[Freeciv-Dev] Re: Development Strategies

[Mike Kaufman <mkaufman@xxxxxxxxxxxxxx>]

On Mon, Dec 03, 2001 at 09:24:33PM +0100, Per I. Mathisen wrote:
> On Mon, 3 Dec 2001, Daniel L Speyer wrote:
> > The current system of flags *hasn't* worked for getting all sorts of
> > interesting rulesets, so let's try changing it.
> 
> A lot of things _can_ be done with the current ruleset structure, but very
> little _has_ been done. Why? I don't think the answer is as obvious as you
> try to make it appear.
> 
> The one biggest problem with the current rulesets is, IMHO, that the
> number of improvements is still fixed, along with some improvement
> effects that are still hardwired (which is why the number is fixed). If we
> fix that before the next version and get some decent isometric tiles and
> some music+sound, I think this would give a lot more interest in writing
> rulesets for the effort we put into it.
> 
> I'm sceptical of a full rewrite of everything. It is a bit too massive
> effort for one person, and design by commitee doesn't work for open
> source. I'm afraid all that talk is going to melt into thin vapour. I

I too am a bit skeptical. I cringe when I read that some project just
_has_ to go OO. I certainly haven't been convinced that is the case
here. For sure, there are subsystems freeciv that need a major rewrite
(I applaud the people doing the AI cleanup. A couple months ago, I had
an AI idea and got scared stiff when I started looking. I also think
that a lot more headway does need to made on generalized improvements,
especially before a lot of new AI work gets done.)

This is not to say that this discussion isn't useless. It serves to
highlight, for example, the weaknesses that not having generalized
improvements presents. Or what a bitch the AI will be to code if
rulesets are changed even a bit. But that would be just as true in a 
c++ environment.

I just happen to believe that if freeciv becomes a c++ project, it's
going to lose a chunk of the developers in the changeover. (certainly me,
but I'm not hot shit or anything...)

I didn't think that I'd be quoting Trent Piepho, but this _is_ supposed to
be a game, not a research project. 

[snip]
> Oh, and if _I_ were to rewrite something, I'd go for an SDL+OpenGL client
> with libuta and various SDL support libraries (PNG,mixer,etc)...
> wooohooo, now we're talkin'!.... instantly multiplatform, even!

Well, I don't know about SDL, but for a good 10 minutes a month back, I 
thought hard about an OpenGL client. I'd probably embed it in a Gtk
frame (you can do that?) and then all the map rendering would be
with OpenGL. It certainly would the cool as all get out. (and it doesn't
have to be a rewrite, just add a /gui-opengl)

-mike

> 
> Yours,
> Per
>