Powers simplification with mathematical correctness

[Claudio L.]

(01-29-2019 07:57 PM)Hlib Wrote:  
Maybe like that?

In your example the expression is not being simplified, √(x^2) is left like that and passed to the solve function, so I guess your answer is don't simplify it? It's a valid option, and the one that will take me the least effort.

(01-29-2019 10:12 PM)ttw Wrote:  Much depends on whether your numbers are integers or floating point. With integers, most simplifications work with maybe one or two branch points. With floating point (AKA reals) there are infinitely many branches.

I wouldn't try to simplify an expression with X^2.37 for example, it is best if it stays that way.

I just tried it in Wolfram Alpha, and it's a bit confusing but consistent I guess:

(X^2)^(1/2) is NOT simplified, it offers ABS(X) as an alternate form but it actually keeps it as √(x^2) (so as to not lose the other roots).

However, (X^(1/2))^2 is simplified to X

I'll have to do the same, wouldn't be consistent otherwise with ->NUM.