Methods of limiting the impact of communication delays on the logical behavior of distributed systems are considered. It is assumed that a distributed system is described in terms of a number of interconnected modules, and each module is described in terms of its possible states and the possible state transitions. Transitions may be initiated spontaneously by a module and may give rise to output messages, which will be received, after some possible delay, by another module as an input. Otherwise, transitions may be initiated by received input. If the system has the property called regularity, its behavior is logically independent of the communication delays. A simple condition for regularity is given. This condition is the basis for the implementation of counter-based synchronization conditions in a distributed environment. Weaker forms of regularity, which make abstraction of internal operations invisible from the point of view of an outside observer, are also considered. The application of these concepts to the design of module interfaces involving 'collisions' and to communication including timeouts is discussed in some detail with examples. >