#2943 complex rootes

[Friedrich Laher]

wxMaxima version: 13.4.0
Maxima version : 5.31.3
Maxima build date: 2013-11-22 16:33:24
Host type: x86_64-suse-linux-gnu
Lisp imple. type: CLISP
Lisp implementation version: 2.49 (2010-07-07) (built on cloud132 [127.0.0.1])

---

look to attachment. If done without inputline %i8 it works,
but i need to evaluate "by hand" that for X > 0 and Y > 0 and Z > 0
the solutions are complex

[Leo Butler]

From 2944 (duplicate report deleted by Laher):

look at attachment:
without input X : 1+abs(x) $ Y: 1+abs(y) $ Z: 1+abs(z) $
it works, but i need to evaluate manually that for
X > 1 and Y > 1 and Z > 1
the roots are complex .
Is the empty solution list %o11 just a sign that Maxima has "given up" ?

[Leo Butler]

Follow-up from Laher (originally in 2945, deleted as a duplicate):

the result is different for different sequzences of equations in the equation list:
for solve([eA,eB,eC],[p,q,r]) the simplest structured is q,
for solve([eB,eA,eC],[p,q,r]) the simplest structured is r,

and the resultlist is empty for [eC,eA,eB] and [eC,eB,eA]

that makes it not very trustworthy

[Stavros Macrakis]

Please send examples using a text format. Sending a png screenshot means that we have to re-type your tests.

[Friedrich Laher]

VA : p*(1-r) $ VB : q*(1-p) $ VC : r*(1-q) $
VI : 1-VA-VB-VC $
eA : VA = X*VI ; eB : VB = Y*VI ; eC : VC = Z*VI;
pqr : solve([eA,eB,eC],[p,q,r]);

and now change the sequence of equations in solve