Complexity measurement algorithms for information systems schemas are considered. Graph representations, based on an objectrelation pardigm and linguistic models, are discussed. Software science metrics are evaluated as complexity measures, as is the cyclomatic complexity measure. The deficiencies of current measures are highlighted. An alternative structural complexity metric is proposed that reflects propagation effects. The system development life cycle is used to determine realms of complexity that provide a framework for evaluation of complexity of designs and for projecting complexity between system development life cycle phases.