I'm reading through a proof of torsion's definition:
$\frac{dB}{ds} = \frac{d}{ds}(T \times N) = \frac{dT}{ds} \times N + T \times\frac{dN}{ds} $
By definition, $N = \frac{dT}{\kappa ds}$ so it is parallel to $dT$ and the cross product is zero. My confusion starts here. It seems that we could also determine $N$ and $\frac{dT}{ds}$ are parallel by the fact that $T$ is of constant magnitude, therefore its tangent vector is always orthogonal to it, which is again a scalar multiple of $N$. My issue is that by this logic it seems to me that the second part of the expanded equation above can be held to similar reasoning:
$T$ is of constant unit magnitude by definition, so $\frac{dT}{ds}$ is orthogonal to $T$ for all $t$. We can unitize $\frac{dT}{ds}$ to obtain $N$, so $N$ is also of constant unit magnitude, therefore the tangent vector of $N$, $\frac{dN}{ds}$ is orthogonal to $N$ for all $t$, meaning that it must be parallel to $T$.
This by all other indicators appears to be incorrect; it would break the proof, but I don't understand where I am going wrong.
As Spock once said, you’re exhibiting two-dimensional thinking: $N\perp T$ and $N'\perp N$ doesn’t imply that $T$ and $N'$ are parallel. Consider the standard coordinate axes: the $y$-axis is orthogonal to the $x$-axis and the $z$-axis is orthogonal to the $y$-axis, but the $z$-axis certainly isn’t parallel to the $x$-axis.