a/compsci posted by smh 6 months ago

A possible indication that an implementation of a type is not strongly linearizable is that the response of one of the operations is not determined at its linearization point. Otherwise events after a linearization could alter the response of the operation, and likely the linearization order will need to change to maintain validity. We can often construct a counter example to an implementation not being strongly linearizable using this observation. The question remains whether this observation directly implies that an implementation is not strongly linearizable. 

Log in or sign up to leave a comment     
Log in
Sign up