Messages

31

Bug reports / Inconsistent handling of logic functions

« on: January 21, 2007, 06:27:13 PM »

In 2.42.9u, on an empty condition stack:
* 'not' is true
* 'and' is true
* 'or' is true
* 'xor' is false
*'(true) and' is true
*'(false) and' is false
*'(true) or' is true
*'(false) or' is true
*'(true) xor' is true
*'(false) xor' is true
This is inconsistent in the sense that it is not possible to assume that a boolean is generated to fill the missing operands, contrarily to what happens with the integer stack (missing operands are replaced by 0).
Since an empty stack is true, as far as gene execution is concerned, what would make the most sense would be to use 'true' as default, fill-in value. Besides, it only causes a change of behaviour in the 2 italicized cases out of 10.

32

Bug reports / Bug in statistics

« on: January 21, 2007, 03:10:33 PM »

It seems my former bug report was incorrect. In 2.42.9t (and probably also in .9u), mutations and mutations/DNA length appear to be incorrectly multiplied by a factor (total number of bots / number of bots in species).

33

Evolution and Internet Sharing Sims / Maca's Zerosim

« on: January 21, 2007, 10:09:09 AM »

Could you post whole genomes, as attachment perhaps? It'd be more informative and much easier to check your explanations.

34

Suggestions / chloroplasts

« on: January 21, 2007, 09:31:00 AM »

I was thinking for a more evolutionary based db we could have the ability for free nrg to be a gene insted of just given to the bot like a chloroplast gene?

Agreed, but the real question is how? You should read this thread. I think it's the best idea we've had so far.

35

Evolution and Internet Sharing Sims / Phylogenetic trees from sim files

« on: January 19, 2007, 09:03:09 PM »

I've created a small Python script to analyse mutation histories in saved simulation files. What it does is building a phylogenetic tree from the mutation details (taking advantage of their being saved in plain text). There must be some useful things to do with this, but I only print the tree with the numbers of bots having a particular genome, or having it as an ancestor.

Limitations : it ignores bots that never mutated, may give nonsensical results if you've ever stripped mutation details or if you have more than one species mutating.

To use it, you need to have Python installed. If you know some Python, I hope you'll figure out what it does, otherwise, I'm also including a handy batch file on which you can drag and drop .sim files.

36

Bugs and fixes / 2.42.9 s>t bug

« on: January 16, 2007, 10:28:02 PM »

Some recent change done when going from s to t version completely kills my current evo-sim.  Try loading this sim in these two different versions and you'll see the difference right away.

Your problem lies obviously with rotational costs. I guess the unit chosen to express it is way out of range (it's labeled as energy per bang, but a bang is a unit of (linear) momentum, not of angular momentum).

37

Bugs and fixes / 2.42.9 s>t bug

« on: January 16, 2007, 06:19:42 PM »

Correct link is: http://www.darwinbots.com/shvarz/01-15%20work.sim

And could you please zip it? 12.3 Mb is a bit much...

38

Suggestions / Genome structure

« on: January 11, 2007, 08:45:30 PM »

It isn't the store the one picking the value, but the start.

So, when the last number in the normal stack is a true value, as it would result from a condition operation, the following start activates. Stores wouldn't change at all in this last suggestion.

It's so a small change it at least could be an option...

Removing the condition stack is not a small change, but I think it's worth considering.

Your proposal isn't totally clear so here is a detailed proposition of implementation:

• cond stack is removed, bots only use the integer stack
• tests output 1 for true and 0 for false
• for flow commands, 0 means false and everything else means true. The first 'start' command after a 'cond' sets a flag that's used by subsequent 'start' and  'else' to decide if the corresponding blocks of code are executed.
• logic functions behave as 'cond' for input and as tests for output

There's one problem with this, though. It becomes impossible to automatically and together all conditions when a 'start' is executed, which breaks backward compatibility.

39

Bug reports / Graphs not correctly resized

« on: January 09, 2007, 03:18:35 PM »

Some graphs aren't resized properly, energy stats seem to be most affected by that.

While I'm on the topic, I'd like to point out also that the fact that all species appear on graphs makes most of them barely useable once a sim has settled somewhat since it forces the use of an unappropriate scale.

40

Suggestions / Genome structure

« on: January 09, 2007, 12:23:44 PM »

I've been taking a look at the Evolve4.0 dna structure, since they seem to have something like this already. I do like how they can have conditions daisy chained to each other ie.

Conditions can easily be chained in DB:

'A
cond
...
start
...
1
stop

'B
cond
0 !=
...
start
...
1
stop

'C
cond
0 !=
...
start
...
stop

41

Darwinbots Program Source Code / Setting up build environment for C++ fork

« on: December 19, 2006, 10:23:01 PM »

OK: I installed VS 2005, vanilla. I opened the solution included in Numsgil's snapshot, and hit build. It then barfed and said it couldn't find two includes: Vectors.h and fx.h.

In some files, the path is correct, "../Common/Vectors.h" but in other files it is just "Vectors.h" so I changed them all to be correct.

I assume fx.h is foxgui? What version of foxgui are you using? I'm looking for the fox gui package on the internet, and the latest stable they have is 1.6.20. The old stable is 1.4.35.

How shall I proceed?

<fx.h> is indeed FOX, you need version 1.4 (I guess the latest stable will be fine), though 1.6 may work. You need to setup your environment so that the FOX headers are in your include path. You'll probably need to link against several libraries, FOX's site should give you an accurate list somewhere.

Warning: I haven't built Nums' source for some time (and never with VS) so the above may be incorrect or incomplete.

42

Bug reports / Bug in statistics

« on: December 19, 2006, 08:49:56 PM »

I think every per species average wrongly divides the sum of the relevant properties by the total number of bots instead of dividing by the number of bots of the species. E.g. if you start a sim with 10 vegs and 5 animals with 5000 nrg each, the average energy for vegs will appear as 3333, and for animals as 1667.

43

Bug reports / 'You need MSSTDFMT.DLL'

« on: December 19, 2006, 08:45:48 PM »

Since 2.42.9L (or earlier), I get a runtime error 713 'Class not registered' whenever I try to access the custom costs panel. It's not really a problem since downloading the offending DLL AND typing "regsvr32.exe msstdfmt.dll" in the command box of Windows' Start menu solves it, but it might bother newbies.

44

Suggestions / League Settings

« on: December 19, 2006, 08:19:13 PM »

I once set the DNA upkeep to something like .000 000 000 1 or something like that.  I don't think any of my bots ever lost nrg.  If you get a half second, can you check on this?

That's 10^-10 ! No wonder you didn't see any effect: say your bots have a 1000-blocks DNA, and live 100 000 turns, over the course of their life, they'll have lost 10^3 * 10^5 *10^-10 = 0.01 energy.

45

Suggestions / Variable length arrays for the bots

« on: December 19, 2006, 08:14:38 PM »

I don't think it's a good idea. Most likely, there will be evolved bots that write to it every turn. Suppose there are 100 of them, living 10000 turns on average  and that they use this foreach every turn. This means you get one million tests to perform every turn... A high price to pay for a dubious benefit!