Are there problems with inductance? radiation? that I haven't considered?
Good choice of controller, Sam
The short answer: Ideally, if not practically, the battery cables need to be as short as possible, with the positive and negative adjacent, or even twisted together, over the entire run. The motor cables only need to be kept short to reduce I²R losses, not because the inductance is harmful (indeed, as major notes, it's helpful) or, even, because the noise emissions are particularly harmful. If you want more of an explanation, keep reading...
I seem to write a post on this every few months, but I'll grant that the topic it isn't exactly easy to search for. Anyway, there are two types of electrical noise - that from a changing voltage, called "E-field", and that from a changing current, called "H-field". Radio transmitters like cell phones, FM stations, etc., primarily emit E-field radiation while induction heaters primarily emit H-field. Both types of radiation can cause noise problems in electronics, but it just so happens that it is nearly impossible to adequately shield against the H-field (magnetic) but almost trivial to shield against the E-field.
A motor controller draws current from the battery in rectangular pulses, but the input capacitor (and batteries) integrate the voltage waveform into a triangular ripple. This still means that there are rapidly changing currents on the battery cables so they will primarily radiate H-field noise. Shielding against this requires lots of magnetic material to concentrate and divert the field lines - ie, lots of iron, mu metal, etc. This is clearly not practical, but you can stop much of the fields from being emitted in the first place by twisting the source and return cables together. One of our early beta testers that is active on this forum, dimitri, had problems with his PakTrakR which were mostly solved by simply twisting the battery cables in the engine compartment (not over the entire run, which would be ideal, but hey - a lot easier to just do so in the last few feet, eh?).
Conversely, a motor controller delivers pulses of voltage to the motor (except when at 100% duty cycle, that is). If the motor were a pure resistance then the current flowing through it would be rectangular, too, and those cables would emit both H-field and E-field noise. However, motor's have significant inductance which integrates the current into a triangular ripple. Thus, the motor cables only have rapidly changing voltages on them and primarily emit E-field noise as a result. This is where a metal shield is effective, as in coaxial cable, though automotive electronics are usually well-protected against E-field noise because they were designed to coexist with spark ignitions systems.