Epigenetic Memory Suggestion

[Endy]

Submit a suggestion and I'll tweak genetic memory or add a few new values that are instantanious. Beats having to rev your own fork every time a new drop comes out...

I've been using 550-554 myself, although it wouldn't make much difference really if the original locations were revised. It's relativly easy to code, just copy the values right at birth.

[EricL]

So, current genetic memory are mem locations 971-990.  The locations are copied one per cycle, in order, over the bot's first 20 cycles of life as long as the birth tie remains in place.  If the birth tie is cut (either intentionally or if the parent dies or similar) during the first 20 cycles, the copying stops.  Bot's need not do anything other than keep the birth tie in place for those 20 cycles for these locations to be copied.

Should I add new locations that are copied instantly or re-use these?  I could make the first five instant then have the next 15 take the first 15 cycles of life or add 5 new ones that are instant...

Voices in favor?

All opposed?

[Jez]

I have never been sure that birth ties are correct, the fact that we can prematurely destroy them using programming techniques but not a sysvars, making that option so infinitely unlikely to be evolved by evo bots, has never seemed the right approach IMO. (A marked distinction between combat and evo bots)

I would prefer something like transmission of body/nrg being the reason for the birth ties existance, although I understand that there are biological reasons for it being this way.

To my mind transferance of information to offspring is no different to the transfer of genetic information, I am not aware of biological comparisons for educating offspring in this way.

Do as you feel fit, you have my agreement for any changes.

[Numsgil]

Just have them do it instantly, it's just not a useful feature otherwise.

[EricL]

I don't know, I kind of like having some of them tied to an infant period.  I also seem to remember PY championing this.

Unless there are strong objections, I'll make the first 5 instant and the other 15 one per cycle.

[Endy]

Sounds good to me  

To my mind transferance of information to offspring is no different to the transfer of genetic information, I am not aware of biological comparisons for educating offspring in this way.

I see it as a sort of phenotype information transfer. It's largely above random mutations, yet at the same possible to change during a bot's lifetime.

[EricL]

FYI, this change will be in 2.42.9q.