Ditch column name requirements in function parameters?

[mhinsch]

Definitely easier for the user and no issue on the C++ side. OTOH it makes it easier to mess up.
IMO this is mainly a question of which behaviour would be expected by most people.
@thibautjombart what's your opinion on this?

[thibautjombart]

I think it doesn't matter to some extent, as if we use epicontacts objects as inputs, then the naming thing is going to happen internally in the function passing the epicontacts to C++, which we write. I suspect the strongest constraint lies in how you want to design things in the standalone C++ library?

[mhinsch]

I have to say, the names requirement gets pretty old on the R side pretty quickly. Let's just drop it.