Cesar Gonzalez-Perez

A powertype-based metamodelling framework

Software development methodologies may be described in the context of an underpinning metamodel, but the precise mechanisms that permit them to be defined in terms of their metamodels are usually difficult to explain and do not cover all needs. For example, it is difficult to devise a way that allows the definition of properties of the elements that compose the methodology and, at the same time, of the entities (such as work products) created when the methodology is applied. This article introduces a new approach to constructing metamodels and deriving methodologies from them based on the concept of powertype. It combines key advantages of other metamodelling approaches and allows the seamless integration of process, modelling and documentational aspects of methodologies. With this approach, both methodology components and project entities can be described directly by the same metamodel.

FAML: A Generic Metamodel for MAS Development

In some areas of software engineering research, there are several metamodels claiming to capture the main issues. Though it is profitable to have variety at the beginning of a research field, after some time, the diversity of metamodels becomes an obstacle, for instance to the sharing of results between research groups. To reach consensus and unification of existing metamodels, metamodel-driven software language engineering can be applied. This paper illustrates an application of software language engineering in the agent-oriented software engineering research domain. Here, we introduce a relatively generic agent-oriented metamodel whose suitability for supporting modeling language development is demonstrated by evaluating it with respect to several existing methodology-specific metamodels. First, the metamodel is constructed by a combination of bottom-up and top-down analysis and best practice. The concepts thus obtained and their relationships are then evaluated by mapping to two agent-oriented metamodels: TAO and Islander. We then refine the metamodel by extending the comparisons with the metamodels implicit or explicit within five more extant agent-oriented approaches: Adelfe, PASSI, Gaia, INGENIAS, and Tropos. The resultant FAML metamodel is a potential candidate for future standardization as an important component for engineering an agent modeling language.

A comparison of four process metamodels and the creation of a new generic standard

Abstract Software development processes and methodologies to date have frequently been described purely textually. However, more recently, a number of metamodels have been constructed to both underpin and begin to formalize these methodologies. We have critically examined four of these: the Object Management Groups Software Process Engineering Metamodel (SPEM), the OPEN Process Framework (OPF), the OOSPICE metamodel for capability assessment and the LiveNet approach for computer-supported collaborative work (CSCW). Based on this analysis, a new, combined metamodel, named Standard Metamodel for Software Development Methodologies (SMSDM) has been constructed which supports not only process but also products and capability assessment in the contexts of both software development and CSCW. As a proof of concept we conclude with a partial example to show how the SMSDM metamodel (and by inference the other metamodels) are used in practice by creating a simple yet usable methodology.

Modelling software development methodologies: A conceptual foundation

Current modelling approaches often purport to be based on a strong theoretical underpinning but, in fact, contain many ill-defined concepts or even contradictions leading to potential misinterpretation. Although much modelling in object-oriented contexts is focussed on the use of the Unified Modelling Language (UML), this paper presents a technology-agnostic approach that analyses the basic concepts of structural models and modelling in software engineering, using an approach based on representation theory. We examine the different kinds of interpretive mappings (either isotypical, prototypical or metatypical) that are required in order to trace model entities back to the SUS (subject under study) entities that they represent. The difference between forward- and backward-looking models is also explained, as are issues relating to the appropriate definition of modelling languages in general based on representation theory. The need for product and process integration in methodologies is then addressed, leading to the conclusion that a mesh of verbal plus nominal nodes is necessary in any methodology metamodel. Finally, the need for a common, cross-cutting modelling infrastructure is established, and a solution proposed in the form of an ontologically universal modelling language, OOLang. Examples of the application of these theoretical analyses to the suite of OMG products (particularly SPEM, UML and MOF) are given throughout, with the hope that awareness of the importance of a better modelling infrastructure can be developed.

Connecting powertypes and stereotypes

Powertypes constitute an advanced OO modelling mechanism that is usually utilized in the form of a specific pattern. Stereotypes comprise the basic customization and extension mechanism in UML, and are also used following a certain pattern. Although different in purpose, these two patterns present some interesting similarities and are shown here to become structurally identical in specific circumstances. This fact can help reduce the apparent complexity of UML and may be of special importance for tools that store and transform models that use powertypes and stereotypes.

Modularization Constructs in Method Engineering: Towards Common Ground?

Although the Method Engineering (ME) research community has reached considerable maturity, it has not yet been able to agree on the granularity and definition of the configurable parts of methods. This state of affairs is causing unnecessary confusion, especially with an ever increasing number of people contributing to ME research. There are several competing notions around, most significantly ‘method fragments’ and ‘method chunks’, but also ‘method components’ and ‘process components’ are used in some quarters and have also been widely published. Sometimes these terms are used interchangeably, but there appears to be important semantic and pragmatic differences. If the differences are unimportant, we should be able to come to an agreement on what construct to promote. Alternatively, the different constructs may serve different purposes and there is a need for them to coexist. If this is the case, it should be possible to pinpoint exactly how they are related and which are useful in what contexts. This panel is a step towards finding common ground in this area, which arguably is at the very core of ME.

Synthesis of a generic MAS metamodel

Method engineering, which focuses on project-specific methodology construction from existing method fragments, is an appealing approach to organize, appropriately access and effectively harness the software engineering knowledge of MAS methodologies. With the objective of applying method engineering for developing an MAS, in this paper we introduce a generic metamodel to serve as a representational infrastructure to unify existing MAS methodologies into a single specification. Our metamodel does not focus on any class of MAS, nor does it impose any restrictions on the format of the system requirements; rather, our metamodel is an abstraction of how any MAS is structured and behaves both at design time and run-time.

An ontology for ISO software engineering standards: 1) Creating the infrastructure

Software engineering standards developed under the auspices of ISO/IEC JTC1s SC7 have been identified as employing terms whose definitions vary significantly between standards. This led to a request in 2012 to investigate the creation of an ontological infrastructure that aims to be a single coherent underpinning for all SC7 standards, present and future. Here, we develop that necessary infrastructure prior to its adoption by SC7 and its implementation (likely 2014). The proposal described here requires, firstly, the identification of a single comprehensive set of definitions, the definitional elements ontology (DEO). For the scope of an individual standard, only a subset of these definitional elements will be needed. Once configured, this definitional subset creates a configured definitional ontology or CDO. Both the DEO and the CDO are essentially foundational ontologies from which a domain-specific ontology known as a SDO or standard domain ontology can be created. Consequently, all such SDOs are conformant to a CDO and hence to the single DEO thus ensuring that all standards use the same ontological base. Standards developed in this fashion will therefore be not only of a higher quality but also, importantly, interoperable.

Templates and Resources in Software Development Methodologies

A great deal of effort is needed to construct software products in a predictable and repeatable manner. Having a precisely defined methodology in place can certainly help, especially if it includes the comprehensive specification of the process to be followed and the work products to be created. However, a convenient integration of these two aspects (process and work product) has not yet been performed. This paper presents a new approach to the definition of methodologies that supports the process and work product domains concurrently through the specification of discrete methodology elements. Some of these elements, called here templates, are designed to be instantiated during the use of the methodology in specific projects, while others, called resources, are intended to be used directly. Theoretical and practical implications of this division, especially regarding metamodelling and the use of powertypes, are explored. The proposed metamodelling approach is shown to facilitate the precise and complete specification of comprehensive methodologies, establishing the foundations for predictable and repeatable results from software development.

A Metamodel for Assessable Software Development Methodologies

Software development methodologies usually contain guidance on what steps to follow in order to obtain the desired product. At the same time, capability assessment frameworks usually assess the process that is followed on a project in practice in the context of a process reference model, defined separately and independently of any particular methodology. This results in the need for extra effort when trying to match a given process reference model with an organisation’s enacted processes. This paper introduces a metamodel for the definition of assessable methodologies, that is, methodologies that are constructed with assessment in mind and that contain a built-in process reference model. Organisations using methodologies built from this metamodel will benefit from automatically ensuring that their executed work conforms to the appropriate assessment model.