Re: [piecepack] Re: Is PP really that ZOG friendly?

[Massimo Biasotto <massimo.biasotto@...>]

Clark D Rodeffer wrote:

> All I've done so far is manipulate Matt Worden's graphics library
> into a format compatible with ZOG. (The easy part.) I haven't used it to
> program any ZRFs yet, except just to see if the pieces overlayed on tiles
> properly. If anyone else on this list has experience using or programming
> with ZOG, I'd be happy to exchange notes.

Hello,

I'm rather new both to piecepack and ZOG, so consider my comments as a
newbie's view.

After subscribing this group and seeing Worden's graphics library, I had
the same idea: adapt it to use in ZOG. Then, as an exercise in ZOG
programming, I've implemented a zrf version of "Easy Slider". It wasn't
difficult and the game works fine in ZOG, so I browsed *all* the over 60
existing PP games in search of something more complex to implement.
Well, I have now given up the idea, since I haven't found very few games
"ZOG compatible", and none of them interesting to me.

These are the main points that in my opinion make difficult to implement
PP games in ZOG (to be played against AI and without custom engines)

1) Of course the "hidden information" problem. In some cases there are
workarounds even for playing against AI. For example, when you need to
setup the board with random face-down coins to be flipped during the game,
you could place generic coin pieces during setup and then have the
Random Player to create a random coin at the moment of flipping, so even
ZOG doesn't know the value beforehand.
Even in this cases, it is always troublesome to workaround the problem,
and it adds unnecessary complexity to the zrf (which affects AI playing
strength).

2) In ZOG it's not possible to have multiple pieces in the same position.
This is a problem for PP games, beacuse often the board is not fixed but
it is built with tiles during setup or the game itself, and then you have
to play pieces over these tiles. In ZOG the board is only a fixed bmp image;
if you create it by placing "tile pieces", you can't then play pawns or
coins over them.
Maybe there are workarounds for this too, I haven't tried anything yet.

3) Goal, or winning condition. Most games use as winning condition something
that is not available in ZOG: scoring. At the end of the game each player's
scoring is computed in a large variety of ways, some of them quite complex.
I have no idea how this could be implemented, apart the most simple cases
where you could probably create a dummy piece for each score point and set
as winning condition to have the majority of these pieces. Even so, I have
the feeling that AI would play very bad.

Conclusion: IMHO Zillions in its current form is only good for abstract games,
and pushing it to handle something else is not worthwhile (a lot of effort to
achieve poor AI play).


The situation is a bit different if you are interested only in network play
against humans, but also here I think there are better tools than ZOG, for
example Thoth (http://digilander.libero.it/zak965/thoth/). Just look at the
wonderful Thoth modules available at http://geoman36.tripod.com/thothgames/
With the graphics library already made, it should be easy to implement many
PP games, and certainly easier than ZOG.


Massimo Biasotto