Code center > Suggestions

Disabling bitwise functions

<< < (2/2)

Endy:
Not sure the dna is efficent or obtuse. Evolution is likely to take the most round-about method imaginable to reach it's solution.

Seems like the stores, inc, dec are the most meaningful parts of the dna.

If the de-bugger could somehow isolate which values are actually stored/inc/dec in the genome it'd be pretty useful.

Alot of what is in the genome really is "junk", mutation correction material, or fodder for new mutations to spring from. While useful to the bot, it doesn't actually do anything so isn't truly vital for us to see.

PurpleYouko:
Did anyone ever try using the partially complete de-bugger in the Excel Snapshot utility?

It is missing a lot of the bitwise stuff but could very easily be updated to include it.
The program contains a stepping function to move through the genome one command at a time and keeps track of all the values in the stack and all memory locations.
It even gives you the ability to change memory locations at will to see what effect it will have.

Jez:
Don't delete 'em! I haven't had a chance to fail to use them properly yet...  

From my POV as a returnee, the bots are almost scary in how mathmatically accurate you can make them now, eye5 is old news now having been replaced by (refxpos refypos angle setaim) or (refxpos refypos distance.) (I figure:) There is a way round aimsx/dx now using .aim. All I need to do now is figure out how to avoid using .up, and I shall have a formula of 'botic beauty!

Seriously, instructions that allow you to replicate old instructions like *.eye5 and avoid the previously relevant pitfalls, (like *.refeye, I can't ref .angle or .refxpos) shouldn't be added just to make the maths easier. It's not meant to be a robot simulating a robot (IMO) it's meant to be a bot simulating a life form.

Just to emphasise the inacuracy of bioforms, try poking a small spot at arms length with the tip of your finger really quickly, most of us will miss by a bit most of the time.

   Robots are 100% accurate 100% of the time, bioforms are a bit more hit and miss and need to practice first.

Numsgil:
Most old ways of doing things should still work.  Primarilly there are just some new ways of doing old things that are a bit more accurate.

Over accuracy is something of an issue.  It doesn't seem to create a problem for the bots as either human machinations or evolving virtual lifeforms, but it certainly strains believability for the bots being analogous to real life forms.

Just changes the combat tactics mostly.  It's difficult to sneak up on an opponent.  It's difficult to hide from a persuer (though Enitor Comesum gives it a pretty good try).  Pretty much all you can do is dogfight your oppenent.

Numsgil:
edit: Just a quick note, when a bot shoots a shot, there is a little innacuracy in its angle.  A spread of something like +/- 8 or 9 degrees.

Navigation

[0] Message Index

[*] Previous page