Chemical CA

[Hunting]

They should bounce off each other as of now, because there are no bonds.

What different types of atoms should there be?

Only one type of them with no special property - let's have a proof-of-concept first. There are quite a few particle CAs already I think.

[Schiaparelliorbust]

Only one type of them with no special property - let's have a proof-of-concept first. There are quite a few particle CAs already I think.

Yes, that make sense. The rule table should be simple enough for me to be able to do it maybe.
Edit: What should happen when two atoms moving perpendicularly to each other? Or a 3 or more way collision?

[BokaBB]

Only one type of them with no special property - let's have a proof-of-concept first. There are quite a few particle CAs already I think.

Yes, that make sense. The rule table should be simple enough for me to be able to do it maybe.
Edit: What should happen when two atoms moving perpendicularly to each other? Or a 3 or more way collision?

I think that those who have more energy should affect others more and change direction at a slower pace.
Have a good day!

BokaBB

[Schiaparelliorbust]

I think that those who have more energy should affect others more and change direction at a slower pace.
Have a good day!

BokaBB

There is no energy yet. We are only making a collision CA. I already said that I think energy should only be for reactions and not for movement. I we want a somewhat wide temperature scale and we use your idea, we would have to define collisions for every single temperature and atom.

[Hunting]

I think that those who have more energy should affect others more and change direction at a slower pace.
Have a good day!

BokaBB

There is no energy yet. We are only making a collision CA. I already said that I think energy should only be for reactions and not for movement. I we want a somewhat wide temperature scale and we use your idea, we would have to define collisions for every single temperature and atom.

Continue. So what are the details? We are going to use 4 states for 4 directions, and when they collides... I can't think of a better idea compared to the gas CA. But that's a lot of states. If it is used, we'll probably run out of states before doing reactions.

[fluffykitty]

How about "any gas particles moving into an occupied spot bounce off, two gas particles moving into the same cell swap directions, three gas particles moving into the same cell allow the central one to continue and reflect the other two, and four gas particles moving into the same cell all reflect"? This can be implemented with one movement/2 steps, and it doesn't require tracking states for filled cells beyond "sense step/move step".

[Hunting]

How about "any gas particles moving into an occupied spot bounce off, two gas particles moving into the same cell swap directions, three gas particles moving into the same cell allow the central one to continue and reflect the other two, and four gas particles moving into the same cell all reflect"? This can be implemented with one movement/2 steps, and it doesn't require tracking states for filled cells beyond "sense step/move step".

Ha! That make sense. I was thinking of something along the lines of "two gas particles moving into the same cell both turns clockwise" as in snakes, but your idea is better.

You can have 1 movement/generation by using the Moore neighbourhood, I think.

[Schiaparelliorbust]

How about "any gas particles moving into an occupied spot bounce off, two gas particles moving into the same cell swap directions, three gas particles moving into the same cell allow the central one to continue and reflect the other two, and four gas particles moving into the same cell all reflect"? This can be implemented with one movement/2 steps, and it doesn't require tracking states for filled cells beyond "sense step/move step".

I think the hardest part is a whole dense cluster of cells colliding into each other. Is there any way we could prevent something like this from happening in the first place?

[Hunting]

How about "any gas particles moving into an occupied spot bounce off, two gas particles moving into the same cell swap directions, three gas particles moving into the same cell allow the central one to continue and reflect the other two, and four gas particles moving into the same cell all reflect"? This can be implemented with one movement/2 steps, and it doesn't require tracking states for filled cells beyond "sense step/move step".

I think the hardest part is a whole dense cluster of cells colliding into each other. Is there any way we could prevent something like this from happening in the first place?

I don't see any problem handling that. Example? (Post in the gas CA)

Or we could use the solution from NTAA. Never allow an empty cell to change into a gas if another gas is adjacent to the empty cell.

[Schiaparelliorbust]

I think the hardest part is a whole dense cluster of cells colliding into each other. Is there any way we could prevent something like this from happening in the first place?

I don't see any problem handling that. Example? (Post in the gas CA)

Or we could use the solution from NTAA. Never allow an empty cell to change into a gas if another gas is adjacent to the empty cell.

So do you think there should be a tail state?

[Hunting]

I think the hardest part is a whole dense cluster of cells colliding into each other. Is there any way we could prevent something like this from happening in the first place?

I don't see any problem handling that. Example? (Post in the gas CA)

Or we could use the solution from NTAA. Never allow an empty cell to change into a gas if another gas is adjacent to the empty cell.

So do you think there should be a tail state?

Obviously no, I agree with you here.

[Schiaparelliorbust]

Obviously no, I agree with you here.

Ok then, what happens when they try to move next to each other? Do they bounce off? If we use this system, I guess we have to find away to bond atoms with heat. Maybe we could imagine that there's a sort of electromagnetic force barrier around each atom and you need heat to overcome it. The barrier is simply the fact that they cannot touch each other.

[fluffykitty]

You can have 1 movement/generation by using the Moore neighbourhood, I think.

This does not work for the case with two particles pointing at each other with one cell in between, and most likely the cases with three or four particles moving into a single cell.

[Schiaparelliorbust]

You can have 1 movement/generation by using the Moore neighbourhood, I think.

This does not work for the case with two particles pointing at each other with one cell in between, and most likely the cases with three or four particles moving into a single cell.

What is your suggestion? I think a NTAA style movement could be useful. As I said earlier, If they're hot enough, they come close to each other and bond. Also, maybe if a cell is stuck, basically bumps into something but then can't go back, it should stop moving entirely, or maybe it can "remember" its direction and moves when it's freed.

[fluffykitty]

Split each effective tick into two CA steps, one for checking which movements are possible, and one for moving. This simulates a range-2 VN CA, which should suffice for a basic gas automaton. Zone-of-influence might be possible with the Moore neighborhood, but there might be impossible edge cases. With my ruleset, a trapped particle will repeatedly reverse direction until freed.

[Schiaparelliorbust]

Split each effective tick into two CA steps, one for checking which movements are possible, and one for moving. This simulates a range-2 VN CA, which should suffice for a basic gas automaton. Zone-of-influence might be possible with the Moore neighborhood, but there might be impossible edge cases. With my ruleset, a trapped particle will repeatedly reverse direction until freed.

Could you give an example please? I can't picture in my head how it's a range-2 VN CA. Is it okay for particles to be out of sync? Basically some particles will be moving while others are checking to see if they can move. What will happen if they interact?

[fluffykitty]

SlowLife in the Rule Table Repository is a rule that simulates Life at half speed by doing this kind of tick-splitting (however, nothing happens during one of the ticks, so the slowdown is pointless in this case). Out-of-sync particles are forbidden and would probably produce those awkward endlessly-growing blobs you sometimes see when using incompletely specified rule tables.

[Schiaparelliorbust]

SlowLife in the Rule Table Repository is a rule that simulates Life at half speed by doing this kind of tick-splitting (however, nothing happens during one of the ticks, so the slowdown is pointless in this case). Out-of-sync particles are forbidden and would probably produce those awkward endlessly-growing blobs you sometimes see when using incompletely specified rule tables.

Then to combat this should there an unchanging wall state where all the reactions happen inside and the empty space inside this wall would be a special cell which syncs cells to a certain phase?

[fluffykitty]

Patterns that start with all cells synchronized should never produce desynchronized cells.

[Schiaparelliorbust]

Patterns that start with all cells synchronized should never produce desynchronized cells.

Maybe not in a simple collision CA but if an atom bonds and the breaks off, it could very well be out of sync.

[fluffykitty]

As long as there are no transitions that violate the phase system (all sense cells become move cells and vice versa) no errors can appear. I don't see how bonding would affect that.

[Schiaparelliorbust]

As long as there are no transitions that violate the phase system (all sense cells become move cells and vice versa) no errors can appear. I don't see how bonding would affect that.

I guess you could sync bonding and breaking off to the searching and moving. Though we would still need to make sure there are no parity-changing transformations.

[fluffykitty]

Cells in the same phase but with different spatial parities shouldn't cause any problems, and avoiding phase issues won't be hard. Note that more phases will increase the state count, and so we should try to avoid having too many of those.

[Schiaparelliorbust]

Cells in the same phase but with different spatial parities shouldn't cause any problems, and avoiding phase issues won't be hard. Note that more phases will increase the state count, and so we should try to avoid having too many of those.

By spatial parity do you mean like a checkerboard pattern? I'm thinking of having 6 elements in the final CA. Is that too much?

[fluffykitty]

By spatial parity do you mean like a checkerboard pattern?

Yes. I was assuming that that was you meant.

I'm thinking of having 6 elements in the final CA. Is that too much?

It's probably a good idea to do tests with fewer elements before deciding how many elements the final CA will have.