Abstract
Program model linking has been largely overlooked and not constrained properly in the extraction of software system models. This often results in inaccurate system models at different levels of abstraction even if programs can be extracted correctly. This paper describes two constrained approaches toward accurate linkage resolution. The first approach is purely based on a set of linking heuristics. The second approach leverages the software build process and also utilizes linking heuristics. We compare these two approaches and discuss their benefits and limitations. The empirical results from a case study of the PostgreSQL database system are also presented. Our study shows that inappropriate linkage resolution leads to a relatively large number of dependency anomalies at higher levels of abstraction. These anomalies can be effectively removed using our proposed approaches.