Two major problems in the engineering of software-intensive, real time and distributed computer systems are, without becoming lost in code details, I) understanding how an implemented system works as a whole, and II) specifying, before implementation starts, how the required behaviour of the whole system is to be achieved. These lead to other problems, such as, long iteration cycles during forward engineering while various code changes are tried in attempts to fix erroneous system behaviour, and inadvertently introducing code changes during maintenance or reengineering that will damage correct system behaviour because there is a lack of backwards traceability to it from the code. This paper illustrates the application of a new technique called use case maps to solving these problems, using as an example a system constructed from a public domain, object-oriented, software framework called ACE.