EMMSAD 2009 was organized in Amsterdam, the Netherlands, on June 8-12, 2009. It was held in conjunction with CAiSE'09.
The field of information systems analysis and design includes numerous information modeling methods and notations (e.g., ER, ORM, UML, DFDs, BPMN), that are typically evolving. Even with some attempts to standardize (e.g., UML for object-oriented design), new modeling methods are constantly being introduced, many of which differ only marginally from existing approaches. These ongoing changes significantly impact the way information systems are being analyzed and designed in practice. EMMSAD focuses on exploring, evaluating, and enhancing current information modeling methods and methodologies. Although the need for such studies is well recognized, there is a paucity of such research in the literature.
The objective of EMMSAD 2009 was to provide a forum for researchers and practitioners interested in modeling methods in systems analysis and design to meet and exchange research ideas and results. It also provided the participants with an opportunity to present their research papers and experience reports and to take part in open discussions. EMMSAD 2009 was the 14th in a very successful series of events, previously held in Heraklion, Barcelona, Pisa, Heidelberg, Stockholm, Interlaken, Toronto, Velden, Riga, Porto, Luxembourg, Trondheim, and Montpellier.
This year we had 36 papers submitted from 18 countries (Argentina, Austria, Brazil, Canada, China, France, Germany, Israel, Italy, Latvia, Luxembourg, The Netherlands, Norway, South Africa, Spain, Sweden, Switzerland, United Kingdom). After an extensive review process by a distinguished international Program Committee, with each paper receiving at least three reviews, we accepted the 16 papers that appear in these proceedings. Congratulations to the successful authors!
Apart from the contribution of the authors, the quality of EMMSAD 2009 depends in no small way on the generous contribution of time and effort by the Program Committee and the additional reviewers. Their work is greatly appreciated. We also express our sincere thanks to the CAiSE Organizing Committee, especially the CAiSE Workshop and Tutorial chairs Paul Johannesson (KTH, Stockholm, Sweden) and Eric Dubois (CRP Henri Tudor, Luxembourg).
Continuing with our very successful collaboration with IFIP WG 8.1 that started in 1997, this year’s event was again a joint activity of CAiSE and WG 8.1. The European INTEROP Network of Excellence has also sponsored this workshop since 2005, as has AIS-SIGSAND.
Business Process annotation with semantic tags taken from an ontology is becoming a crucial activity for business designers. In fact, semantic annotations help business process comprehension, documentation, analysis and evolution. However, building a domain ontology and annotating a process with semantic concepts is a difficult task. In this work, we propose an automated technique to support the business designer both in domain ontology creation/extension and in the semantic annotation of process models expressed in BPMN. We use natural language processing of the labels appearing in the process elements to construct a domain ontology skeleton or to extend an existing ontology, if available. Semantic annotations are automatically suggested to the business designer, based on a measure of similarity between ontology concepts and the labels of the process elements to be annotated.
Ontologies are commonly used in computer science either as a reference model to support semantic interoperability, or as an artifact that should be efficiently represented to support tractable automated reasoning. This duality poses a tradeoff between expressivity and computational tractability that should be addressed in different phases of an ontology engineering process. The inadequate choice of a modeling language, disregarding the goal of each ontology engineering phase, can lead to serious problems in the deployment of the resulting model. This article discusses these issues by making use of an industrial case study in the domain of Oil and Gas. We make explicit the differences between two different representations in this domain, and highlight a number of concepts and ideas that were implicit in an original OWL-DL model and that became explicit by applying the methodological directives underlying an ontologically well-founded modeling language.
While formal methods are focused on some particular parts of software systems, especially secure ones, graphical techniques are the most useful techniques to specify in a comprehensible way large and complex systems. In this paper we deal with the B method which is a formal method used to model systems and prove their correctness by successive refinements. Our goal is to produce graphical UML views from existing formal B specifications in order to ease their readability and then help their external validation. In fact, such views can be useful for various stakeholders in a formal development process: they are intended to support the understanding of the formal specifications by the requirements holders and the certification authorities; they can also be used by the B developers to get an alternate view on their work. In this paper, we propose an MDE framework to support the derivation of UML class and state/transition diagrams from B specifications. Our transformation process is based on a reverse-engineering technique guided by a set of structural and semantic mappings specified on a meta-level.
The Model Driven Architecture (MDA) is an architectural framework for information integration and tool interoperation that could facilitate system modernization. Reverse engineering techniques are crucial to extract high level views of the subject system. This paper describes a reverse engineering approach that fits with MDA. We propose to integrate different techniques that come from compiler theory, metamodeling and formal specification. We describe a process that combines static and dynamic analysis for generating MDA models. We show how MOF (Meta Object Facility) and QVT (Query, View, Transformation) metamodels can be used to drive model recovery processes. Besides, we show how metamodels and transformations can be integrated with formal specifications in an interoperable way. The reverse engineering of class diagram and state diagram at PSM level from Java code is exemplified.
Conceptual modeling traditionally focuses on a high level of abstraction. Even if geographical aspects such as location is included in several enterprise modeling frameworks [26], it is not common to have geographical aspects included in conceptual models. Cartography is the science of visualizing geographical information in maps. Traditionally the field has not included conceptual relationships and the primary focus is on a fairly low abstraction level. Both cartography and conceptual modeling have developed guidelines for obtaining high quality visualizations. SEQUAL is a quality framework developed for understanding quality in conceptual models and modeling languages. In cartography such counterparts are not common to find. An attempt to adapt SEQUAL in the context of cartographic maps has been performed, named MAPQUAL. The paper presents MAPQUAL. Differences between quality of maps and quality of conceptual models are highlighted, pointing to guidelines for combined representations which are the current focus of our work. An example of such combined use is presented indicating the usefulness of a combined framework.
Recently a great number of methods and frameworks to develop multiagent systems have appeared. It makes difficult the selection between one and another. Because of that the evaluation of multiagent system software engineering techniques is an open research topic. This paper presents an evaluation framework for analyzing and comparing methods and tools for developing multiagent systems. Furthermore, four examples of usage are presented and analyzed.
Languages for specifying information systems should not only contain a data definition (sub)language (DDL), i.e., a part for specifying data structures, but also a data retrieval (sub)language (DRL), i.e., a part for specifying queries, and a data manipulation (sub)language (DML), i.e., a part for specifying transactions. The language ORM contains a DDL and a DRL (ConQuer), but it does not contain a sufficient DML as yet. We therefore propose an extension of ORM with a DML, for specifying transactions to be easily validated by domain experts. We introduce the following set of standard classes of specifiable transactions: add an instance, add a query result, remove a subset, and change a subset. We also treat compound transactions in ORM. In ORM there are usually several ways to specify something. For all transactions we therefore propose syntaxes, verbalizations, and diagrams as well. They allow for type-checking and easy validation by domain experts.
In this paper we will illustrate how the fact-oriented approach, e.g. ORM, CogNiam can be used in combination with OMG's Semantics of Business Vocabulary and Business Rules' (SBVR) standard. Within the field of modeling methods for information systems analysis and design, this standard can become a dominant force, when it comes to expressing initial domain requirements for an application's ontoloy and business rules, for domain analysis as well for design
Today's large-scale computing systems are deployed in open, changing and unpredictable environments. To operate reliably, such systems should be able to adapt to new circumstances on their own to get them running and keep them running. Self-adaptive software system has been proposed as a good solution for this demand. However, very few techniques are available to date for systematically building such kind of system. Aiming at this requirement, this paper presents a sound approach to derive a self-adaptive software architecture model from the requirements goal model in systematic way. At the same time, we illustrate our approach by applying it to a simplified on-line shopping system.
Self-contextualizability refers to the system ability to autonomously adapt its behaviour to context in order to maintain its objectives satisfied. In this paper, we propose a modeling framework to deal with self-contextualizability at the requirements level. We use Tropos goal models to express requirements; we provide constructs to analyse and represent context at each variation point of the goal model; and we exploit the goal and context analysis to define how the system satisfies its requirements in different contexts. Tropos goal analysis provides constructs to hierarchically analyse goals and discover alternative sets of tasks the system can execute to satisfy goals; our framework extends Tropos goal model by considering context at its variation points, and provides constructs to hierarchically analyse context and discover alternative sets of facts the system has to monitor to verify a context. A self-contextualizable promotion information system scenario is used to illustrate our approach.
This paper shows typical security and consistency challenges regarding the models of the business and the IT universe of the dynamic service-, process- and rule-based environment at Credit Suisse. It presents a theoretical solution for enterprise engineering that is implementable, and fits smoothly with the daily needs and constraints of the people in the scenario. It further enables decentralized modeling based on cognitive and mathematical or logical concepts. Normative aspects of the models are analyzed by graph constraint checks, while consistency is checked and ensured by model integration and transformation. To cope with theoretical and practical necessities, the presented solution is kept sound and usable as well as extensible and scalable. All techniques are based on one theoretical framework: algebraic graph theory. Therefore, the techniques are compatible with each other.
Advantages and shortcomings of different process modeling languages are heavily debated, both in academia and industry, but little evidence is presented to support judgements. With this paper we aim to contribute to a more rigorous, theoretical discussion of the topic by drawing a link to well-established research on program comprehension. In particular, we focus on imperative and declarative techniques of modeling a process. Cognitive research has demonstrated that imperative programs deliver sequential information much better while declarative programs offer clear insight into circumstantial information. In this paper we show that in principle this argument can be transferred to respective features of process modeling languages. Our contribution is a pair of propositions that are routed in the cognitive dimensions framework. In future research, we aim to challenge these propositions by an experiment.
In current business practice, an integrated approach to business and IT is indispensable. In many enterprises, however, such an integrated vi ew of the entire enterprise is still far from reality. To deal with these challenges, an integrated view of the enterprise is needed, enabling impact and change analysis covering all relevant aspects. This need sparked the development of the ArchiMate language. This paper is concerned with documenting some of the key design decisions and design principles underlying the ArchiMate language.
Meta models are the core of enterprise architecture, but still few methods are available for the creation of meta models tailored for specific purposes. This paper presents two approaches, one focusing on the stakeholders' information demand of enterprise architecture and the other driven by causal analysis of enterprise system properties. The two approaches are compared and a combined best-of-breed method is proposed. The combined method has merged the strengths of both approaches, thus combining the stakeholder concerns with causality-driven analysis. Practitioners will, when employing the proposed method, achieve a relevant meta model with strong, and goal-adapted, analytic capabilities.
This research captures best practices in the business-to-business (B2B) domain as a means of competitive advantage and innovation for organizations striving to adopt B2B environments. We present a case of developing and validating a set of patterns for B2B adoption and then discuss the case in the context of a number of other cases where organizational patterns have been used to capture, document and share competitive organizational knowledge.
Enterprise Modeling (EM) is used for a wide range of purposes such as developing business strategies, business process restructuring, business process orientation and standardization, eliciting information system requirements, capturing best practices, etc. A common challenge impeding the value and impact of EM is insufficient model quality. Despite substantial attention from both researchers and commercial vendors of methods the current situation in practice with respect to the quality of models produced is not satisfactory. Many modeling projects produce bad models that are essentially useless. The objective of this paper is to introduce a format, anti-patterns, for documenting critical don'ts in EM and to demonstrate the potential of the format by using it to report a set of common and reoccurring pitfalls of real life EM projects. We use the format of anti-pattern for capturing the bad solutions to reoccurring problems and then explain what led to choosing the bad solution. The anti-patterns in this paper address three main aspects of EM - the modeling product, the modeling process, and the modeling tool support.
Apart from the contribution by paper authors, the quality of this workshop depends on no small way on the generous contribution of time and effort by the program committee. Their work is greatly appreciated. We also express our sincere thanks to the CAiSE organizing committee, especially the CAiSE Workshop organizers.
Additional reviewers