In 2000 the UK Engineering Professors' Council (EPC) drafted an output standard to describe first-cycle engineering programmes that would prepare students for practice and further professional learning. The standard described what is authentic and worthwhile in engineering education--it identified complex outcomes of learning. This poses practical and theoretical challenges: how can we know if a student has met this standard? This paper argues that complex outcomes like these often resist measurement and that it is appropriate to use other forms of judgement when trying to assess student achievement. This differentiated approach to assessment, which values formative assessment (or feedback) as well as summative assessment (or feedout), is summarized. It becomes necessary, then, to think about the assessment arrangements for whole programmes, not just for individual modules. Some implications of this programmic and differentiated approach are explored. (Contains 1 table and 1 figure.)