Comment by mdmoss

I think the author's statement of the Liskov substitution principle is backwards; An object must be replaceable by a subtype, not a supertype.