Model-Checking Based Data Retrieval An Application to Semistructured and Temporal Data

This thesis deals with the problems of characterizing the semantics of and assuringe'cient executionfor databasequery languages, wherethedatabase contains semistructured and time-varying information. This area of techn- ogy is of much interest and signi'cance for databases and knowledge bases; it also presents many challenging research problems deserving an in-depth investigation.Thus, the topic of Elisa Quintarelli's dissertation is well chosen and totally appropriate to the current research trends. In her thesis, Elisa addresses a number of related problems. However, her work and contributions concentrate on two main problems. The ?rst is the de'nition of an e'ective graph-based approach to the formalization of query languages for semistructured and temporal information. In her approach, query execution is viewed as the process of matching the query graph with the database instance graph; therefore, query execution reduces to sear- ing the database for subgraphs that are similar to the given query graph. The search for such matches can be supported through the computational process of bisimulation. This approach is used to de'ne the semantics of s- eral languages, including graphical languages, such as G-Log and GraphLog, semistructuredinformationlanguages, suchasLorel, andtemporallanguages, such as TSS-QL. Both graph-based approaches and bisimulation had been used by previous authors for de'ning query languages and their semantics; however, this work goes well beyond previous approaches by integrating and re'ning these techniques into a ?exible and powerful paradigm that Elisa demonstrates to be e'ective on a spectrum of languages and a suite of alt- native semantics.