[casual_games] languages... (that's an 's' at the end!)

[Joe Pantuso]

I have had experience in the past with requiring specific MS components
(very large ones) to be installed.  Specifically service packs and IE5.

IE5 is a good analogy for the .NET situation.  In our case (McAfee Personal
Firewall, I used to lead that dev group) we needed specific OS components
upgraded.  IE5 was a big deal from a development standpoint, not the browser
itself, but all the plumbing that was changed in Windows to support it.  It
was a 'stealth service pack'.

Somewhat less than 1% of customers had a problem with this.  Having a
problem with it was always a matter of personal preference and not a
technical issue.  The thing is though, that we didn't need the browser
itself, we needed the fixes to the OS that installing IE5 achieved,
including a couple nasty exploits.

Size is not a consideration when you are talking about a component that gets
installed *once*.  Games written in .NET are no larger, and in most cases
could be significantly smaller.  There is room for disagreement with that
position.  My experience suggests it is not a factor.

Further the reach of .NET is considerable.  I was being facetious before
when I threw out the 50% number.  .NET capable systems with a moderns OS
underneath (meaning XP and 2K) represent 80-85% of all desktop PCs world
wide.  Remember a Win98 box was purchased 8 years ago!

Now ignore .NET for a moment.  If we look at doing real QA on an
application, and we are using something that can be run on that other 15-20%
of desktops in the world, our testing matrix explodes.  Instead of having 1
or 2 configurations to test it goes up to 4 or more.  My costs of
development go up, and/or my support costs go up.  The economic proposition
does not make sense in many cases.  (Ignoring that many people do really
well selling their games on non-Windows machines due to the paucity of games
there).