[Freeciv-Dev] Re: Is the city names patch good? (PR#1126)

[Raimar Falke <hawk@xxxxxxxxxxxxxxxxxxxxxxx>]

On Tue, Dec 11, 2001 at 01:53:35PM -0800, jdorje@xxxxxxxxxxxxxxxxxxxxx wrote:
> Raimar Falke wrote:
> 
> > On Tue, Dec 11, 2001 at 02:38:57PM -0500, Jason Short wrote:
> > 
> >>Jason Short wrote:
> 
> >>Upon further thought, I have come to the conclusion that the natural 
> >>city names sytem _is_ fatally flawed, and should be replaced (preferably 
> >>by a better natural city name system, perhaps one like I describe above).
> >>
> >>The flaw is this: it discards the city order from the ruleset.  This 
> >>will mean that it is impossible to convert the ruleset back to on 
> >>ordered system.  (By "impossible", I mean it would be just as much work 
> >>as building the ruleset from scratch.  It would still be possible to 
> >>revert to a previous CVS version, but all additions to the ruleset would 
> >>be lost.)
> >>
> >>If the 64-nation limit is relaxed before the next release, then after 
> >>the release we're going to see an influx of many more nations.  And all 
> >>(or most) of these nations will use natural city naming.  It will then 
> >>become infeasible to change systems at a later date.
> >>
> >>All I'm really arguing for is a change in the ruleset syntax (perhaps as 
> >>I describe above).  If somebody (preferably the original author, maybe 
> >>me) provides a patch to do this, would it be accepted?
> >>
> > 
> > I see the problem. However the other problem (putting more info on a
> > name) is the more important reason (at least for me now). So I will
> > accept such patches.
> 
> 
> Here is a preleminary patch to accomplish the task.
> 
> 
> It is intended to accept rulesets of the form
> 
> cities =
>    "Washington(coastal,river),New York(coastal,river)", ...
> 
> However, the string parsing code needs work (it can't handle whitespace 
> or bad capitalization, for one thing).
> 
> It should work as-is with the old (pre-natural-names) nation rulesets 
> (do cvs up -D "2 weeks ago" to get this).  However, to get real results 
> you need to redo the natural names in the rulesets to match what it 
> wants (I'll provide some separately).
> 
> Also, note that you can still label any "terrain" type in the rulesets. 
>   This terrain will be associated with the city if the city is "near" 
> it.  With a complicated enough ruleset, you could end up getting some 
> pretty well-named cities, I think.
> 
> All in all, I think it will do well.  It certainly needs more work, though.

It goes in the expected direction.

        Raimar

-- 
 email: rf13@xxxxxxxxxxxxxxxxx
 "Reality? That's where the pizza delivery guy comes from!"