Rebuild the "lost code"

[PurpleYouko]

Try playing a sim in 2.37.6.

Choose Alga Stratificus as the veggie and Anon Terrifica as the animal.

Enable "pond mode", corpses and bouyancy. set to size 2, disable all torroidal settings and run it.

Make sure to set the corpse decay to a reasonable value like about 5 points with a 5 cycle interval.
Set the light intensity to 30 - 50 and the gradient to 1

The veggies will all float to the top where they will recieve sunlight and grow. The robs will eat some and then you will see little white corpses falling to the ground whle spitting out shots in random directions.
When the sun goes down the veggies will drop to about mid screen and bob up and down a bit. Next morning, up they go again.


How about making 2.4X do this stuff too?

[EricL]

You got it.  On the list for 2.42.7.

[Elite]

PY, we really value your input on this  

Eric's aim is to have 2.4 completely backwards compatable and capable of doing everything 2.37.6 does, so if there's anything you want to see in 2.4 that 2.37.6 does, or any differences between the two versions, then please go ahead and say.

How about first making ties do exactly what ties in 2.37.6 do, before making alterations?

Eric, I recomend running a few sims in 2.37.6 just to get a feel for the program, and taking a look at the following multi-bots:
Inchworm (and it's variants)
Hexagonis
And also looking at some tie feeders

I have nearly finished my exams so I might give that tie tutorial on the wiki another go

[PurpleYouko]

How about first making ties do exactly what ties in 2.37.6 do, before making alterations?

That is a very good suggestion. It will give a nice starting point to move on to bigger and better things such as the possibility of infinite ties etc.

[EricL]

So I do want to work torwards making 2.4X have no take backs relative to 2.3X.  To that end, yes I will continue to put in (significant) time on making ties work to everyone's expectations and also keep filling in the gaps w.r.t. things like pond mode, etc.  Have no fear, this remains a primary goal for me.  That said, I want to make some points:

1) It's a lot more fun to write new code then it is to try to understand someone elses code and find/fix the bugs there-in.  I'm happy to spend significant time on the latter, but to keep my sanity and interest, I have to spend some percentage of each each release on my own stuff.  Each release from me will be a balance.  Each may have some new stuff, but also will hopefully continue to take steps towards no take backs.

2) The physics of 2.4 are just very different than 2.3X and changing this is not within the scope of things I am willing to do in 2.4.  There will be cases where sysvars work correctly in both versions, but the results will be different.  Inchworm for example, depends upon the underlying tie physics of 2.3 - that ties have a certain default springyness, that bots have a certain mass, that setting the tie angle or tie length when one bot is fixed torques or pushes the other bot in a certain way.  It is quite possible (indeed, I beleive it to be the case today for inchworm) that all the underlying sysvars that the bot uses work and work correctly, but that the marco behaviour is different due to the underlying physics differeces in the two versions.  I am willing to enertain suggestions to close such gaps, but only up to a point.  I'm not sayign this is the case yet, but at some point, we may need to see a 2.4 inchworm which in order to work "correctly" has to do things a little different, say set the tie stiffness explicitly becuase the default beahves irrconcilably different than that in 2.3X.

Those points made, I will continue to plug away at closing the gap.

[PurpleYouko]

Eric

I have no problems with making minor changes to robots to fine tune them to work in a newer version. I don't expect you to tune the software to make the bot work to its greatest potential.
Every new version requires some of the older bots to be tweaked and rewritten to some degree. just so long as the principles work.