I'm seeing production crashes where the Location class is trying to allocate 320kb, presumably for either the latitude or the longitude. Is there a reason why unlimited precision is necessary? Wouldn't double be good enough?
I agree that double would be good enough. In general: The accuracy of astronomical calculations cannot be increased by choosing BigDecimal but by choosing another better astronomical algorithm.
I'm seeing production crashes where the Location class is trying to allocate 320kb, presumably for either the latitude or the longitude. Is there a reason why unlimited precision is necessary? Wouldn't double be good enough?