. This paper gives a cohesive approach to modeling and implementation with graphs. This approach uses extended entity relationship #EER# diagrams supplemented with the Z-like constraint language GRAL. Due to the foundation of EER#GRAL on Z a common formal basis exists. EER#GRAL descriptions give conceptual models which can be implemented in a seamless manner by e#cient data structures using the GraLab graph library. Descriptions of four medium size EER#GRAL-applications conclude the paper to demonstrate the usefulness of the approach in practice. 1 Introduction Using graphs as a means for discussing problems, as a medium for formal reasoning, or as a paradigm for data structures in software is folklore in today's computer science literature. But most of the di#erent approaches that use graphs are not used in a coherentway. There are di#erent models in use based on undirected or directed graphs, with or without multiple edges or loops. Sometimes graph elements are typed or at...