Difference between revisions of "WhiteWhale (robot)"
m (→Performance: Comma) |
(Note on the controversy) |
||
| Line 13: | Line 13: | ||
The nanobot's most obvious shortcoming is its inability to do anything remotely object-oriented. There is no way to instantiate arbitrary objects. | The nanobot's most obvious shortcoming is its inability to do anything remotely object-oriented. There is no way to instantiate arbitrary objects. | ||
| + | |||
| + | The architecture has proven to be a bit controversial. | ||
== Performance == | == Performance == | ||
Revision as of 05:20, 3 October 2009
This page is reserved for Kinsen's latest nanobot. This robot builds on the work from NanoInterpreter.
The robot is in the RoboRumble.
NanoRumble ‒ APS: 77.96% (1st), PL: 147-4 (1st), Survival: 89.87%
The current source code can be downloaded from http://sites.google.com/site/dcvqksyb/robocode/kinsen.nano.WhiteWhale_1.0_src.jar.
Architecture
The robot understands a very small set of instructions: set, get, subtract, divide, if, call, and characterToInteger (CMD_SET, CMD_IF, CMD_SUBTRACT, CMD_DIVIDE, CMD_IF, CMD_CALL, and CMD_LITERAL). This instruction set provides the foundation for basic mathematical expressions. The expressions use prefix notation (e.g. + 5 2 instead of 5 + 2).
WhiteWhale has a large, but limited memory. The calls to Java functions are stored in STRING_TABLE and the commands are stored in COMMANDS.
The nanobot's most obvious shortcoming is its inability to do anything remotely object-oriented. There is no way to instantiate arbitrary objects.
The architecture has proven to be a bit controversial.
Performance
The interpreter robot is quite speedy compared to other bots due to various neat tricks, which you'll have to read the code to find out for yourself. :) Let's just say that it wasn't exactly a fast bot for a very long time.
As far as the nanorumble goes, WhiteWhale currently beats all the bots except for 4.
