pg. 172

[alextzik]

The example mentions "The contour lines of f are lines of constant f . Thus, if a contour line of f is tangent to h, then the directional derivative of h at that point, along the direction of the contour h(x) = 0, must be zero". Shouldn't it be that the directional derivative of f at that point, along the direction of the contour h(x) = 0, must be zero?

[tawheeler]

You are correct! Good catch.

[tawheeler]

Looks like this is already fixed:

If the point $x^*$ optimizes $f$ along $h$, then its directional derivative at $x^*$ along $h$ must be zero.