Re: More Core War
[Martin] >You guys are making really rapid progress
Well, I think the reason I was able to get a compiler out so
fast was I really liked Rod's approach - it was simple.
I think we started with the wrong idea in mind: ICW88 compliance.
That introduced more complexity into the project than we could
handle. I certainly couldn't have written it as quickly without
the things I learned last summer tho :)
>My recollection is that the routine ID's were dynamically set each
>time the program was run so the actual values set would vary from
>run to run, but still the constant (say) MOV would always point to
>the correct routine. But I'll have to check.
Yeah, you're probably right; I've lost all my notes too <g>
The reason we can't use routine_id() here is that the compiled
warriors require static numbers. DAT's routine_id() might be
0 today, but change to 7 with the next update. In order for the
warriors' compiled code to work, the numbers have to stay the same.
It worked for us because we didn't try to compile the warriors into
an intermediate format (although we were toying with the idea, right?),
we just took the Redcode straight to the core.
I'm still trying to figure out how to implement labels in my compiler :)
>I was also developing some ideas for displaying the evolution of the
>game. I'll see if I can hook that up to whatever Corewar system you
>guys get together.
I can't wait to see your code Martin, it's probably gonna be awesome!
Adding on to an existing system is alot easier than writing one from
scratch, so it shouldn't take too long.
----->Buddy
budmeister1 at juno.com
|
Not Categorized, Please Help
|
|
