Program re-write!

"In the first few months of life, Frida is realising she can control her own body. At around 4-5 months she realises thats all she can control directly, specifically that she can't control her mother. Then she realises she is herself and not a part of her mother." Vera Vasas said that.

The program is a mess. I can't understand it. It will need to be re-written. I didn't get enough sleep last night. The issue seems to be the storage of a dictionary that links atom parameter vectors (numbers) to names in ALMemory.

1. Each atom has a unique ID
2. Each time an atom is sampled in the loop it writes to an ALMemory structure with a key that is the ID of the atom, and data that is a vector, the first value of the vector being whether that atom is active or not. One atom ONLY writes to a single place in ALMemory keyed by the ID of that atom.
3. Atoms point to each other using their unique IDs
4. Sensory inputs to atoms have totally separate numbers.
5. Some atoms generate motor outputs internally.
6. Three kinds of atom are first initialised:

a. A function atom that takes an input sensory state and is not activated by any other atom, and writes a function of the input to its location in ALMemory.

b. A SHC atom that reads that location in memory and tries to minimise the value of that location in memory by writing a set of parameters to ALMemory.

c. A motor atom that reads the ALMemory location written to by atom b, and moves a set of motors.

OK, its not so bad. 21st March 2013, and I've massively simplified the previous code to get this..

https://github.com/ctf20/DarwinianNeurodynamics.git

See LivingMachines 2013 for the first draft 1. Now to add replication with mutation to the atoms within a single molecule, that has a functional and meaningful effect.

Next step. How should replicants be functional? Consider the 3 atom molecule above.

a. If atom A replicates then there are many possible sensory function atoms feeding sensory states into the SHC atom. How should the SHC atom use these states? 1. It could choose one of them, 2. sensory atoms feeding into the SHC atom could compete for being able to send the message to the SHC atom. This could be achieved by using mutual inhibition between competing atoms such that two equivalent atoms compete for being activated if sent signals from the same inputs. 3. In a sensory atom, only one will ever be activated at a time, this will be the least used one. So in a way this problem is solved in this case.

b.