A recent study has found the number of '+1's and '-1's in a codebase is around 15 times higher if you use 1 based indexing.
Also I wonder if it's highly dependent on whether the domain is largely algorithmic (e.g. video codecs) or business logic (e.g. supply chain).
And how much is simply the language being "misused", e.g. writing "<= n - 1" rather than "< n".
> e.g. writing "<= n - 1" rather than "< n"
Wouldn't such errors further increase the ratio?
But if you ran such an analysis across public GitHub repositories per-language and wrote the results up in a blog post, I'm sure HN would love to see it. Definitely front-page material.
> But if you ran such an analysis across public GitHub repositories per-language and wrote the results up in a blog post, I'm sure HN would love to see it.
I'll see what I can do. I do have some repos in sight that could be used for it.
But even then this won't be as straightforward as you say since different languages have different applications (eg cpp for games, julia for scientific computing). This would require writing the same code in both indexing patterns and then comparing them.