Re: Simples
Software copyrights are evaluated partly on structure, treating variable names used in the same way as equivalent. I recall looking at the infamous "rangecheck" code for a bit under a minute and concluding that there were, in practical terms, exactly two ways to code it, unique up to choice of variable names. This means, of course, that if three programmers of modest skills addressed the problem rangecheck solves, it is a near certainty that at least one of them was an infringer.
Automating checks for such things on github seems a singularly bad idea.