Manipulating multiple ties at once

[Elite]

It's a little complicated ...

1 .tienum store
200 .tielen store
2 .tienum store
50 .tielen store
314 .tieang store

^ How about, when the program detects that different number is being stored into .tienum, it stores all the actions that have been requested so far into the 'to-do stack' and then sets .tienum. Multiple ties can be manipulated each cycle.

Currently, the above gene would act the same as:

2 .tienum store
50 .tielen store
314 .tieang store

But I'm suggesting that rather than the new .tienum overwriting the previous, the program should execute what has been requested first

Understand?

Is this doable?

[Numsgil]

Have you looked at the "New Tie Paradigm" on the wiki?  It's how ties are going to work in 2.5.  It does something very similar (basically only differences in semantics).  Though instead of tienum being a sysvar it's a command (and called settie).

[Elite]

Oh, so it is

I should really check things like that before posting

Is there going to be any changes with viruses or am I free to post some new virus suggestions (maybe even an entirely new paradigm)

[EricL]

I don't think it would be too difficult to do this just for .tienum.  But the special casing for this one sysvar makes me nervous from a design and elegance perspective.  I think we might be better served to invent some truely parallel gene execution mechanism.  Maybe codules?

[Numsgil]

Eric - Originally I had an idea for "chromosomes" behaving like simoltaneous threads, each with their own instsance of the 1000 memory array.  (Or you could think of it as the chromosomes "filter" reading through their past actions, if that makes any sense ).

The problem that I could never find a satisfactory solution for was how to resolve conflicting commands.  -1 .up store and 1 .up store.  There are some obvious solutions, but they all have various conceptual problems depending on the sysvar.

Elite - I haven't thought about viruses except that viruses should be codule based and conspec specific (to a limited degree).  ie: pass codule 2 to another bot, which happens to have code to pass itself on to other bots.  Bots that don't execute codule 2 will have nothing to worry about (non-conspecs).  These transfered codules replace whatever was there before them.

Anyway, that's as far as I've gotten.

[Elite]

Oh good, I've gone further with viruses

I'm formulating an entirely new virus mechanics that should be more realistic than our current system. Conspec horizontal gene transfer can then be handled by another method (as my viruses are aimed at non-benevolent pathogens).

I'll start a thread when I'm ready.