On 07/09/2011 01:32 PM, Karanbir Singh wrote:
> yes, we all clearly take that on board - I hope the changes we are
> bringing in helps clear that, and prevent this sort of a situation. But
> there are still lots of places for improvements, and over the next few
> months lets try and address all of those.
>
> - KB
Sorry for thread-jacking, but I wanted to start this thread in relation
to your comment.
As I understand it, a lot of the delay came from reproducing Red Hat's
build environment. That being needed for the binary compatibility. With
each new major release, the number of packages, and in turn, the amount
of complexity grows.
Is that a correct understanding? If so, then EL7 will be even harder to
sort out and will lead to an even longer delay in release.
I think there is a business case to be made for CentOS, from the point
of view of Red Hat. My experience has been that a lot of
people/companies start out on CentOS. After a while, those that succeed
and do well eventually want to switch to Red Hat proper. As good as
CentOS is, by it's very nature, it will always lag behind RHEL in so far
as updates are concerned.
Given all this; I think there is an argument for Red Hat wanting to
assist CentOS. As we saw with this release, the delay drove people away
from EL. I am sure many went to Debian or other non-EL distributions.
Each of these defections is another potential future customer lost to
Red Hat.
If Red Hat could be convinced to help the CentOS team with things like
setting up their build environment, they would help foster this
potential customer base while investing minimal time and effort. Has
anyone in the CentOS team approached Red Hat to discuss some sort of
arrangement like this?
As an anecdotal example; We've built our entire infrastructure on
CentOS. Now, our clients who are doing well, we are moving to Red Hat
proper while still using a lot of CentOS internally and for smaller
clients. It's a very smooth fit and transition, thanks to CentOS's
binary compatibility.
Just an idea. Thanks for the hard work and I'm anxious to play with
CentOS 6!
If Red Hat really wanted or cared about the customers you list here, it could simply make RHEL a free download with security updates. That would require very little spending on their side compared to duplicating their build infrastructure at CentOS and supporting both environments (eg. transfering their knowledge, what makes their product tick, to a open source project where it could be copied by companies seeking to profit from it).
One could make a point that doing that would be a burden for Red Hat in terms of additional head count required to support the non-paying customers and the infrastructure costs, something they would have a hard time promoting internally to shareholders. Let's imagine that all CentOS contributors could be motivated to help RH in such imaginary efforts... RH would be giving direct control of the quality of its product to outsiders. Something already accomplished with Fedora.
Your idea is nice and it's looking at the right perspective, IMHO. However, I don't feel it'll have much traction within Red Hat.
Right now I think it'd be more practical to request any help that is needed (besides servers and hosting) and organize this work to reap the benefits of a larger contributor base. But I'm just a CentOS user that hasn't contributed anything besides promoting it and helping other users, so my opinion should be taken with a grain of salt.