Catálogo de publicaciones - libros

In object-oriented programming, reentrant method invocations and shared references make it difficult to achieve adequate encapsulation for sound modular reasoning. This tutorial paper surveys recent progress using auxiliary state (ghost fields) to describe and achieve encapsulation. Encapsulation is assessed in terms of modular reasoning about invariants and simulations.

This paper considers an integration of asynchronous communication, virtual binding, and multiple inheritance. Object orientation is the leading paradigm for concurrent and distributed systems, but the tightly synchronized RPC communication model seems unsatisfactory in the distributed setting. Asynchronous messages are better suited, but lack the structure and discipline of traditional object-oriented methods. The integration of messages in the object-oriented paradigm is unsettled, especially with respect to inheritance and redefinition.

Asynchronous method calls have been proposed in the Creol language, reducing the cost of waiting for replies in the distributed environment while avoiding low-level synchronization constructs such as explicit signaling. A lack of reply to a method call need not lead to deadlock in the calling object. Creol has an operational semantics defined in rewriting logic. This paper considers a formal operational model of multiple inheritance, virtual binding, and asynchronous communication between concurrent objects, extending the semantics of Creol.

Object calculi have been investigated as semantical foundation for object-oriented languages. Often, they are object-based, whereas the mainstream of object-oriented languages is

Considering classes as part of a component makes instantiation a possible interaction between component and environment. As a consequence, one needs to take information into account.

We formulate an operational semantics that incorporates the connectivity information into the scoping mechanism of the calculus. Furthermore, we formalize a notion of equivalence on traces which captures the uncertainty of observation cause by the fact that the observer may fall into separate groups of objects. We use a corresponding trace semantics for full abstraction wrt. a simple notion of observability. This requires to capture the notion of for traces where classes may be instantiated into more than one instance during a run and showing thus twice an equivalent behavior (doing a “replay”), a problem absent in an object-based setting.

Hard real-time systems need methods to determine upper bounds for their execution times, usually called worst-case execution times. This paper explains the principles of our Timing-Analysis methods, which use Abstract Interpretation to predict the system’s behavior on the underlying processor’s components and use Integer Linear Programming to determine a worst-case path through the program. Under the assumption that non-trivial systems are subject of the analyses, exhaustive analyses can not be performed and some uncertainty about the system’s behavior remains. Uncertainty, i.e., lack of information about a system’s execution states incurs cost in terms of precision of the upper and lower bounds on the execution times. Some cost figures are given for missing information of different types. These are measured in machine clock cycles. It is (intuitively) argued, that component-based software design and the use of middleware may induce intolerable costs in terms of precision.