Minutes per Kilometer

Two things. Assuming that you're trying to insert that code into the definition of toSpeed...

1. kmPerHour is a Float, and % on a Float does not work in MonkeyC
2. kmPerHour is a unit of speed, but you're trying to display it as time. You have kmPerHour, but what you really want is minutes per kilometer.

Assuming that s is meters per second, you probably want something like this:

var secondsPerMeter = 1.0 / metersPerSecond;
var secondsPerKilometer = 1000 * secondsPerMeter;

this can be written more simply in a single step as:

var secondsPerKilometer = 1000.0 / metersPerSecond;

Once you have a unit of time, you can break it down into minutes/seconds.

// round to nearest whole second
secondsPerKilometer = (secondsPerKilometer + 0.5).toNumber();

// now that we have a Number, we can use %
var minutes = (secondsPerKilometer / 60);
var seconds = (secondsPerKilometer % 60);

var formattedTime = Lang.format("$1$:$2$", [ minutes, seconds.format("%02u") ]);

Floating point numbers (both Lang.Float and Lang.Double) have advantages and disadvantages when compared to fixed point numbers (Lang.Number and Lang.Long). It is probably not a bad idea to go find a book on the topic, but I think I can give a brief summary:

1. Floating point arithmetic is typically costlier in terms of CPU cycles than fixed point. For purposes of this conversation, mixed arithmetic (where one operand is a floating point and the other is fixed point) can be considered a floating point operation.
2. Floating point number systems have more precision than fixed point when the values are relatively small. A 32-bit fixed point number cannot represent values between 0 and 1 at all, but a 32-bit float can represent something like 1,056,964,610 unique values in the range [0, 1].
3. Floating point number systems have less precision than fixed point when the values are relatively large. A 32-bit float using IEEE754 format is able to represent whole numbers from -16,777,216 to 16,777,216 exactly. For values outside that range, you might want to use something that has more precision (like a double) or used fixed point values.
4. There is almost always error with floating point, and sometimes it is important to know about the representation and how to determine the error.
5. A 32-bit floating point value is no costlier to store than a 32-bit fixed point value; both take exactly 32 bits. Same for a 64-bit float vs 64-bit fixed. In MonkeyC a 32-bit value (Float or Number) do not require a heap allocation, but a 64-bit value (Double or Long) do.