[Custom] Enterprise version.
I haven't seen anything much on the list concerning CDDs lately, but on
a visit to the site, it is noted that discussion is still centred here,
so, a few thoughts:-
With an enterprise version, I don't think that we need to worry about
straight out of the box installation on a newbie level. The new Debian
installer will have that covered for the workstation/desktop level.
In the server environment, a sys. admin. would be handling the
situation, and I feel that there is even less to worry about.
Installation in Debian has been greatly simplified in recent times, and
I don't see it as being the hurdle it has traditionally been assumed to
So, skipping over those, the 2.6 kernel is in Sid, I would recommend
starting with that, with a particular accent on security, considering
the intended environment. It already has the very marketable performance
For the server environment, the AMD 64Bit project obviously must be
included, dependent on the individual environmental application.
Boot, -Grub handing off to Lilo.
File systems, - Ext2 is history, Ext3 has possible application, but I
would consider just Reiserfs4 and XFS to cater for small and large file
Rather than manufacturing the hardware detection from a conglomeration
of packages (discover, kudzu, detect, etc.) as has been suggested, I
understand there is to be collaboration with Knoppix/Debian on this? And
we may be better off allowing that to happen first, especially with a
view to clusterknoppix.
As far as the office suite scenario goes, the drive would appear to be
Even allowing for the code headstart from star office, Open Office has
come a long way in three years, and has the busiest development team I
know of. Sun has just started to recognise their work with an offer to
finance, and they are not based on non-free Qt. Whereas KDE have a
bubblegum GUI advantage, and that is all I can see, with the Open Office
attitude, I don't see that as being a major difference for long. Gnome
at this point, unfortunately, does not have what the target market would
consider a suite. Individual packages like Gnumeric are up there, but
not a coherent interoperable whole.
I would also look at things like the deployment of the new XFce4, and
that type of menu structure, rather than assuming the performance
overhead of KDE or Gnome.
In the open source/free software environment, marketing aspects are
rarely considered, and traditionally not looked on favourably. But if
that viewpoint drives M$ into the dirt before their touted
Longhorn/Linux offering, I'm all for it. Marketing concepts maximise
profit. There are many definitions of profit.
Favourable marketing concepts, within the enterprise environment are,
stability, which we have, security, which we now have again, speed must
also be achieved in order to further cast those that would appear to be
competition, in the eye of the target market, into the shade by
comparison. In order to do this, each package, and its' dependencies,
must be chosen with care.