Architecting Systems with Trustworthy Components [electronic resource] : International Seminar, Dagstuhl Castle, Germany, December 12-17, 2004. Revised Selected Papers / edited by Ralf H. Reussner, Judith A. Stafford, Clemens A. Szyperski.

Invited Articles -- Audition of Web Services for Testing Conformance to Open Specified Protocols -- A Core Theory of Interfaces and Architecture and Its Impact on Object Orientation -- Making Specifications Complete Through Models -- Bus Scheduling for TDL Components -- Refinement and Consistency in Component Models with Multiple Views -- Articles by Participants -- A Taxonomy on Component-Based Software Engineering Methods -- Unifying Hardware and Software Components for Embedded System Development -- On the Composition of Compositional Reasoning -- Trustworthy Instantiation of Frameworks -- Performance Prediction of Component-Based Systems -- Towards an Engineering Approach to Component Adaptation -- Compatible Component Upgrades Through Smart Component Swapping -- Exceptions in Component Interaction Protocols – Necessity -- Coalgebraic Semantics for Component Systems -- A Type Theoretic Framework for Formal Metamodelling.

Software components are most generally viewed as a means of software re-use and, as such, much past research has been devoted to the study of problems associated with - tegrating components into cohesive systems. However, even when a collection of tru- worthycomponentshavebeensuccessfullyassembledthequalityoftheresultantsystem is not guaranteed. In December 2004, 41 experts on this topic from around the world, from research as well as industrial organizations, came together at Dagstuhl to discuss pressing issues related to architecting software systems from trustworthy components. Duringthecourseofthecold,yetsunny,DecemberdaysinDagstuhl,discussions- sions addressed topics such as compositionalreasoningon varioussystem-level prop- ties (such as deadlocks, live-locks etc.), compositional prediction models for different quality attributes (such as performance or reliability), blame analysis, interaction p- tocols, and composition frameworks. Using the liberal form of Dagstuhl Seminars, the days of the seminar were ?lled mostly with discussion in a variety of settings: in wo- ing sessions, around the table at meals, small groups in a corner, and also all together in the main meeting room.