Final Activity Report Summary - XMLDATA (Designing, querying and exchanging XML data)
Research. The main scientific results achieved in this project can be subdivided into five categories.
1. Design and constraints (how to organise XML documents? how to describe desired properties of XML documents?) We have concentrated on the following aspects: producing consistent specifications (it had previously been shown by the chair holder and his co-authors that some very reasonably looking specifications of XML data may actually be inconsistent), design of XML documents, for both native and relational storage techniques, and handling of missing and incomplete information in XML documents.
2. Languages for unranked trees. As unranked trees form a theoretical model of XML documents, it is important to understand the power and the complexity of languages for querying them. We have done a complete analysis of the complexity of the languages proposed so far, and looked at languages based on a model-theoretic approach to unranked trees.
3. Data exchange and schema mappings. XML was introduced as a standard for data exchange on the web, and yet most data translation/exchange/integration tools go via relational representations. To build a theory of schema mappings and exchange/translation of data for XML, we first developed missing parts of the relational theory (accounting for incompleteness and open/closed world distinctions in data exchange), and then applied them to XML, producing descriptions of schema mappings, their analyses, operations on them, and query answering algorithms.
4. Static analysis. This underlies both reasoning about schemas and descriptions, and basic query optimisation tasks. We concentrated on connections between classical program analysis techniques and XML documents, contributing to both fields. We worked on automata-based approaches to XML static analysis, used XML languages to produce verification techniques for programs with recursive procedure calls, and used tree transducers to answer questions about verification of infinite-state systems.
5. Foundations. Most of the theoretical work on XML is grounded in logic and automata theory, and we produced theoretical results in those fields as well.
Teaching/training. Two new courses were designed and introduced at the University of Edinburgh. Three PhD students and four postdoctoral fellows were supervised.
In numbers: 5 conference keynotes, 2 conference chairmanships, 11 public lectures, 2 new courses, 4 new grants (3 as PI, total over Euro 2.5M) 1 book, 1 edited book, 7 journal papers, 13 conference papers.
1. Design and constraints (how to organise XML documents? how to describe desired properties of XML documents?) We have concentrated on the following aspects: producing consistent specifications (it had previously been shown by the chair holder and his co-authors that some very reasonably looking specifications of XML data may actually be inconsistent), design of XML documents, for both native and relational storage techniques, and handling of missing and incomplete information in XML documents.
2. Languages for unranked trees. As unranked trees form a theoretical model of XML documents, it is important to understand the power and the complexity of languages for querying them. We have done a complete analysis of the complexity of the languages proposed so far, and looked at languages based on a model-theoretic approach to unranked trees.
3. Data exchange and schema mappings. XML was introduced as a standard for data exchange on the web, and yet most data translation/exchange/integration tools go via relational representations. To build a theory of schema mappings and exchange/translation of data for XML, we first developed missing parts of the relational theory (accounting for incompleteness and open/closed world distinctions in data exchange), and then applied them to XML, producing descriptions of schema mappings, their analyses, operations on them, and query answering algorithms.
4. Static analysis. This underlies both reasoning about schemas and descriptions, and basic query optimisation tasks. We concentrated on connections between classical program analysis techniques and XML documents, contributing to both fields. We worked on automata-based approaches to XML static analysis, used XML languages to produce verification techniques for programs with recursive procedure calls, and used tree transducers to answer questions about verification of infinite-state systems.
5. Foundations. Most of the theoretical work on XML is grounded in logic and automata theory, and we produced theoretical results in those fields as well.
Teaching/training. Two new courses were designed and introduced at the University of Edinburgh. Three PhD students and four postdoctoral fellows were supervised.
In numbers: 5 conference keynotes, 2 conference chairmanships, 11 public lectures, 2 new courses, 4 new grants (3 as PI, total over Euro 2.5M) 1 book, 1 edited book, 7 journal papers, 13 conference papers.