Re: THE problem with xconq (long)

[Stan Shebs <shebs at apple dot com>]

OK, I've been successfully trolled... :-)

"A. Rick Anderson" wrote:
> 
> At 02:18 AM 2/1/2002 +0100, Hans Ronne wrote:
> > >I have been following XConq from the days of XConq 5 quite passionately.
> 
> >Stan fixed most of the problems back in 2000
> >before he started on the SDL interface.
> 
> This is also about the time when he rejected suggestions that if he was
> going to do a major re-write effort, that he consider moving the code base
> to Java.

More accurately, I wanted to see where the arrows were in somebody
else's back first, in other words what were the pitfalls in doing
something on Xconq's scale in Java.  And as far I've been able to
tell by watching the various announcements about game projects and
the like, there still haven't been any open source Java strategy
games that are complete to hand out to users.

The activity closest to reality I know of is FreeCiv, which has a
considerable developer community despite also written in plain C,
plus a few people working on a Java version.

> >THE problem with xconq is that nobody is working on it any more. Every
> >project needs a critical mass of interested developers in order to keep it
> >alive, particularly if it is done on a purely hobbyist basis.
> 
> I've been following XConq since before Stan left Utah, but I have no
> interest in working with computer technology that was obsolete back
> then.  Obviously, unless he's getting paid for it, neither is Stan.  He
> managed to convince his employers to pay him to write a game based on 'C',
> using a Lisp-based scripting engine.  Congratulations to him for pulling it
> off for as many years as he did.

Ahem, I was never paid to work on Xconq.  In fact I did it in C
despite being a grad student researching and promoting Lisp
technology, because Lisp was already failing to deliver.  I had
written a graphical Wooden Ships & Iron Men clone in Lisp the
previous year, but it was pretty slow even though I knew a lot
about how to optimize Lisp, and the graphics code was completely
wired to the HP Chipmunks that the game was written on (yes, HP
had a Lisp-everywhere initiative once upon a time).  The initial
C version went together quickly and easily, and my fellow students
were having fun with it within weeks of me starting the project.

> I'd love to learn how to create a multi-player, RTS with Java and I would
> be glad to contribute to a project being led by someone who has been there,
> done it and has the T-Shirt and obviously has a lot to teach me about this
> area.  However, there is no way that I can afford to dig-out my K&R book
> (1st Edition), just to putter around with a piece of code that even its
> creator can't figure out what is wrong with it.  Why would I want to ride
> around on the weekends in a busted up, smoggy, diesel two-ton, when I can
> cruise on an Electra-Glide?

Heh, but as it turns out there is a whole subculture of automobile
enthusiasts who do nothing but fix up old cars and drive them around.
One of my coworkers owns two Ferraris, but spends considerable effort
on restoring a Model T...

The key thing here is lack of motivation.  There is nothing stopping
you from recoding Xconq in Java, indeed it ought to be easy, because
you're handed all the known working algorithms, and since you only
need to read the code, the minutiae of C shouldn't matter (your K&R
book won't help you BTW, Xconq has been in ISO C for some time).
If Xconq in Java would be as popular as you suggest, then people would
flock to work on it, and your version would have the active developer
community.

But the reality is that even you, as the most vociferous proponent
of rewriting Xconq in Java, apparently can't be persuaded even to
lift a finger to start on it.  If you won't do anything, what reason
is there to believe anybody else will?

Stan