Viktor Söderqvist

**My interpretation:** The "maybe" is referring to the improperness. Thus, for the *maybe_improper* list types, the list may also be proper. * Properness: proper | improper | maybe improper *...

I agree, very good summary of docs and dialyzer behaviour Ulf! > I think your arguments for the YesNo interpretation are quite compelling. Personally, I've always assumed the YesYes interpretation...

Aha, I understand how you are thinking. The purpose of the second parameter is to be the type of the "terminator". `[a | b]` is an improper list according to...

> Is that a fair characterization of your standpoint? Yes, exactly. Regarding `nonempty_maybe_improper_list`, it may be proper but it may not be empty. ```Erlang -type nonempty_maybe_improper_list (a, b) :: nonempty_improper_list(a,...

More evidence... ``` 1> is_list([a|b]). true 2> is_list([a]). true 3> is_list([]). true 4> is_list(b). false ```

[off topic] @OvermindDL1 It seems what you have found is the `is_list` C macro, which is checking if a term is a cons cell, nothing else. The Erlang BIF `erlang:is_list/1`...

I think a consensus would be nice, rather than a fast decision. Of course we can take a preliminary decision if we want to, but there is no hurry to...

> @zuiderkwast fwiw I don't find `is_list(b) == false` convincing argument for the NoYes interpretation. After all, `is_list/1` is only concerned about proper lists Of course `is_list/1` is only looking...

I ran in to this problem today and added a known problem test case in beb5a775dab61813febb90d73800efe43ca8ea54

Sorry for the commit message and the filename mentioning "exhaustiveness", which the issue is not about. Unreachable clauses is similar though, in a way. I'll rename the file one day...