Both hardware and software emulators have been built by hobbyists using only the original specs. They assemble and run the source listings provided by MIT. You can quibble about a lot of Apollo technology, but it's very difficult to argue the AGC and its code were a whitewash. These days it is within many people's grasp to verify the original work. We can't take the surviving LMs out for a spin, but we can run the original computer code.
Trying to impose modern software practice on engineers of 1965 is just silly. Although software-specific design methods were in their infancy, the software was considered essentially part of the hardware and electronics design and was actual configuration-controlled along with that -- to a much more rigorous degree than most software today is controlled.
Keep in mind that software development cycles were glacially longer in 1965 than they are today, and far more rigorous and introspective. There was no such thing as "agile methods" and no urge to crank out 50,000 lines of PHP over the weekend for the next great web site. Software was serious stuff, seriously engineered.
|