common-lisp Comparison operators on numeric values


Numeric values can compared with = and the other numeric comparison operators (/=, <, <=, >, >=) that ignore the difference in the physical representation of the different types of numbers, and perform the comparison of the corresponding mathematical values. For instance:

(= 42 42)
T ;; => both number have the sme numeric type and the same value
(= 1 1.0 1d0)
T ;; => all the tree values represent the number 1, while for instance (eql 1 1d0) => NIL
  ;;    since it returns true only if the operands have the same numeric type
(= 0.0 -0.0)
T ;; => again, the value is the same, while (eql 0.0 -0.0) => NIL
(= 3.0 #c(3.0 0.0))
T ;; => a complex number with 0 imaginary part is equal to a real number
(= 0.33333333 11184811/33554432)
T ;; => since a float number is passed to RATIONAL before comparing it to another number
  ;; => and (RATIONAL 0.33333333) => 11184811/33554432 in 32-bit IEEE floats architectures
(= 0.33333333 0.33333334)
T ;; => since the result of RATIONAL on both numbers is equal in 32-bit IEEE floats architectures
(= 0.33333333d0 0.33333334d0)
NIL ;; => since the RATIONAL of the two numbers in double precision is different

From these examples, we can conclude that = is the operator that should normally be used to perform comparison between numeric values, unless we want to be strict on the fact that two numeric values are equal only if they have also the same numeric type, in which case EQL should be used.