While working on a parser involving LLVM IR, I started wondering whether there is a fundamental distinction between data and logic. Both can be represented as bits and both seem to contain information. In languages such as C++, functions can be stored as members, while functional programming techniques can represent state through constructs like monads, making the boundary feel even less clear. I understand that a CPU interprets some bit patterns as instructions and others as data, but I'm looking for a deeper explanation. What does "information" mean in this context, and are data and logic fundamentally different or merely interpreted differently?
3 Answers
Data and logic can use the same underlying representation. The important difference is how an interpreter, compiler, or processor assigns meaning to that representation. A sequence of bits might be treated as a number, an image, a function, or machine instructions depending on the rules applied to it. In that sense, data is information considered by some process, while logic is information that describes transformations or relationships.
At the hardware level, executable code can sometimes be reused as ordinary data. For example, old cartridge programmers could display the bytes of a program as a bitmap. Nothing about the bits themselves permanently labels them as “code” or “data”; their role comes from the context in which they are read. The distinction is mainly semantic and operational rather than physical.
Some programming languages make the equivalence especially explicit. In Lisp, programs are represented as ordinary data structures, and those structures can then be evaluated as programs. The same object can be inspected as data or executed as logic, depending on what operation is applied to it.
Lua also has a similar idea in a broader sense: functions are values that can be stored, passed around, and returned. It is not quite the same homoiconic design as Lisp, but it still demonstrates that executable behavior can be handled as data.