Axel Thimm wrote:
> Hi,
> 
> On Thu, Mar 30, 2006 at 11:06:51AM -0600, Troy Dawson wrote:
> 
>>Hello,
>>I would like to start a conversation, that will hopefully grow into some 
>>type of policy.
>>
>>What makes a release Scientific Linux compatible?
>>
>>What I'd like to get out of this discussion is a way so that a user, or 
>>site developer can say "This package, or groups of packages, have 
>>changed, therefor this release is not compatible."  Or "I can make 
>>changes up to this point and still be compatible."
>>
>>I don't expect this discussion to be finished quickly, possibly months, 
>>or maybe even the fall Hepix, but it needs to get started at some point. 
>>  I will be bringing it up at this spring Hepix.
>>
>>And this isn't to single out any one site release.  There are plenty of 
>>sites or branches.  Yes, it will be able to help Cern, and Fermi, say 
>>"yes, we are compatible" but it could also help those people who have 
>>been mixing and matching atrpm's, dag, CentOS and others.
> 
> 
> I think dag, dries, kde-redhat and atrpms are not in the same league
> as SL, CentOS, etc. The former are simple add-ons to RHEL and RHEL
> compatibles and the latter are true distributions.
> 
> If a site decides to break compatibility with the base SL that would
> be bad. After all the strength of all RHEL compatible distributions is
> the fact that something built for RHEL will work as well on SL
> including not only the mentioned oss projects, but also ISV/IHV with
> their proprietary bits (I'm now thinking of HP and proliants,
> e.g. simple sensor readouts for temperatures sometimes require special
> closed source modules).
> 
> Of course you sometimes don't have a choice if the RHEL bits don't
> serve your needs (4kstacks, no xfs etc).

Well, I wasn't thinking that atrpm's and dag is going to break 
compatibility, but for some admin's that's just the first stop.  They 
then go to fedora-extra's, and then down to fedora core.  That is why 
I'm asking, just where should we draw the line.  Or is there a firm 
line, and maybe just have guidelines.

Troy
-- 
__________________________________________________
Troy Dawson  [log in to unmask]  (630)840-6468
Fermilab  ComputingDivision/CSS  CSI Group
__________________________________________________