[Axiom-developer] [#12 radicalSolve fails to find all roots ?] (new)

[Bob McElrath]

anonymous address@hidden wrote:
> When Axiom cannot find solutions, it is (presumably) a PROOF that the other
> solutions are NOT solvable by radicals (using *real* numbers), or at least,
> there is no known proof that it is solvable at the time of implementation. 
> (That
> is why I am surprised at the above result for z^7=2).

Given Axiom's assumptions about input in this problem, why cannot I do
this:

\begin{axiom}
    z:Complex(Float)
    radicalSolve(z^7=1)
    z:Integer
    radicalSolve(z^7=1)
    z:Variable(Complex(Float))
    radicalSolve(z^7=1)
    z:Symbol(Complex(Float))
    radicalSolve(z^7=1)
\end{axoim}


...and get the appropriate answers?

Also this behind-the-scenes behavior where the answer depends on the
input type or assumptions is undesirable, and surprising to casual
users.  When algorithms must make assumptions about the type of a
Variable or Symbol, at the very least a message should be printed
indicating that the assumption was made.

An even better algorithm would print a message, then keep that
assumption for the remainder of the calculation...


--
forwarded from http://page.axiom-developer.org/zope/mathaction/address@hidden