On 06/22/2014 11:53 AM, Johnny Hughes wrote:
All we are trying to do here is be completely honest and complete open about the distro as a whole and the process to get it as a whole.
I for one applaud the very positive changes toward more openness.
Now, as far as I am concerned, the release number is relatively unimportant. I've never tried to align any particular RHEL release to a CentOS release, especially after I've seen a bit of the other side of dealing with this through doing my own rebuild for IA64 of C5. But if I were to have a preference, it would be to specifically make the release number different, to highlight the, well, difference. CentOS != RHEL at some arbitrarily chosen abstraction level (no matter where you draw the line at '100% compatible' there's always a gadlfy or three out there that will say 'but you didn't build it this way, or that way, or it differs by three bytes, or the buildtime is different or .....') Yes, this is a change in perception, but not a change in reality. It is more in line with SL with their 'rolling' repository, which concept I find very attractive.
(For those in the peanut gallery, I am not on the CentOS board, do not have a centos.org or redhat.com e-mail, and never have had.) I've not liked the whole idea of 'point releases' for some time now, since people will artificially stay there (and I know the reasons why: partly due to historical breakages when going to a new point release, and partly due to vendors who are a bit antiquated in their support policies; I've even posted on how those are reasons people use to stay at a point release before (grep the archives, it's there)).
I'd prefer it to not be a 'point release' for another reason. In some folks' minds, 6.2>5.10 when that is definitely not the case.
However, I'm absolutely positively sure I don't like the Ubuntu-ish date format, but that's primarily because it would be Ubuntu-ish. I'd prefer just a basic build number: CentOS 7 build 756, for instance (no decimal points!). I'm sure that there would be at least one person chime in and try to map a build number to an RHEL point release, but that would be their own business. Of course, a date format makes it much clearer how that 6.2!>5.10, and that would be a good thing, and a straight build number won't help that particular case.
And I know lots of people will disagree with me; that's fine, disagree all you want. Replies in disagreement to /dev/null, please, as more than likely your point has already been stated elsewhere in the thread. :-)