If integers are used, then store as integers. If doubles are used, then store as doubles. This way precision issues are prevented.
