Bugs in NEW 2.4!

[Botsareus]

There is somthing wrong with the collisions when the program runs out of memory, otherwise they work fine. Its the evil vb ran out of memory bug. its 256MB  RAM (ram name is dumbus)
Otherwise the collision work perfict.

I love it when a robot only has 600 energy , it looks like a little dot, the only way I know its there is because its skin is sticking out ( witch is in a way another bug: the skin does not get resized), I prefer marking them with cross hairs better. If we don't mark them at all then no one can see them. They will be pixels on the screen, how the hell do you select that?

Delta mutations - I like that.

[Numsgil]

Delta mutations and Point Mutations occurr independantly of reproduction.

That is, there's a chance each cycle.

[Botsareus]

We still get ties stratching across the screen due to robots incorrectly jumping from one side of the screen to another. With big fat vegs you also see a lot of blinking on the sides for all the movment going on. That code needs a more realistic model in the future , we talked about this on its own thread somewere.

[Numsgil]

Yeah, it needs to be worked out so torroidal sims are really torroidal.  But that invovles some ugly math in alot of places.

[Botsareus]

I pinned it down the program running out of memory after I restart a simulation twice... This might be a huge problem for me in the future, my ram is fine...

And don't forget to fix the skin and make small bots easyer to select...

I gtg.

[Numsgil]

You can always zoom in.  In fact I'd encourage it.

Sounds like the program has a memory leak.

I can't figure out the skin bug.  Any who wants to show their worth as a programmer should totally isolate the function and fix it.

[Welwordion]

hmm when I make changes like disabling mutation or gravity they simulation tends
with bots that do not move or anything else, even ,more if I restart the simulations.

Ah by the way is there a changelog what new in 2.4 ?

[Numsgil]

There was a changelog, but the changelog was getting so long I abandoned it.

Assume everything is new.  Then ask specific questions somewhere and I can remember if I changed that or not :P

Or you can get The windiff ignore capitals program on the FTP and compare source code between 2.37.4 and 2.4.  (Haha, have fun :P)

Your physics settings are probably causing bots not to be able to move.  Open up the "advanced physics" panel and read me back all the sliders' values, etc.

[Welwordion]

^^ to know all the commands I can use would be good enough for the start.

I suppose if nothing is set, the settings are random?
Well cause I tried only with zero gravity.
with all kinds of fluid and all kinds of solids.
Is the rest random  enough to szop movement in so many cases?

oh and when I said nothing I also meant the do not shot ties.

hmm I tried to use another bot, since I remembered that other bots behaved different and when I changed back the problematic bot coudl do something again.

Bye the way usally first the unlearn to tie before they unlearn to move.

[Numsgil]

Some values are inherited from the old 2.37.4 physics page, some are set to 0.

What's your fluid set at?  I've noticed "thick fluid" is about impossible to move through.

I'm trying my best to get the wiki updated.

[Welwordion]

hmm strange different bots behave different on this matter.
Hunter seems to be immune against it while my bot niemand(currant test object for surrounding strategy) is quite weak against that effect:

Well I only wanted to show that in my eyes they programm behaves so random.

viscosity 0,01
bang efficeny 80%
all other 0 zero gravity
all other random
(just notice that mutations always seem reseted on enabled when I look into it,
mutation rates Button:all enabled-->all disabled right)

(just for the interested)
improved version moved to Bottavern multibot theories

[Numsgil]

I can't debug bot DNA, makes my head hurt.  That's sort of funny, isn't it.  :P

[Welwordion]

I did not want you to debug it ^^ just thought you might want to try the bot out for fun. Oh I edited the above post, put in my settings.

[Numsgil]

I'm in the library, but I'll bring a floppy later and copy it to see what it does...

I'll check the mutations page, I think I've noticed the thing you're talking about.

[Botsareus]

The Energy Proportional to robot size is nice and all. But in some ways it is too impractical and sertanly not realistic.

Lets say you have a robot with 32000 energy , and another one with 10 energy. The one with 10 energy ends up being a cell to another cell. We just did celluler strinking ^ 2. Zooming on that is just too impractical even if the zoom was to infinity (btw zoom is kinda week right now due to the limit witch needs to be adjusted too).

A real cell has tissues it has to sustain independent of its energy. A real cell can build up in pressure as well. So I prepose instead we use somthing like size = N + ScurveEquation were N is the tissue every cell has and ScurveEquation instead of liner is a pressure effect.

ex:

Private Sub Form_Load()
ScaleMode = 3
Show
For x = 0 To ScaleWidth
o = (ScaleHeight / ScaleWidth * x) * 0.2
Max = ScaleHeight * 0.2
'begin: shows you how mutch I hate plotting
y = (1 + 100 * 1.1 ^ -o / 10) / (1 + 50 * 1.1 ^ -o / 10) * ScaleHeight * 0.5
y = y * 2.1 - ScaleHeight - 30
'end
Line -(x, y)
Next
End Sub

Another serios problem (this it turns out has nothing to do with the momory overflow) is that the robots stop getting energy from the plants after a while. Run FB in F1 settings, see what I mean. dont forget high mutation rates.


I can't even find were the size of the robot gets related to the skin, I am pritty sure it's the problem with a memory leak because it is very rear.