The problem was written so that it can be solved using either integer or floating-point arithmetic. As an easy-ish problem, truncation was chosen for the final output rather than rounding, because truncation is the default for both integer division and float-to-int casts. The input data was chosen so that there should be no issues with roundoff error. Eric's "official" solution uses doubles (linear.cpp), but Ron wrote an alternate solution using floats (linear_2.c), and Andy then modified Eric's solution to use only ints (linear_3.cpp). All three solutions agree.