know everything or be perfect, I only display a willingness 
to continue to learn and develop). I merely seek to contribute back to a 
community from which I have taken something of value. Yet at every step 
of the way I have been rejected and knocked back. Never once has a 
CentOS dev approached me with an offer of mentorship or advice or 
anything else. As I said, this is absolutely not about me - my 
circumstances are not unique. For every person like me who is knocked 
back or rejected, there must be dozens more onlookers who see that and 
don't even bother trying to engage with the project.

Another example is the forums. I started engaging with the CentOS 
project back in 2005 in the CentOS forums. For years I worked diligently 
  helping users there and was "rewarded" for my efforts in 2008 being 
made a forum moderator/administrator. My fellow forum moderators both 
have @centos.org email addresses, something I was denied? How is one 
supposed to represent the project when one isn't given the tools to do 
so? It's only an email alias - why would some be afforded that and 
others be denied? You may think this is a moot point and I'm complaining 
for the sake of it, but it's about how you make people feel - do they 
feel valued and worthwhile or are they made to feel like their efforts 
don't really matter. It was mentioned earlier in this thread that only a 
handful of people actively participate on bugs.centos.org. Do you ever 
wonder why no one contributes on bugs.centos.org? Perhaps if you 
understand my point in this paragraph then you will start to understand 
why you have no contributors on bugs.centos.org? Generally when people 
are made to feel like they are worthless inconveniences they will go 
elsewhere. You can't just treat people like shit, say CentOS is a 
meritocracy and you "cannot help their wounded feelings" and then wonder 
why they don't want to help you :-)

Which brings me to the point that great software developers do not 
always make great managers of people or people's expectations - the two 
skill sets are not mutually inclusive. I have said it before, and Marcus 
has said it herein, maybe the CentOS Project could use a Community 
Manager role - someone to be the public face of the project and 
interface between the community and the core development team. Someone 
like Akemi has time and time again demonstrated the required skills and 
would be my nomination should such a role ever exist. This would also 
free up valuable developer time to concentrate on development issues 
rather than responding to media nonsense all over the place - let the 
project speak with one voice and let that voice be one who understands 
how to deal with the public. Just an idea.

We had large parts of this discussion nearly a year ago in private and 
were told things would improve. Since then nothing has changed and we're 
having the same discussion again, this time in public. And again we are 
told things will improve. Last time a lot of the concern stemmed around 
the lag of the 5.3 release, being 10 weeks. We were told that processes 
were being put in place to improve that situation. Now we're having this 
discussion again in the midst of an ongoing 12 week lag (and counting) 
for the release of 4.8. It seems to generally be moving in the wrong 
direction.

Russ is right, people are engaged in this discussion because they care 
passionately about the project and want it to succeed. IMHO if CentOS 
continues to ship a product based on a ~6 month update cycle where 3 
months of the year security updates are late (more than the stated 72h) 
and the other 3 months there are no security updates at all due to the 
update lag period (now running at 12 weeks) required to produce the 
product, then at least to me that is a major cause for concern.

There - I feel so much better getting that lot off my chest :)