Data integration often requires a clean abstraction of the different formats in which data are stored, and means for specifying the correspondences/relationships between data in different worlds and for translating data from one world to another. For that, we introduce in this paper a middleware data model that serves as a basis for the integration task, and a declarative rules language for specifying the integration. We show that using the language, correspondences between data elements can be computed in polynomial time in many cases, andmay require exponential time only when insensitivity to order or duplicates are considered. Furthermore, we show that in most practical cases the correspondence rules can be automatically turnedinto translation rules to map data from one representation to another. Thus, a complete integration task (derivation of correspondences, transformation of data from one world to the other, incremental integration of a new bulk of data, etc.) can be specified using a single set of declarative rules.

Structured data stored in les can bene t from standard database technology. In particular, we show here how such data can be queried and updated using declarative database languages. We introduce the notion of structuring schema which consists of a grammar annotated with database programs. Based on a structuring schema, a le can be viewed as a database structure, queried and updated as such. For queries, weshow that almost standard database optimization techniques can be used to answer queries without having to construct the entire database. For updates, we study in depth the propagation to the le of an update speci ed on the database view of this le. The problem is infeasible in general and we present anumber of negative results. The positive results consist of techniques that allow to propagate updates e ciently under some reasonable locality conditions on the structuring schemas.

Relationships among different modeling perspectives have been systematically investigated focusing either on given notations (e.g. UML) or on domain reference models (e.g. ARIS/SAP). In contrast, many successful informal methods for business analysis and requirements engineering (e.g. JAD) emphasize team negotiation, goal orientation and flexibility of modeling notations. This paper addresses the question how much formal and computerized support can be provided in such settings without destroying their creative tenor. Our solution is based on a novel modeling language design, M-Telos, that integrates the adaptability and analysis advantages of the logic-based meta modeling language Telos with a module concept covering the structuring mechanisms of scalable software architectures. It comprises four components: (1) A modular conceptual modeling formalism organizes individual perspectives and their interrelationships. (2) Perspective schemata are linked to a conceptual meta meta...

The increasing amount of information available in the web demands sophisticated querying methods and knowledge discovery techniques. In this study, we introduce our architectural framework WIND for a data warehouse over a domain-specific thematic section of the Internet. The aim of WIND is to provide a partially materialized structured view of the underlying information sources, on which database querying can be applied and mining techniques can be developed. WIND loads web documents into several complementary local repositories like OODBMSs and text retrieval systems. This allows for a combination of attribute and content-oriented query processing. Special interest is paid to domain-specific document formats. To support conversion between (semi-)structured documents and database objects, we consider a technique for the generation of format converters based on the notion of object-grammars.

We investigate data organization in the World Wide Web. In order to extend traditional database techniques to the Web, we concentrate on stvctved 14eb sevwvs, those servers (now very common) in which data are organized according to precise structures and pages present strong regularities. The paper introduces several tools for the management of structured servers. We present a page ov{ented data model, called the ARANEUS )(t( Jod, inspired to the structures typically present in these servers. The model allows us to describe the scheme of a structured server, in the spirit of databases. Then, we develop a view definition language, called the ARANEUS /{w L(ng(g, based on the notion of v{gat{o in the scheme. It has two features: first, it allows to build relational views of the Web, which can then be queried using any relational query language; second, it allows the definition of derived Web structures from relational views. This can be used to restructure query results and access them as an hypertext or to build derived servers.