If δ(R) = 0 we have the traditional polynomial ring R[y].
By the distributive law, y*(a+b) = ya+yb. Therefore δ(a+b) = δ(a)+δ(b). In other words, δ is a group homomorphism that respects addition in R.
By associativity, (y*a)*b = y*(a*b). Therefore δ(ab) = δ(a)*b+a*δ(b).
Conversely, assume δ is a function from R into R with these two properties. Build the aura polynomials as described above, and prove the result is a ring. The properties of δ make multiplication associative and distributive, thus producing a ring. I'll leave the details to you.
A polynomial is in normal form if each term is a coefficient in R, times a power of x, times a power of y. When two such terms are multiplied, normalize the result by pulling powers of y past powers of x. Thus xy2 times 3xy = 3x2y3 + 6x2y2.
Imagine multiplying y by p(x), where p is a polynomial in R[x]. By the distributive law, we can approach this term by term. Start with yxn and replace yx with xy+1 n times. The result is xny + nxn-1. Therefore yp = py + p′, where p′ is the formal derivative of p with respect to x.
View this as an aura polynomial in R[x][y], where δ implements differentiation. If p and q are two polynomials in R[x], δ(p+q) = δ(p) + δ(q), and δ(pq) = δ(p)q + pδ(q). These are the standard rules for differentiating a sum and product. Since δ has the required properties, the result is a valid ring. The weyl algebra is indeed a ring.
Note that R is a subring that commutes with x and y. If R is commutative, the weyl algebra meets the technical definition of an R algebra.
Let H be an ideal in the weyl algebra. Let t be a term of a polynomial in H. Note that yt-ty = t∂x, and xt-tx = t∂y. The ideal H is closed under partial differentiation.
Notice the chain of descending left ideals generated by the powers of y. If the simple ring were left artinian it would be the n by n matrices over a division ring. Obviously this is not the case. This is an example of a noetherian simple ring that is not artinian.
If R has characteristic p, multiply by xp on the right. This differentiates, with respect to y, p times. At some point a multiple of p is introduced as a coefficient, and the result is 0. Therefore xp commutes with everything, and generates a proper ideal. The same holds for yp. This is valid even if R is simple, e.g. Zp.
Note that x is not invertible, hence the result is not a division ring.