Your main arguments against the definition $a\times b := |a||b|\sin(\theta)\hat n$ seem to be (1) that to use this definition, we have to be able to find the right-handed normal vector $\hat n$ first. I don't see that as much of a problem. There are several known methods of finding such a normal vector. And (2) that you seem to believe the cross product is only useful for finding the normal vector. I disagree. A non-exhaustive list of some of the other useful applications of the cross product includes
- finding the area of the parallelogram with sides as the vectors.
- describing various physical quantities including torque and the magnetic field.
- together with the dot product in the form of the scalar triple product, finding the volume of a parallelopiped.
- describing (/modelling) the Lie algebra $\mathfrak{so}(3)$.
Now here are some arguments for that particular definition of the cross product (BTW, there are other definitions):
- It's entirely coordinate-independent. So, when using this definition, not only will not need to worry about having to change our coordinates at some point in the middle of working through a problem, but it makes the notation far more compact than say the formulas in user247327's and Robert Israel's posts.
- It's a very geometric definition. It's of the form (scalar) times (unit vector) so we know immediately upon looking at that formula that the length of the cross product is $|a||b|\sin(\theta)$ and the direction that it points in is the right handed normal direction. Again, compare this to the formulas given in user247327's and Robert Israel's posts. It is not at all clear just looking at those definitions what the length or direction of $a\times b$ is.
- It suggests a connection with the dot product as defined by $a\cdot b :=|a||b|\cos(\theta)$. For instance, Lagrange's identity becomes very clear: $$(a\cdot b)^2 + |a\times b|^2 = |a|^2|b|^2\cos^2(\theta) + |a|^2|b|^2\sin^2(\theta) = |a|^2|b|^2\big(\cos^2(\theta) + \sin^2(\theta)\big) = |a|^2|b|^2 \\ \bbox[5px,border:2px solid red]
{(a\cdot b)^2 + |a\times b|^2 = |a|^2|b|^2}$$
Postscript:
Note that the definition of an object is not always going to be the best way to calculate it in practice. Now that we have this definition, we can show that it implies the formulas in user247327's and Robert Israel's posts under mild assumptions. So if we want to put some numbers to it, there's no reason not to use those formulas instead.
Compare this to the Riemann integral. The definition of the Riemann integral is pretty complicated, right? But luckily we rarely have to use that definition to actually evaluate an integral -- we use the fundamental theorem of calculus.
It's the same thing for the cross product. The definition might be a little more difficult to use, but for the reasons above (and some more complicated reasons I didn't get into) it's a pretty good definition for proving theorems and other things we might need to do with a definition. But it's completely fine to use some other method to actually calculate it.