AMDG interval_concept.hpp:75 It would make more sense to me to make it a precondition of construct that low_value <= high_value. I would like to see a specialization of interval_traits for boost::interval. interval_concept.hpp:190-191 I don't think this is correct. Consider calling flip with an interval [1,2] and an axis 1. I would expect the result to be [0,1]. flip as written will give [1-2,1-1]=[-1,0]. The correct formula is newLow = 2*axis - high(interval); newHigh = 2*axis - low(interval); interval_concept.hpp:205-206,219-220,231-232,246-247, maybe more Please use static_cast, instead of a C-style cast. interval_concept.hpp:247 Is it necessary to set high in terms of low instead of adding displacement to high? What's the difference between convolve and move? interval_concept.hpp:366 I'm inclined to doubt whether this is any faster than if(position < low(interval)) return low(interval) - position; else if(position <= high(interval)) return 0; else return(position - high(interval)); which is easier to understand. interval_concept.hpp:396 is the use of the bitwise & operator instead of the logical operator intentional? I didn't see get_half and join_with in the documentation. In Christ, Steven Watanabe