I working through the practice problems in Davie's book (by myself) and there is one that is not quite clear to me: "Construct an abstract data type for a queue. Attempt to supply an implementation (hidden from the user of the type) which allows access to the head of the queue and the construction of a new queue by adding elements to the end of an existing one. Both these interface functions should operate in time independent of the length of the queue. What implications does the this exercise suggest for search and 'updating' (by copying all unchanged nodes) a general graph?" There are two parts to the problem here: 1) defining the data type and the interface functions, and 2) indicating the implications for general graphs. On the first part I'm a little unclear because I am not sure what he means by "allows access to the head of the queue"; i.e., does he mean to just get the value in the head of the queue, or pop the value off the head of the queue and return the new queue? If the former, a solution is not hard, but if the latter, then I am not sure how to solve this without somehow transversing the queue, which of course is not a time-independent activity. (Even in the implementation for a queue in the Data.Graph.Inductive.Internal.Queue, eventually a "reverse" call must be made.) Anyway, maybe if someone could explain the relationship between this problem and graphs, I'll see the bigger picture and have a clearer view of the problem. -- frigidcode.com theologia.indicium.us