[Freeciv-Dev] Re: Server/ruleset unification [Was [RFC PATCH] init_techs]

[Raimar Falke <hawk@xxxxxxxxxxxxxxxxxxxxxxx>]

On Fri, Sep 28, 2001 at 09:42:14AM -0700, Arien Malec wrote:
> 
> --- Raimar Falke <hawk@xxxxxxxxxxxxxxxxxxxxxxx> wrote:
> > > simple_set_statement is intended for 
> > 
> > > toplevel attributes:
> > 
> > And that is the point. What makes such attributes toplevel attributes? 
> > Why can't "huts" be under a "map" object? Why can't "ailevel hard" be
> > under "aicontroll"? Why can't "min_dist_bw_cities" in another category
> > (event if this is named "misc")? IMHO they are special because of
> > historical reasons.
> 
> The distinction I mean between "toplevel attributes" and "object attributes" 
> is
> that, regardless of whether you call ailevel "aicontrol.ailevel" or "ailevel"
> or "game.ailevel",

>  you can treat the whole string as a unique identifier that >
>  identifies the one and only one ailevel variable you want to
>  modify.

Who says that there is only one AI?

> However, for nation.greek.init_techs, you have to find the greek
> nation *first*, before you can modify it.

Yes. But this distinction is still technical nature. We don't have to
force the user to also make this distinction.

> I would strongly recommend that if we want to use "dotted syntax" to organize
> variables, that we use a different syntax to identify objects.
> 
> e.g.
> 
> game.ailevel
> 
> vs., e.g.
> 
> nation[greek].init_techs
> 
> 
> I'm also leaning towards the simpler set/create syntax, because it makes the
> lexting and parsing code much easier. For instance, all commands can be 
> newline
> terminated.

At least for me this isn't the great unification.

        Raimar

-- 
 email: rf13@xxxxxxxxxxxxxxxxx
 "That's fundamental game play!  My main enemy is *ALWAYS* fighting 
  a 4-front war.  I make sure of it!"
    -- Tony Stuckey, freeciv-dev