[Cumbria] For discussion - Is Linux really ready for the Enterprise environment?

[Ian Linwood]

Hello cumbria,

The deamons advocate here!

Firstly, My Experience ---------------------- I've been using Linux
since the early 90's. My first experiences with Linux was with the
only distro available at the time - Slackware. I have a virtual book
of correspondence between myself and Patrick Volkerding, who patiently
led me by the nose, through my first experiences.

I've used several distro's. Slackware, RH, Suse, Turbolinux, Mandrake
(thanks Dave), and a couple of more obscure offerings. I've used these
in a domestic and some in a corporate environment.

Distro's don't really matter, really. They are all based around the
Linux kernel.  Unfortunately some vendors 'hack' the standard kernel
to suit their own agenda. This is OK, as this is what the GPL allows.
But in some circumstances this has led to a very diverse fork of the
'Linus' branch. RH is the main culprit. In my opinion, this is an
attempt to hijack Linux (is Alan Cox to blame?). RH has recently
engaged in patenting their software (loud booing and meaningful
hissing encouraged).

Suse at least try to get their changes incorporated into the Std.
kernel and contribute to other OS projects. Yast is an good thing, as
it makes life very easy for newbies (and people like myself!). But
Suse has it's own problems! Again, custom kernels (to a lesser extent)
and GUI schizophrenia. To be honest, I am not a GUI fan. I broke my
teeth on Tandy's and Sinclair ZX81/Spectrum/QL's (god I wish I'd kept
that quiet!). Then glorious IBM's with CGA monitors, 8088's, and
Winchester drives (at the time, seriously - WOW!) CP/M and early
MSDOS, then serious 32bit systems (motorola) with Xenix, HPUX.
Sun/Sparc has been my light over the last few years. Current releases
(7, 8&9) have delighted.

I contributed code, when I was fluent in C, (indirectly, none credited
- for 3Com 905 driver to fix transceiver selection bug) to older linux
kernels, after which the driver was completely re-written).   :-)

So, as you can see, I'm not a newbie. To summarize, I've had
exposure/experience with most commercial OS's out there, I've done
some serious development work too (not mentioned above).

I've said all this in an attempt to prevent being flamed by all and
sundry (and to avoid the crap of - "my distro is better than yours").
As you can see I work with everything (dos, win - all evos, VMS, Unix
variants and clones)

I am currently in BSD mode (Free/Net/Open). This is a real haven for
me, as these geeks really know their stuff.  Unfortunately this is a
really hostile place for people who do not know how to RTFM.

Back to Linux; NG's seem to be flooded with users stuck in RPM hell.
There is no standard Desktop to implement across corporate enterprises
(or SME's). Distro's have different ways of installing packages.
Enterprise hardware is not well supported, SMP support is truly CRAP.

Is UnitedLinux the way ahead? Or will this be another 'proprietary'
implementation which we will have to pay a royalty (in more ways than
one) for?

Should Linux be wrestled away from Linus and put in the hands of, say,
IBM? For Linux to grow any further, it needs a clear path to stride.
Does anyone know where this leads, can anyone see through the fog
ahead?

Please don't get me wrong, I like Linux, but after a decade of
development, it seems to me, that the trust still doesn't seem to be
there.

Sun are contemplating Open Sourcing Solaris. They are working on doing
this with Java (oh bliss!). Where will this leave everyone else?

Linux, BSD, SysV, Win32, VMS, AS400 and countless more derivatives
exist. Does any one really need to be dominant? Co-operation and
communication are what is important. Common methods and standards are
what is needed. Open architectures, to allow all to exist and work
together. It doesn't matter what car you drive, as long as you can
turn right safely.

Crickey, I'm sounding like RMS ... "excuse me while I go off and shoot
myself" (c) Douglas Adams