I've tried reading unannotated C and C++ and it's really hard. The problem with reading code is you're anticipating all kinds of input that doesn't necessarily

I've tried reading unannotated C and C++ and it's really hard. The problem with reading code is you're anticipating all kinds of input that doesn't necessarily exist at the time of review, and the code communicates not with a person but a code interpreter that reads in a hyper-precise and idiosyncratic way. Annotations don't work with legal code because there is a real risk the interpreter might use the escaped comments unexpectedly as production code(!). Imagine your C and C++ comments could generate production errors. You probably wouldn't use them much. There's an argument for giving people a sheet of hypothetical inputs and projected code outputs when we're talking contracts. Biz gives you the 'code' and then in a separate sheet say: hypothetically, a shady backwater adtech company asks for your mother's maiden name; are we allowed to sell it without further consent? And then they give the code output: "Yes, we can do whatever we want." And they go through a list of these hypothetical. The counterargument is such a sheet could be even more mind-numbing and unwieldy than the original code if not written in an entirely new spirit. And FWIW, my personal view is that these 70-page contracts of adhesion are a travesty, an evolutionary step, and a weak fiction at this point. The industry does not have a right to force 3 months of mandatory bedtime reading for their userbase every few months, the whole thing is broken and we're all just waiting for courts to coalesce around a better model.