From: Matt Hargett (

I would like to add, to the Gartner group section:

ZDNet and several other online and printed publications were quoting a
'paper' by the Gartner Group that stated a statistic saying 35 million
lines of code have been added to NT 5.0 since NT 3.51 in reference to the 
operating system's likelihood of stability. We cannot view this
'paper' on the Gartner Group's website, as we have to be paying members;
we will assume that the quote is accurate.

This is what is referred to as "cooking the numbers", I believe.  Let's
define some terms here. A 'distribution' of an Operating System,
such as Redhat Linux, includes the kernel and a collection of common
utilities; an Operating System, to me, refers to necessary kernel and
usermode components required for practical functional use.

Firstly, to compare NT 5.0 distribution to NT 3.51 distribution in terms
of lines of code is one of the most ridiculous things I have ever heard in
my life. IIS, MMC, device drivers, the Win9x-style explorer, Internet
Explorer, and quite a few other components have been added to the NT 5.0
distribution since NT 3.51. Bottom line -- telling us the number of lines
of code added to the distribution tells us more about the grwoth of the
distribution rather than the potential (in)stability of the operating
system. If they wanted
to really inform the industry of NT 5.0's probable instability, they would
give us statistics on lines of code added that run in kernel mode; not a
vague statistic that is indicative of the total size of the distribution
and infers that that has something to do with stability of the operating

Secondly, why compare NT 3.51 to NT 5.0? Wouldn't it be more relevant to
compare NT 4.0 (which was released to the general public 3 years
ago) to NT 5.0? Or would the numbers not be big enough to cause panic?

Shame on the Gartner Group for insulting our intelligence with such

Matt Hargett

main page ATTRITION feedback