The communicating sequential processes (CSP) formalism, introduced by Hoare [Hoa85], is an event-based approach to distributed computing. The action-system formalism, introduced by Back & Kurki-Suonio [BKS83], is a state-based approach to distributed computing. Using weakest-precondition formulae, Morgan [Mor90a] has defined a correspondence between action systems and the failures-divergences model for CSP. Simulation is a proof technique for showing refinement of action systems. Using the correspondence of [Mor90a], Woodcock & Morgan [WM90] have shown that simulation is sound and complete in the CSP failures-divergences model. In this thesis, Morgan's correspondence is extended to the CSP infinite-traces model [Ros88] in order to deal more properly with unbounded nondeterminism. It is shown that simulation is sound in the infinite-traces model, though completeness is lost in certain cases. The new correspondence is then extended to include a notion of internal action. This allows the ...

A Functional Database Phil Trinder D.Phil. Thesis Wolfson College Michaelmas Term, 1989 This thesis explores the use of functional languages to implement, manipulate and query databases. Implementing databases. A functional language is used to construct a database manager that allows efficient and concurrent access to shared data. In contrast to the locking mechanism found in conventional databases, the functional database uses data dependency to provide exclusion. Results obtained from a prototype database demonstrate that data dependency permits an unusual degree of concurrency between operations on the data. The prototype database is used to exhibit some problems that seriously restrict concurrency and also to demonstrate the resolution of these problems using a new primitive. The design of a more realistic database is outlined. Some restrictions on the data structures that can be used in a functional database are also uncovered. Manipulating databases. Functions over the database a...

. Conceptual Structures (CS) Theory is a logic-based knowledge representation formalism. To show that conceptual graphs have the power of first-order logic, it is necessary to have a mapping between both formalisms. A proof system, i.e. axioms and inference rules, for conceptual graphs is also useful. It must be sound (no false statement is derived from a true one) and complete (all possible tautologies can be derived from the axioms). This paper shows that Sowa's original definition of the mapping is incomplete, incorrect, inconsistent, and unintuitive, and the proof system is incomplete too. To overcome these problems a new translation algorithm is given and a complete proof system is presented. Furthermore, the framework is extended for higher-order types. Key phrases: logical foundations of Conceptual Structures; OE operator; inference rules; logical axioms; higher-order types; meta-level reasoning. 1 Introduction The logical foundation of CS Theory, as presented in [...

Contemporary business process systems are built to automate routine procedures. Automation demands well-understood domains, repetitive processes, clear organisational roles, an established terminology, and predefined plans. Knowledge work is not like that. Plans for knowledge intensive processes are elaborated and reinterpreted as the work progresses. Interactive process models are created and updated by the project participants to reflect evolving plans. The execution of such models is controlled by users and only partially automated. An interactive process system should - Enable modelling by end users, - Integrate support for ad-hoc and routine work, - Dynamically customise functionality and interfaces, and - Integrate learning and knowledge management in everyday work.

Syntax Q = fT N jEg E = E EjE Ej:EjT 2 RjA!Aj9T : R:Ej8T : R:E A = N jBjT [N ] op =! j ? j j j 6= j = 5.3 Semantic Domains T uple = List V al V al = Num+Bool + Unb Unb = T ide X Num Dbase = RIde ! Set T uple Env = T Ide ! T uple 5.4 Semantic Functions 5.4.1 j : Nmls ! V al Not specified 5.4.2 fi : Bools ! V al Not specified 5.4.3 Q : Query ! Dbase ! SetT uple Q[jfT N jEgj] ffi = fae[jT j]j ae 2 E[jEj] ffi fT ! unbtuplegg where unbtuple = null [jT j] j[jN j] 5.4.4 E : Exp ! Dbase ! Env ! Set Env E[jE 0 E 1 j] ffi ae = fae 1 jae 0 2 (E[jE 0 j]ffiae) ae 1 2 (E[jE 1 j]ffiae 0 )g E[jE 0 E 1 j] ffi ae = fae 0 jae 0 2 (E[jE 0 j]ffiae) ae 0 2 (E[jE 1 j]ffiae)g E[j:Ej] ffi ae = f ilter (E[jEj]ffiae) = OE ae E[jT 2 Rj] ffi ae = (unbtuple? ae[jT j] !fae \Phi fT 7! vgj v 2 ffi[jRj]g; f ilter (ae[jT j] 2 ffi[jRj]) ae) E[jA 0 ! A 1 j] ffi ae = \Theta [j!j] (A[jA 0 j]ae) (A[jA 1 j]ae) ae E[j9T : R:Ej] ffi ae = fae 1 jv 2 ffi[jRj] ae 1 2 (E[jEj] ffi ae \Phi fT 7! vg)g E[j8T : R:E...

Many software engineering courses include all-term projects to convey principles relating to large-scale multi-person development. But even such projects will easily be too small and simple, unless a sufficient amount of study time is allocated to them. This time may be hard to find, especially in strictly programmed profession studies where a lot of general theory courses have to be taken. This paper reports on the experiences from a software engineering project where the solution to the above problem has been to have several courses share one project. This had some advantages. First of all, it allows time for a bigger and more complex project with reasonable sacrifices of "own time " in each of the participating courses. Equally important, it is possible to show connections between the courses. In spite of these advantages, there have also been problems with the project, still leaving room for improvement. 1.

We study Interval Temporal Logics over infinite intervals. First, the ordinary possible worlds models are extended to the infinite possible worlds models. Accordingly, an axiomatic system is proposed and it has been proved complete. Secondly, infinite intervals are included in a logic over abstract intervals. An axiomatic system is given and proven to be complete. Wang Hanpin is a faculty member of department of computer science and technology, Peking University, P. R. China. He is a fellow of UNU/IIST from February 1997 to August 1997 and from May 1998 to August 1998. His research interests lie in Mathematical Logic(especially Model Theory), Computational Complexity and logics of Programs. E-mail: whp@iist.unu.edu Xu Qiwen is a Research Fellow of UNU/IIST. His research interest is in Formal Techniques of Programming, including concurrency, verification, and design calculi. E-mail: qxu@iist.unu.edu Copyright c fl 1999 by UNU/IIST, Wang Hanpin and Xu Qiwen Contents i Contents 1 Int...

Abstract. In an organizational context the norms that apply to an agent depend on the roles he holds in the organization. The deontic characterization of structural roles is defined when the organization is created. But an organization is not a static entity. Among the dynamic phenomena that occur in an organization there are interactions between agents consisting in a transference of obligations or permissions from an agent to another. These kind of interactions are called delegation. In this paper we analyze different ways in which delegation occurs in an organizational context. We argue that the concept of “agent in a role ” is relevant to understand delegation. A deontic and action modal logic is used to specify this concept. 1

We propose and discuss a complete sequent calculus formulation for Signed Interval Logic (SIL) with the chief purpose of improving proof support for SIL in practice. The main theoretical result is a simple characterization of the limit between decidability and undecidability of quantifier-free SIL. We present a mechanization of SIL in the generic proof assistant Isabelle and consider techniques for automated reasoning. Many of the results and ideas of this report are also applicable to traditional (non-signed) interval logic and, hence, to Duration Calculus. 1

Model At this point, while the model is defined formally, the notion of event is not. The notion of event is an intuitive idea, and is meant to be identified with some occurrence of the system being modeled that is of interest to the user. While this is a useful idea, it is not a formal definition. The semantics of the previous section charac- 21 terize an event by its properties, namely, what kind of event is it, and when does it happen. That information is sufficient for the RTM, and what follows. This section, however, will recast the model by attempting to give a more explicit definition of event, and show how the semantics can be built up from this definition. The definition in this section will not be referred to again in what follows though, and is intended primarily as an illustration that if necessary, the intuitive notion of event can be defined, although the intuitive notion may be more satisfying, and is more useful from the perspective of a specification writer. The abst...