prel: will give you exactly the same whole number value as the existing prel coords.
pos: gives you the actual position as a real number.
I primarily made the distinction to reduce comparison confusion. But I'm also sure each will prove useful in different scenarios.
single: mainly for comparing numbers of different precision. variables are stored as double precision real (floating point) numbers, and in some cases (i've forgotten what they are) equals [=] comparisons were failing because the source of one number was from a single precision number. so the comparison was fixed by converting the double precision number to single precision beforehand. hopefully it won't be needed often.