One important aspect of software reuse is the organization of collections of reusable software artifacts. The Software Information Base (SIB), developed within the ESPRIT project ITHACA, provides a directory to reusable software by storing information about software objects concerning the entire software life-cycle, namely requirements, design and implementation descriptions, as well as aggregate representations of complete systems and application domains. The SIB has an attributed graph structure. The selection of artifacts from the SIB, either directly or through other software development tools, is performed using a specialized Selection Tool (ST). In this paper we present the design and functionality of the Selection Tool. The main search mode supported by the Selection Tool is browsing. It is a flexible navigation process that takes full advantage of the knowledge representation mechanisms underlying the SIB semantic network, and provides local search of controllable size, direct access to specific areas or objects in the SIB, filtering mechanisms, and orientation aids. The information stored in the SIB and displayed by the Selection Tool is multimedia. The representational issues addressed by the SIB — ST system, as well as the relationship between the ST and hypertext systems are discussed.