Specify Logic Resource.

[cqc-alec]

No description provided.

[cqc-alec]

In two minds about including this. It's rarely used as a logical operation and can be constructed from the others if needed ... then again why not ... then again again why not include nand as well... Any thoughts?

[cqc-alec]

Couple of changes I'd like to make here on reflection:

• remove xor;
• allow and and or to take any number of inputs.
  (Then true is just and with 0 inputs and false is just or with 0 inputs and we just have three basic op types.)

[ss2165]

why is not not N-ary?

[ss2165]

also where do you see operations like xor being defined? It's too primitive to be defined as a library function no?

[cqc-alec]

Presumably that would mean applying N nots in parallel -- so N inputs and N outputs? Yes we could define this, but it seems less useful. N-ary and and or save us from chaining a bunch of them together in a tree, taking advantage of associativity and commutativity, but I don't think there's a particularly compelling reason to combine parallel nots into one.

[cqc-alec]

also where do you see operations like xor being defined? It's too primitive to be defined as a library function no?

Well, I wonder about this. As far as I can see xor is very rarely used as a logical operation (I can't think of any examples in tket), and on occasions when it is it can be implemented as something like "(A or B) and not (A and B)". It is much more often used in bit-twiddling or as addition mod 2; in both these cases I imagine one would use int<1> from the arithmetic resource (where addition is xor).

But I will add it if you think it's useful here.

[cqc-alec]

I don't think any of C, C++ or Python have a logical xor operation (analogous to && or ||, or and and or) (though they all have bitwise xor).

[ss2165]

that makes sense - happy with current state

[ss2165]

Comments already added as single comments