Darwinbots Forum
Code center => Bugs and fixes => Topic started by: Sprotiel on December 06, 2007, 05:34:50 PM
-
I thought this couldn't happen anymore, but I currently have a large number veggies atop each other in the lower right corner of my sim (see image). If you want the corresponding sim, I'll upload it.
-
Please post the sim. This can certainly still happen. Depends on collision elasticity.
-
There it is.
-
So what is happening here is that the starting position for Alga_Minimalis.txt is set to be a tiny tiny area in the upper left hand corner of the field. Was that intentional on your part?
Veggy repopulation is piling the veggies up, materializing them on top of each other. They do slowly separate (you could speed this up by increasing the collision elasticity if you wanted).
I'm not sure I want to dive into the collision physics to try to improve this right now. The problem is that we run the risk of injecting movement nrg into the system that can result in all sorts of strange oscilations and such. But I'll take a look at the overlap case once more and see if I can tweak it.
-
So what is happening here is that the starting position for Alga_Minimalis.txt is set to be a tiny tiny area in the upper left hand corner of the field. Was that intentional on your part?
Certainly not! I didn't put A Minimalis in the species list and didn't want it to be repopulated.
-
But you were running the sim in internet mode and you do have veggy repop turned on.
When a new species teleports into a sim in internet mode, I create a new species entry for it and wire that up to the DNA file if the names are the same (soon this will be unnecessary).
And veggy repop now chooses randomly between all veggy species in the sim for which a local DNA file is available.
So, one bug (besdies the overlap) is that I don't default the starting position correctly for new species that are teleported in. That is why they are all getting repopulated into the corners.
But the larger issue of course is what veggy repopulation should imply when there are more than one species of veggy in the sim.
I have long argued that veggies are too powerful, that we should extract some price for all that free nrg. I have been consistant in this and I have been consistantly opposed. People appear to want author and breed super veggies. So be it. But this implies that in internet mode that veggies are not just part of the environment, cannot be unique to each sim, must be teleported and must be given a level playing field accross all sims. They are active combantants and as such, should not be given a home field advantage due to repopulation favoring the home team. It won't be long before I allow repopulation direct from the running DNA without a bot DNA txt file and thus allow repopualtion of any veggy, even really nasty potentially mutated killer veggies people write and first introduced in some far off sim. I must, to keep the internet mode playing field level.
I'm sure people object. Let's hear it.
-
It won't be long before I allow repopulation direct from the running DNA without a bot DNA txt file and thus allow repopualtion of any veggy, even really nasty potentially mutated killer veggies people write and first introduced in some far off sim. I must, to keep the internet mode playing field level.
I'm sure people object. Let's hear it.
I won't object. It makes perfectly sense to me. I've been a bit annoyed by the fact that the repopulation only chose to work for one veg species in the sim. Sometimes when I start a sim with more than one veggie species I make sure I have repopulation turned off and only rely on the cost reinstatements to prevent extinction. So I guess if you implement this it will be the vegs with the lowest numbers that will be repopulated?
-
So I guess if you implement this it will be the vegs with the lowest numbers that will be repopulated?
Actually, this part of it is already implemeneted in 2.43u. If you have multple veggy species in your sim for which DNA files exist, repopulation will choose evenly between all of them whether you are in internet mode or not. Their order in the species list does not matter.
-
The starting position for teleported in bots issue is addressed in 2.43v.
-
Can we get mutated veggies to repopulate? I've long been frustrated that my long term evo sims can't have the veggies coevolve because they all die off.
-
Can we get mutated veggies to repopulate? I've long been frustrated that my long term evo sims can't have the veggies coevolve because they all die off.
My thought was to randomly select an extant individual of the species to use as the repopulated genome in the case where the species is allowed to mutate. In the event there are no extant individuals left at re-pop-time, I would use the "base" genome which would either be the DNA txt file or the genome of the first bot of that species teleported in.
-
Well, better than nothing I suppose
-
Always interested in better ideas...
-
I'm not sure of the logistics, but if you take all the existant bots, plus X recently recorded dead bots, and choose randomly among them for each new bot to place, I'm thinking you can either eliminate or drastically reduce the effects of muller's ratchet on repoped veggies.
-
When a new species teleports into a sim in internet mode, I create a new species entry for it and wire that up to the DNA file if the names are the same (soon this will be unnecessary).
That is a problem and opens many possibilities of abuse in Internet Mode. The correspondence between a genome and a species name is only valid at sim startup and for a specific computer.
-
That is a problem and opens many possibilities of abuse in Internet Mode. The correspondence between a genome and a species name is only valid at sim startup and for a specific computer.
Which is exactly why I plan to break the dependency between a DNA txt file and the bot for the purposes of veggy repopulation.
An alternative would be to completely disable repopulation in internet mode. Veggy reproduction would be it. If the bots wipe out the veggies in a specific sim, nrg flow into the sim stops until a veggy teleports in from elsewhere. Terrarium worked this way. People coded their bots to not eat veggies all the way down to nothing. Preventing other species access to veggies became a battground for competition. It also allows veggy species to go extinct and stay extinct in the internet mode megasim, which is important if people want to compete by authoring veggies.
-
I agree with that. If vegs just keep repopulating it won't be a progress in IM. It's more interesting to see how it evolves and which bots will be dominant. If some species gets extinct they shouldn't just magically appear again. The problem is though that there is a limited amount of bot species and I'm bad at writing new ones. I only have 2 zerobot species, so it's hard not to restart with a new sim and try again, so I don't know.... Restarting is kind of the same thing as repop. Hmm...
-
I'm going to resolve this bug for now sicne the original bug was fixed. If people want repopulation turned off in internet mode or some other design change, let's track that in the suggestions forum.