Re: [GD-Linux] ANN: Candy Cruncher for Linux shipped

[Brian H. <bri...@py...>]

Jeez, I didn't expect this to turn into such an argument.  But I'll toss in 
my $0.02.

The Linux distribution situation is a complete disaster right now.  If Ryan 
hadn't handled the Linux port for us, it wouldn't have gotten done period, 
irrespective of the actual porting process.  Just seeing the system 
requirements alone freaked me out (kernel version, XFree86 version, sound 
access, glibc version, etc.) -- I thought it was bad enough under Windows.

The multiple distros + emu layers that exist are problematic for many 
reasons -- kernel versions, XFree86 versions, default locations of "stuff", 
how sound is handled, etc.  It's complete anarchy, and this is going to be 
one of the things that holds back Linux from more desktop penetration.  I'm 
not a Linux user (last time I used it on a daily basis was 1994 when it was 
like 0.995 and we were bitching that Diamond wasn't releasing dot clock 
information for their SpeedStar video cards), but the few times I've tried 
to use it and get help from my friends it's been disastrous.  Not because 
Linux itself was a mess, but because I'd have one friend that knew RH cold, 
but couldn't navigate my Mandrake install.  And then there's Debian (and 
FreeBSD) thrown into the mix.  So someone that "knows Linux" may not 
actually know how to admin some distro they haven't dealt with before.

At least our Windows users can download CandyCruncherDemo.exe and just 
double click and go (the MacOS X distribution was a bit of disaster, but 
nothing as complicated as the Linux stuff).  That's just not possible in 
the Linux world (if it were, I assume someone would have done it).

Our BeOS version is going to probably end up in the same limbo as Linux 
commercial games -- you can get it and play it, and you can get a 
registration code for _another_ version (Windows, OS X, MacOS), so 
technically you're not buying the BeOS version of Candy Cruncher.  That way 
we don't have to support it because, frankly, we're not a position to.

If I was developing a serious, large budget commercial game for Linux, I'd 
probably have to do something similar -- either make it officially 
unsupported (on-line games can get away with this -- you pay for the 
account, not the client, and you can choose which client works for you), or 
make the officially supported subset so limited that A.) you have a good 
chance of compatibility and B.) it will still work, but people using a game 
"For RH 7.x ONLY" on a Slackware install are on their own.

Anyway, I'm incredibly happy with the job that Ryan did.  If given the 
choice between ease of use and minimal size; or "standard" (read: obtuse to 
the uninitiated) and a bit smaller, well, I'll take the former.

Brian