All symbols in macro-language are stored as a string; however there are supported several math operations as well. Macro-language allows to work both with real-numbers (with double precision) and integer-numbers (unlimited width from version PE 5.02).
For integer operations, string is converted to integer number. If the string represents real number with non-zero fractional part, error is reported.
For operations, that are supported with both real and integer operands, the real numbers are preferred. If both operands are integers, operation is executed above integer type.