Algebra through geometry
by t0rajir0u, Apr 19, 2007, 6:33 AM
Problem: Given reals
such that



Find
.
Solution: Well, this is a pretty intimidating system. You might be tempted to try a difference-of-cubes sort of thing, which does lead to a solution, but I took a different path motivated by a problem in Engel and by the observation that
:
Suppose we have triangle
with
the Fermat point; this means
. Let
. Then by Law of Cosines on
we have



We've produced a right triangle! Now, how do we go about calculating
?
Well, consider the area of
. Since the triangle is right we know that this is just
. But recall that the area of a triangle with sides
and angle
between them is
. This tells us that
$\begin{eqnarray*}[ ABC ] &=& \frac{ \sqrt{6}}{2}\\ &=& [AFB]+[BFC]+[CFA] \\ &=& \frac{\sin 120^{\circ}}{2}\left( xy+yz+zx \right) \\ \end{eqnarray*}$
From which we readily deduce that
.
If you're a computational masochist you might realize that the rightness of the triangle is not necessary, and that we could've used Heron's if the right-hand constants had been something else.
Practice Problem: Find a straightedge-and-compass construction for a pentagon.




Find

Solution: Well, this is a pretty intimidating system. You might be tempted to try a difference-of-cubes sort of thing, which does lead to a solution, but I took a different path motivated by a problem in Engel and by the observation that

Suppose we have triangle








We've produced a right triangle! Now, how do we go about calculating

Well, consider the area of





$\begin{eqnarray*}[ ABC ] &=& \frac{ \sqrt{6}}{2}\\ &=& [AFB]+[BFC]+[CFA] \\ &=& \frac{\sin 120^{\circ}}{2}\left( xy+yz+zx \right) \\ \end{eqnarray*}$
From which we readily deduce that

If you're a computational masochist you might realize that the rightness of the triangle is not necessary, and that we could've used Heron's if the right-hand constants had been something else.
Practice Problem: Find a straightedge-and-compass construction for a pentagon.