> The most obvious upper bound for K(X) is obviously the length N of X itself.
This seems false as you'd have to account for the header of the program. Otherwise how can you tell programs and texts apart?
This seems false as you'd have to account for the header of the program. Otherwise how can you tell programs and texts apart?
By opening them either with a text editor or a compiler/interpreter.
Homoiconic languages have no problem being edited and run.
Since this overhead is constant, it becomes negligible as the sequence length increases. Yet we should take it into account, e.g. by saying the upper bound is 'N + O(1)'.