Beyond its contracts and invariants, Meyer made Eiffel a very "opinionated" language, with those opinions argued forcefully in his Object-Oriented Software Construction. It seems to me that well-argued, almost "ideological" designs tend to have outsized influence. The original relational database papers by Codd come similarly to mind.

It isn't just having a good mathematical model or strong abstract theory. Eiffel got co/contravariance wrong (or at least, incomplete), after all.