I was recently reading a newspaper in the train and there was a little math riddle, I thought “how funny, that’s gonna be easy, let’s do it” and yet…
The problem goes as follow : in a barn, there is a 1 meter cubic box against a wall and a 4 meter ladder is leaning against the wall, touching the box at its corner. Here is a picture :
So, the big triangle has a hypotenuse $FE$ of $4$, the square $ABDC$ has sides of length $1$ and is basically “insquared” at the right angle, i.e. $D\in \overline{FE}$. The question is “what is the length of the biggest cathetus”, here $AF$.
So far, no problem.
Now here are my solutions:
 By Thales’ intercept theorem, $\frac{FB}{FA}=\frac{BD}{AE}$, by hypothesis, $FB=FA1$ and $BD=1$. Now by Pythagoras, $FA^2+AE^2=FE^2$; by hypothesis, $FE=4$, so we end up with a system of equations, letting $h=FA, d=AE$:
Which solves (removing 3 nonrelevant solutions) into $d \cong 1.3622$ and $h \cong 3.76091$.

Now, if I consider the “function” of the line : $f(x)=\frac{h}{d}x+h$, I know that $f(1)=1$ and I end up with Pythagoras with the system :
$$ \begin{aligned} \frac{h}{d}+h&=1 \\[2ex] h^2+d^2&=4^2 \end{aligned} $$
It solves again into the same, again removing 3 nonrelevant solutions
Okay, this means that using Pythagoras is no good since it ends up giving a quartic equation (4 answers, of which 3 are “nonrelevant”).
 Now if I consider the length of the arc $f(x)$ between $0$ and $d$ it has to be $4$ and again $f(1)=1$ I end up with the system:
Which solves again into the same answers, but this time removing only 2 nonrelevant solutions (i.e. it gives a cubic equation instead of a quartic).
I tried also using the areas and the smaller trangles $FAD$ and $AED$ for example :
$$\frac{h \cdot d}{2} = \frac{h\cdot 1}{2}+\frac{d\cdot 1}{2}$$Yet I wasn’t able to get to any “hand solvable” solution : if I were able to bring it down to some quadratic equation, that would be nice, since it is a common assumption, here, that everybody has seen the “general formula for solving quadratic equations” in school and so would be able to solve this, I may then see how it is seen as a funny riddle in the newspaper.
My best trial, with “just” a cubic equation, is way too complicated for the normal readers of this newspaper, so it’s bugging me.
So, I asked around! And somebody proposed me the following solution, using a clever variable change:
Let $FB=x$. By similarity of triangles we then have $CE=1/x$. Pythagoras thus gives
$$ 4=\sqrt{x^2+1}+\sqrt{1+(1/x)^2}=\sqrt{x^2+1}(1+\frac1x). $$Squaring this gives us
$$ 16=(x^2+1)(1+\frac2x+\frac1{x^2}), $$but he preferred to move one factor $x$ from the former factor on the right hand side to the latter, so
$$ 16=(x+\frac1x)(x+2+\frac1x). $$Almost there: write $u=x+1/x$. We can solve $u$ from the quadratic
$$ 16=u(u+2), $$and then solve for $x$ from the equation
$$ x+\frac1x=u. $$