Lecture 2
---------

Where do processes come from?

One idea: a running process creates a new process by specifying an executable to run in the new process.

That's how Windows does it.  That isn't how UNIX-like systems work.

Orthogonal design in APIs
 - only one way of doing a specific operation
 - complex operations are built out of simple ones

Consider the use cases for making a new process
 - run a new program (binary)
 - duplicating an existing process


Difference?

Process A copied to Process B

vs

Process A executing a binary to create B (which is the one used to make A)

In the first case, B's variables have the same state as A's variables

Process duplication
* useful operation for concurrency
* Simple interface
* can be combined with other operations to run other programs

To duplicate a process: fork
To run a binary into the existing process: execve