UWN

Now, double bars would nicely fit.

What surprised me is that even round brackets are unable to resolve this. Could be some stateful thing.

This may help: ``` ?- write_canonical({(a).>} .>). outputs(".>(.>(a))"), unexpected. outputs(".>({}(.>(a)))"). % expected, but not found

``` ?- writeq(a.> .>). outputs("(a.> ).>"), unexpected. outputs("a.> .>").

Please refer to the original quad: ``` ?- writeq(((a) .>) .>). outputs("a.>"), unexpected. outputs("a.> .>"). % expected, but not found ```

Posted. This is a top level issue which is out of scope (1Scope, Note f). And even `portray_clause/1` handles this correctly in both Trealla and Scryer. ``` ?- portray_clause(a.>). a.>...

Do I understand this correctly that now this bitvector is updated all the time? It only needs to do so during GC.

For argument, see 3.7. A compound term has arguments. A goal has arguments similarly. A parameter, more specifically a lambda parameter, is not an argument (or, strictly speaking it is...

`?- X+\(X=a).` No parameters. Only renaming of variables (in this case none) within the goal. Note that λ is deconstructed to `\` and `^`. The `^` is responsible for parameter...

Better leave such changes to the author.