Neato. Yeah, range checking would be nice. I kno...

2008-07-11T11:40:00.000-04:00

Neato. Yeah, range checking would be nice. I know I've seen it done with attributes but it used a special compiler and cost $$ to use I think.

I like that C# allows me to be lazy by using implicit operators ;) you can just use it as if it were a regular integer without any problem (but you can only update the reference with another UserId so it still gaurentees the value makes sense)

This was one of the safety features of super-safe ...

2008-07-10T16:39:00.000-04:00

This was one of the safety features of super-safe Ada. In Ada you can simply declare a new primitive type:
type UserID is new Integer;

Better yet you can do ranges to prevent overflows and negative values:
type UserID is range 0 .. 1000;

This does not work like typedef in C, in Ada new types are distinct and not interchangable, ie the compiler will complain.

To get this behaviour in C, you should not use typedef, but instead create a struct:
typedef struct {
int id;
} UserID;

The disadvantage is that you have to type user_id.id to get the actual value, but there is no runtime overhead for passing by value because UserID is the same size as int.

I am not sure if the C# compiler/runtime has any overhead for structs but I am willing to bet GCC (and any decent compiler) will compile both int and UserID to the same code.

Comments on The Macadamian Files: Compile Time Semantic Checking

Neato. Yeah, range checking would be nice. I kno...

This was one of the safety features of super-safe ...