This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
Section 5.2.3 (Static Typing Feature) says: If an implementation does not provide the Static Typing Feature, then it MAY report type errors during the static analysis phase only in cases where the inferred static type and the required type have an empty intersection (that is, where evaluation of the expression is guaranteed to fail). It MAY defer some or all type checking until the dynamic evaluation phase. The use of an RFC MAY here is very misleading. The sentence is not intended to be permissive, it is intended as a constraint. "MAY... ONLY" here means "MUST NOT.... EXCEPT". I suggest: If an implementation does not provide the Static Typing Feature, then it MUST NOT report type errors during the static analysis phase except in cases where the inferred static type and the required type have an empty intersection (that is, where evaluation of the expression is guaranteed to fail). It MAY defer some or all type checking until the dynamic evaluation phase.
I agree. I believe this is editorial. Unless someone objects, I will make this change.
The Working Group agreed with this resolution at the joint teleconference on 2015-09-01.
Fixed in the spec.