[Sussex] Re: Gentoo problems.

[Steve Dobson]

Mark

On Mon, Feb 23, 2004 at 07:27:52PM -0000, Mark Harrison wrote:
> > > Of course, there is always the flip side in that companies are afraid of
> > > change.
> >
> > A much more likely scenario.  A recent study showed that for every
> > windows XP installation in British businesses there are 4 windows 2000
> > installs and 150 NT4 installations.
> 
> An even more likely scenario which explains both the low takeup of XP and
> Linux in the "FTSE 100" space:
> 
> IT managers in large companies who value their jobs don't do anything as
> damnably stupid as choosing an Operating System based on the cost of the OS
> in isolation.

Those numbers would, I assume include old legacy systems.  How many of them
are in companies and how many in homes?  Legacy systems are not normally 
changed until they become obsolete and the system is turned off for the 
last time.
 
> For what it's worth, I define a "large company" as having more than a few
> thousand staff, typically in more than a few hundred locations - this pretty
> much holds for 95% of the companies I've ever worked for, either as an
> employee or a consultant.

I have the same view.  But then I worked for GEC back in the late '80s when
it was a very, very big company.

> The normal way to write an IT strategy is to start with "what do I need to
> do", work from there to "how do I do it" (business processes), work from
> there to "how do I implement it" (applications) and finally figure out what
> operating system is required.

There is also the factor of risk.  It is more risky to introduce something
new, and there are ways of giving that a cost value.  Hell I'm guilty of 
that myself.

> The general cost model for any IT budget is that software licencing
> typically makes up less than 20% of the budget for any IT project. The
> largest costs are user-retraining, and support.

I would agree that that is true today.  But there is one sector where even
a 5% licencing cost is too high, Education.  With Unis switching over to
Linux, *BSD etc to save costs.  How long is it going to before there is a
very significant cost in using Linux?

<snip>
> One of the other main barriers to OpenSource adoption in the corporate marke
> place is the lack of widespread support infrastructures. While a number of
> us on the list have helped various businesses move to OpenSource, I'm not
> aware of many companies which could realistically provide a UK-wide, or
> European-wide support solution for an organisation with many hundreds of
> branches... but there are many companies who will sell exactly this service
> for offerings based around commercial software packages ranging as widely as
> from MS Office to SAP.

I disagree.  IBM in putting a large amount of cash into Linux, as are HP.
Sun and Novel have just started.  Red Hat are exactly in that market space
and have drop there other lines to concentrate on its Enterprise Solutions.
Morse would be another such a company, which I know employees Linux experts.

The fact that these departments are not as big as the M$, Solaris, HP-UX
etc sectors reflects the share that Linux has not lack of support.

> This is changing, and I think that IBM's new view of linux will change the
> penetration in large organisations dramatically.

While I'd give credit to IBM for being early into the Linux camp I'm sure
that Linux was just an experiment that has paid off.  Think of all the 
projects that IBM have started and failed with - OS2 for example.

> However, many small IT consultancies dramatically underestimate the amount
> of time and complexity involved in a major migration project... or fail to
> understand the complexity of priority and decision-maker networks in
> companies with more than a few thousand staff.

Here I completely agree.  One of the reasons I decided not to go into the
Linux support area is that I couldn't help companies migrate from Windows
as my N$ expertise just isn't good enough.

Steve D