The use case is as follows. Say I am exploring some idea and have several versions of it in work in parallel. So I open a new file and call it for example 'Version04". Then I introduce the same class (at top level) as shown above as a routine and start writing functions. By the way: I do not have a collision with function names, as long as I mark them private.
So, I could definetely introduce a class “Version04” and have the rest within that class. But it would over time lead me away from using top level objects in general. I think the ‘private’ marking should be more effective in keeping anything marked as private just that: private and never visible or effective in any way outside the file.
I understand the behaviour now a bit better and there might be some underlying architectural reason, why it can’t be different. I just wonder why it is implemented that way and would like to understand why.