Re: [tlug] Eating bugs...

[Jonathan Q <jq@example.com>]

On Fri, Apr 18, 2003 at 12:26:50AM +0900, Stuart Bouyer wrote:

>If you read the original MS sponsored article you notice that personnel
>costs were where most of the costs came from. The article focused on a
>company already using MS products, the costs came from "training" their
>MS orientated staff in using Linux and related software. The of course
>used the most costly training available.

It's been a while since I've seen that rag.  Yes, that's quite a 
slant they put on it.  If IDC were being honest instead of
paid by MS to say what MS wanted them to say, the study would have
found something like, "It is cheaper, over the short term, to
continue running Windows rather than to convert to Linux and incur
staff-retraining expenses and the cost of duplicating some hardware.
However, the five-year prognosis is that you will save quite a bit
on your software costs by moving to Linux."

>If I remember correctly there was something about RedHat support - like
>when you buy RedHat you get 90 days support, so they kept adding the
>cost of buying the disto so that they were continually covered.

Gee, and I thought I was being charitable to the MS case in my
analysis :-p  Few companies actually buy a copy of the distro for
every server on which they install it, or even on any of those servers.
IIRC that support is pretty much just install support.

And of course, in the real world, a company moving to Linux would be
best off hiring a few experienced Linux admins and having them
impart their knowledge to the Windows admins.  In a large organization,
that approach plus the purchase of some books would be more cost
effective than sending all the Windows admins to expensive training.
Oh, and there's that little matter of all the money they'll save
on not sending people to expensive Windows training any more.  This
is starting to look more and more like a moneysaver, or at least a 
wash :-)

>Also included were costs of training all the users to use linux (who of
>course were happily using MS at home so of course they had no problem
>using it at work - amazingly there were no Mac users in the office!)

A lot of offices do have no Mac users, or very few (chucking the
Windows boxes and buying Macs would be a good option for many
companies, too), but since the point was supposed to be about
which is cheaper to run as a certain kind of server, the study
is raising a big, fat, red herring here.


>All in all an amazing article that shows the FUD that can be generated
>by statistics and number crunching if you generate the right
>"conditions". As my math teacher used to say - "You can prove anything
>with statistics - as long as choose the correct numbers and the correct
>questions".


Yes, lies, damned lies, and statistics rear their ugly heads again.
My statistics teacher (and that was by far the most interesting and
useful math class I ever had) said the same thing about proving
anything with statistics.

Jonathan