On 07/05/08 14:01, Andy Buckley wrote:
> Jan Iven wrote:
>> On 06/05/08 18:38, Andy Buckley wrote:
[..]
>> From my perspective, what matters with SL(C) is the validation of
>> experiment software, and the (current) assumption at CERN that desktop
>> == public interactive cluster == batch service == grid (i.e. HEP-wide
>> compatibility).
> 
> In my opinion, the homogenisation of HEP batch services has been a
> double-edged sword. On one side, it has definitely made the above
> assumption valid, which is nice for non-technical users (except that
> their desktop functionality is then limited to that of a 3 year old
> batch/Grid node... hence the Ubuntu/Fedora creep among more techie
> physicists).

(lets not forget the RedHat-6/7 "seep" with the more physicist 
physicists - "128M are enough for anybody, whats wrong with FORTRAN66, 
and my old X-Terminal never had sound issues...")

> But simultaneously, the single supported target platform has led to
> non-portable experiment software and LCG middleware. My experience is
> that *having* to ensure portability results in better code, and easier
> portability between major releases of SL, but the enforced SL domination
> means that this often isn't done.

Don't get me wrong, I am all in favour of portable software (and would 
encourage developers / release managers to build&test on a zoo of 
platforms [more so since I don't develop a lot myself :-) ]).
My point was purely from a support perspective - I definitively don't 
want to troubleshoot marginal one-off issues due to somebody's failed 
download of the über-geek PPC slackware update from some shady 3rd-party 
repository..
And right now it doesn't look like we have a clear winner for a 
"supportable" laptop/desktop Linux in our community..

Regards
jan