camelCase for non-source-code files. I find camelCase faster to “parse” for some reason (probably just because I’ve spent thousands of hours reading and writing camelCase code). For programming, I usually just use whatever each language’s standard library uses, for consistency. I prefer camelCase though.