[Freeciv-Dev] Re: (PR#3936) introducing native coordinates

[Jason Dorje Short <vze49r5w@xxxxxxxxxxx>]

Raimar Falke wrote:
> On Sun, Apr 13, 2003 at 04:51:46AM -0500, Jason Dorje Short wrote:
>
> > Raimar Falke wrote:
> >
> > > On Sat, Apr 12, 2003 at 08:24:06AM +0200, Raimar Falke wrote:
> > >
> > >
> > > > > but square_iterate and other local functions cannot be implemented
> > > > > in these coordinates
> > > > Why not? In the worst case I calculate all the vectors by
> > > > hand and hardcode them. Something like:
> > >
> > > Actually there is a "portable" way to do square_iterate: using
> > > MAPSTEP.
> > >
> > > At the start you go radius times NW.
> > >
> > > At the body you do a STEP W. This radius times. Than a STEP S. Than
> > > radius steps E and so on.
> > This is a clever(ish) way to define an operation on "native" coordinates 
> > that really uses map coordinates.  Unfortunately, in addition to 
> > being extremely inelegant
> I disagree. It is one place less which is topology dependent.
Actually it is one more place which is topology dependent - since this 
(DIR_DX/DIR_DY) would not otherwise be a topology-dependent operation. 
The only places that care whether the topology is isometric are 
native_to_map_pos, map_to_native_pos, and the mapview window-clipping 
function.
> > Additionally, you are defining map directions in terms of map
> > coordinates, rather than native coordinates - even though you
> > propose to remove map coordinates from the code.
> In the current code everything is a map coordinate. Map coordinates in
> your proposal are similar to the effective form in my proposal. My
> coordinates however are similar to your native coordinates.
In the current code everything is a map and a native coordinate, since 
the two are identical for non-iso maps.  The effective form in your 
proposal is equivalent to native coordinates.
> > If the directions don't even line up with your map, how is anyone
> > supposed to understand what's going on?
> I think it is easier to understand one set of coordinates where
> MAPSTEP does a magic thing as multiple sets of coordinates.
The problem is (1) the set of coordinates is itself very non-intuitive 
and (2) the hacks you have to do to get local operations to work are 
error-prone and non-intuitive.  Better to introduce a single, 
freeciv-wide solution that is easily understandable (explained in 
doc/HACKING) and works for everything.
> > Note, all of these are trivial if you are simply willing to do them
> > in map coordinates.
> Yes. But you don't have to scare people by introducing multiple
> coordinates (you have scared and confused me) and if these are only
> used in one function (map_distance_vector). square_iterate and MAPSTEP
> can be easy done without them as outlines before.
My original patch requested feedback on the explanation it provided in 
doc/HACKING.  Perhaps we should start this discussion over, and you can 
tell me what scares/confuses you.
jason