C :: Application That Uses Two Different Timing Techniques

I am looking at a same application in c that uses two different timing techniques. One involves the nanosleep() function which suspends thread for given nanoseconds. The other one uses a combination of gettimeofday() which returns number of microseconds since epoch as a start value. Then it uses var counter of for loop as max value. And it uses ellapsed time for each call of gettimeofday in microseconds in a do while loop to increment diff, until diff is not less than var. Why would someone choose to use this timing techinque be used instead of nanosleep?

Code:
unsigned long getTime() {
struct timeval detail_time;
gettimeofday(&detail_time,NULL);
return (detail_time.tv_usec);

[Code] .....