On Wed, 2 Jun 2004, csieh wrote:

> John,
>
> On Wed, 2 Jun 2004, John Franks wrote:
>
>> On Wed, 2004-06-02 at 11:09, Steve Traylen wrote:
>>> On Wed, 2 Jun 2004, Corrie Kost wrote:
>>>
>>>> I would tend to favour following the Redhat choices for what is loaded
>>>> by the core distributions. To do otherwise would eventually cause problems.
>>
>>> Just to second this the absolute ideal as I see it for this is that the core
>>> version be as much as possible identical to redhat. Including for instance
>>> bug for bug matches even if they may be trivial to fix.
>>>
>>
>> I did not expect this to be contentious and I don't feel that strongly
>> if people want a "bug for bug match".
>
> I sure do not want all the RedHat bugs.

I also would rather the redhat bugs did not exist but I believe an aim for SL
is to appeal to as wide an audience as possible? To do that doing as
absolutly as little as possible to the source code, ie nothing, is the way
to go.

If it is modified in any way then I belive people and sites will be less
willing to use things. They would rather have the same problems as redhat
than any potential new ones.

Moving forward with extensions rather than replacements that contain
bugfixes and the excelent change log to RHEL that already exists is clearly
an excelent thing towards a middle ground beyond this extreme.

Anyway I'm feeling guilty for turning up on a list of a fantastic effort
and critisising, Rutherford and I want SL to succeed as it is definetly the
best solution that exists to problems we now face within HEP and wider
community, it in fact represents something better than what we had previously
speaking as a non-accelerator lab second to CERN, Fermi and SLAC to which
we have always tried to follow in the past.

What can Rutherford do now of assistance to the process as I am very happy
to help.

    Steve

>
>>
>> Just for the record though, I believe in this instance it is a bug:
>> See http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=113485
>>
>> This bug is fixed in Fedora Core 2.  Perhaps it will get fixed in RHEL
>> also.  Maybe it has, I have no way of checking.
>>
>> Since the packages appropriate for a default scientific workstation will
>> be substantially different from those for a default corporate desktop it
>> might be helpful if there were a simple mechanism to produce a good base
>> scientific workstation, e.g. a list of packages to add/remove.
>>
>
> Clearly things can be done here that will NOT break the "core" packages of
> binary compatability.  In this case the packages already exist.  The only
> question is if they are installed by default.  Note that there are 4
> "group" choices.  There are custom, personal desktop, workstation and
> server.  It would be appropriate to add these tetex changes to
> workstation.
>
>> --
>> John Franks <[log in to unmask]>
>> Dept of Mathematics, Northwestern Univ
>>
>
> -Connie Sieh
>

--
Steve Traylen
[log in to unmask]
http://www.gridpp.ac.uk/