ROBINSON@BITNIC.BITNET (Andrew T. Robinson) (02/05/90)
What is needed to create protocol-independent specifications is some patience and some discipline. It doesn't happen quickly and it doesn't happen without some tradeoffs--but it does happen. Certainly, we can strive for the GOAL of system independence even if we take the attitude that the goal can never be fully achieved. As far as the platypus problem goes, if the platypus doesn't walk and talk like a duck, it's out of spec and hence unsupported. If someone decides that the platypus has some desirable features that the duck doesn't have, they can submit a proposal for an extension to the specification that allows the platypus to exist--but if and only if the platypus is a superset (does not alter the basic structure) of the duck. Andy