Bruno Wolff III <[EMAIL PROTECTED]>:
> I would suggest avoiding floating point as there can be rounding issues.
> Using rational numbers (which is essentially your scale factor suggestion)
> will preserve exact values.
In IRC discussion, Sirp shot down the idea of fractional MPs pretty
thoroughly. Others objected strongly to MP scaling. I don't think either
approach is viable. There may be a third way, however...
> I also have a question for Eric. Is the assumption here going to be that
> the entire hex is filled with road? Because another way I could see
> proceding with this is to make road type separate from the base terrain
> type, so that the terrain type would set the defense values for units,
> but having having the road type would replace the normal movement cost
> for the terrain with that of a road. I though (but could easily be wrong)
> that bridges worked like that.
They look like they do, but they don't. :-) For gameplay purposes, a bridge
is a road (was grassland before my change). The visuals are handled
separately from the movement cost.
I said there may be a third way. I think Ivanovic was groping for this
idea a couple days ago; I was going to write it up in my proposal, but
spaced it (it was the missing third alternative on my list).
It might be doable to give units an ability that is, in effect, a reservoir
of 1 or 2 points that they can only use on roads. If I implement this,
the WML would look something like this:
#define ABILITY_ROAD_MOVEMENT_1
[ability]
id=road_movement_1
name="Road Movement +1"
deescription="Unit has 1 extra MP per turn usable on roads only"
apply_to=road
value=1
[/ability]
I think it turns out that given the movement range of non-fliers,
an {ABILITY_ROAD_MOVEMENT_1} and {ABILITY_ROAD_MOVEMENT_2} would cover
all the interesting cases. This would avoid all the headaches associated
with fractional MPs and scaling MPs. The only drawback is that the C++
support may be rather tricky, depending on how movement calculations are done.
Sirp is skeptical about the road bonus idea but has not ruled it out; he
says he'd need to see it tested in some custom content before wanting
it anywhere in mainline.
Fair enough. I guess that makes it my responsibility to implement
something like the above, then write a proof-of-concept scenario or
two to demonstate that it gives interesting results.
--
<a href="http://www.catb.org/~esr/";>Eric S. Raymond</a>
_______________________________________________
Wesnoth-dev mailing list
[email protected]
https://mail.gna.org/listinfo/wesnoth-dev
