John Cremona describes an use of the algebraic QQbar domain :
Then I test
a=sqrt(2)-sqrt(3)
b=sqrt(3)-sqrt(2)
QQbar(a).minpoly() ; QQbar(b).minpoly() # seems right. The same even
polynom.
But the test and the numerical values are True. I get +0.31 in both cases.
QQbar(a)==QQbar(b)
This me
Hello John
thank you for this. I tried the same thing on mathematica,
which managed to simplify 'c' back to 'a'.
I don't quite understand the culture of sage-support yet.
Is commenting on mathematica's ability to do a particular
task a useful thing to say? Or does it just annoy everyone?
best
With d=c-a, not even d.simplify_radical() gives 0.
Simplifying "nested radicals" is a notoriously hard problem in
symbolic computer algebra. As this example shows (unless there are
other tricks to try which I do not know about), Sage's symbolic system
is not up to examples like this.
As an algeb
