The paper discusses the challenges of modeling and querying complex objects in database systems. It highlights that complex objects, often found in new applications like engineering, require specific database management system (DBMS) capabilities. However, previous work has not clearly linked these requirements to the necessary DBMS features, often focusing on relational models. The paper aims to clarify the minimal set of capabilities needed for handling complex objects in a model-independent way. It outlines how these capabilities are being integrated into PROBE, an object-oriented DBMS being developed at CCA. The paper emphasizes that many new applications involve highly structured objects composed of other objects. For example, a part in a part hierarchy is made up of other parts, an integrated circuit module is composed of other modules, pins, and wires, and a geographic feature like an industrial park is made up of buildings, smokestacks, and gardens. These complex objects are often the units for storage, retrieval, update, integrity control, concurrency control, and recovery. For instance, in a design application, an entire part assembly may need to be locked when redesigned, and the deletion of an integrated circuit module must be propagated atomically to all its components. The basic problem is that in conventional database systems, complex objects are typically represented by many tuples scattered across several relations. For example, a 4-input AND gate is built from three 2-input AND gates, requiring multiple relations to represent the gate type, pin type, gate instance, and wire instance. The paper shows how PROBE is being developed to address these challenges by providing a more efficient and integrated approach to modeling and querying complex objects.The paper discusses the challenges of modeling and querying complex objects in database systems. It highlights that complex objects, often found in new applications like engineering, require specific database management system (DBMS) capabilities. However, previous work has not clearly linked these requirements to the necessary DBMS features, often focusing on relational models. The paper aims to clarify the minimal set of capabilities needed for handling complex objects in a model-independent way. It outlines how these capabilities are being integrated into PROBE, an object-oriented DBMS being developed at CCA. The paper emphasizes that many new applications involve highly structured objects composed of other objects. For example, a part in a part hierarchy is made up of other parts, an integrated circuit module is composed of other modules, pins, and wires, and a geographic feature like an industrial park is made up of buildings, smokestacks, and gardens. These complex objects are often the units for storage, retrieval, update, integrity control, concurrency control, and recovery. For instance, in a design application, an entire part assembly may need to be locked when redesigned, and the deletion of an integrated circuit module must be propagated atomically to all its components. The basic problem is that in conventional database systems, complex objects are typically represented by many tuples scattered across several relations. For example, a 4-input AND gate is built from three 2-input AND gates, requiring multiple relations to represent the gate type, pin type, gate instance, and wire instance. The paper shows how PROBE is being developed to address these challenges by providing a more efficient and integrated approach to modeling and querying complex objects.