[computer-go] How to prevent Copyright violations?

[Chrilly]

Dear Don,
>
> I think I sold about 50 copies in Europe but I found out later that half
> the population had a copy (or so it seemed to me at the time.)
>
That seems to be the number of honest chess-users in Europe. Fritz sold 
about the same.
Yes I know that Rexchess was for a time very popular. But I never had one. 
It was somewhat before the time I was really active in computer-chess.

>
> I have a palm GO program that has made some extra "spending money" for
> me.   You must purchase a registration key that I send out to the user
> when PalmGear tells me.   It has a simple hash protection scheme and
> it's easy to break with a debugger.   I know it has been successfully
> hacked, but I have no idea what effect the hack has on my sales.   It
> has made a pretty good chunk of money - but it was spread out over a lot
> of time.   Not nearly enough to make a living but probably enough to pay
> for my time developing it.
>
Is this a full 19x19 programm or a 9x9? I assume its a "normal" programm and 
not MC?
Are the commercial conditions in Go really better than in chess?

I make with Suzie some progress, but its harder than I thought. Everything 
must be programmed very precise. E.g. conditions for links or eyes. If one 
thinks that condition is not so important, make a shortcut, it shows up 
already in the next game. This is in my experience not so dramatic in chess. 
In the contrary its in chess even sometimes counterproductive to make very 
precise and detailed rules. Maybe it was the same when the programms 
searched only 5 plies like in Suzie.

I had an interesting search experience. Suzie searched the first plies 
full-width, the last ply at the horizont only the best 15-moves of the 
history-heuristic. The programm tried then to escape at higher search depths 
from ladders, although the eval said correctly, that the prey can be 
catched. This happened, because the the Ladder-Killer was usually not in the 
best 15 moves and the hunter did not play the move on the last ply. I search 
now full-width again and this works better. Well full-width with the usual 
pruning mechanism like Null-Move, Multi-Cut and Futility. But this has at 
least no effect on such things as ladders,. It has of course 
horizont-effects for things the eval does not know.

Chrilly