My solution was (given a precision factor in C):

B -> X:For(Z,0,C:For(Y,3,A:AnsX:End:X+(B/Ans-X)/A -> X:End:X

I don't normally use the solve( function.

]]>`Prompt A,B solve(B-prod(X+0rand(A)),X,1→X`

Please edit your first post rather than adding information in a new post.

]]>(1) Find A xroot B using only +,-,*,/,().

(2) Optimize it using logarithms, if possible. ]]>

My solution was

`Prompt A,B solve(ln(B)-Aln(X),X,1→X`

which is untested and may not work.

I haven't golfed a solution without log( yet.

How much precision is required? What if the solution only gives, say, 10 sigfigs?

]]>What is your 20-byte solution, by the way? The algorithm should be able to calculate up to A=100 and B=1e12. ]]>

I think I have 20 bytes.

]]>Given A and B, calculate A xroot B without using any root symbols or exponent symbols.

A and B are integers, but bonus points if you can make it work with noninteger indexes and radicands. ]]>