Objectives To define the formalisms for expressing the Integration Information Model (IIM), and to define the user query language.

Description of work TASK 1: Specification of the formalism for expressing the IIM (Task Leader UNIROME1) The formalism for expressing the IIM will be formally defined. Such a formalism must allow for expressing, in a declarative way, the knowledge about the information sources to be integrated, and for specifying the global information schema, while balancing the trade-off between such expressiveness requirements and the issue of providing efficient information integration algorithms. TASK 2: Specification of the user query language (Task Leader UNICAL) The query language of the INFOMIX system will be formally defined. The language is to be fully declarative, highly expressive, and able to deal with semi-structured data to some extent. TASK 3: Handling inconsistency and incompleteness in the IIM. (Task Leader UNIROME1) The problem of dealing with inconsistency and/or incompleteness in the IIM specification will be analyzed, and formal methods for handling various forms of inconsistent/incomplete specifications will be defined.

Deliverables D3.1- Formalism for IIM specification (Report). This report provides the formal definition of the formalism for specifying the Information Integration Model. D3.2- Query language (Report). This report provides the formal definition of the query language of the INFOMIX system. D3.3- Techniques for handling inconsistency and incompleteness in the IIM (Report). This report provides the definition of formal methods for handling inconsistency and incomplete information in the IIM specification.

Milestones and expected results M2 - Acceptance of the theoretical foundations of the INFOMIX Modeling Formalism. The workpackage provides the detailed specification of the formalisms for expressing the Information Integration Model and the user queries in the INFOMIX system. This is a critical project step, since such formalisms must be designed in order to match one of the main project goals, i.e., expressing very rich forms of knowledge about the information to be integrated, while taking into account the complexity of the subsequent information integration activity. Formal specification of the formalisms for expressing both the Integration Information Model and the user queries will be provided. Moreover, dealing with inconsistent and/or incomplete information specifications is a crucial issue in the development of a general framework for information integration, since such aspects arise, in various forms, in the vast majority of real information integration scenarios. Formal methods for dealing with inconsistent and/or incomplete specifications of the Integration Information Model will be provided.