Refactor file interpreter to use binary files

At the moment, the files to be parsed are actually text files with one operation per line.

An appropriate refactor to use a binary representation would be great.

The issue provides no file or test names. Start by locating the interpreter's current text-file parsing entry point and tracing how parsed operations reach the virtual machine. Define the binary representation and its compatibility expectations before changing code; done means the interpreter can parse and execute the binary files consistently.

Written by the indexing model from the issue text.