(I) A Declarative Framework for ERP Systems(II) Reactors: A Data-Driven Programming Model for Distributed Applications

Christian Oskar Erik Stefansen

4968 Downloads (Pure)

Abstract

This dissertation is a collection of six adapted research papers pertaining to
two areas of research.
(I) A Declarative Framework for ERP Systems:
• POETS: Process-Oriented Event-driven Transaction Systems. The paper
describes an ontological analysis of a small segment of the enterprise
domain, namely the general ledger and accounts receivable. The result is
an event-based approach to designing ERP systems and an abstract-level
sketch of the architecture.
• Compositional Specification of Commercial Contracts. The paper describes
the design, multiple semantics, and use of a domain-specific language
(DSL) for modeling commercial contracts.
• SMAWL: A SMAll Workflow Language Based on CCS. The paper shows
how workflow patterns can be encoded in CCS and proceeds to design a
macro language, SMAWL, for workflows based on those patterns. The
semantics of SMAWL is defined via translation to CCS.
• Using Soft Constraints to Guide Users in Flexible Business Process Management
Systems. The paper shows how the inability of a process language
to express soft constraints—constraints that can be violated occasionally,
but are closely monitored—leads to a loss of intentional information
in process descriptions. This in turn makes it difficult for a process
execution engine to help its users adhere to established practices. The
paper then describes how soft constraints can be used to capture preferred
practices explicitly in process descriptions.
• A Work Allocation Language with Soft Constraints. Based on the idea
of soft constraints the paper explains the design, semantics, and use of
a language for allocating work in business processes. The language lets
process designers express both hard constraints and soft constraints.
(II) The Reactors programming model:
• Reactors: A Data-Oriented Synchronous/Asynchronous Programming Model
for Distributed Applications. The paper motivates, explains, and
defines a distributed data-driven programming model. In the model a
reactor is a stateful unit of distribution. A reactor specifies constructive,
declarative constraints on its data and the data of other reactors in the
style of Datalog. An attempt to update the data of a reactor starts a
reaction during which other reactors may be affected. The reaction ends
when all constraints of all affected reactors are satisfied or when it is
clear that this is not possible (conflict).
Original languageEnglish
Place of PublicationKøbenhavn
PublisherDepartment of Computer Science, University of Copenhagen
Number of pages307
Publication statusPublished - 2008

Fingerprint

Dive into the research topics of '(I) A Declarative Framework for ERP Systems(II) Reactors: A Data-Driven Programming Model for Distributed Applications'. Together they form a unique fingerprint.

Cite this