<p>This thesis have two tasks: one is to help in the development of the ArchWiztool at ABB and the other is to find a way of how to formalise security related architecturalresponsibilities in a general way so that they can be incorporated in the tool. This thesis report focuses on thesoftware quality attribute security.The ArchWiz tool is developed at ABB Corporate Research in Västerås. The scopeof the tool is to serve as a software architecture guidance and knowledge tool for all softwarearchitecture professionals within ABB. The ArchWiz tool is a way of helping the architects of ABB to matchtheir product requirements with a list of general reusable requirements. The matched productrequirements can then use the reusable requirement's architectural solutions, which are also presentedin the tool. The tool focuses on usability, security and safety in this first version but it isconstructed so that the user can add their own general requirements regarding any quality.The architecture design phase in the development of a software system is a keypart in the development process, it gives the first design decisions and gives informationon if the system will have potential to meet its key requirements. Security is a software quality thathas grown in importance for the architectural design of the system. There exist a number of potentialthreats and attacks that might breach the security of the software and these threats needs to beprotected against. The ArchWiz project closed in December 2009 and at that time the tool was notfinished. A good foundation and a GUI framework for further implementations were developed but toget a fully functioning tool more implementations need to be made. Security is one of thequality attributes, which the ArchWiz is supposed to support. Suggestions on how to formalise securityresponsibilities in the tool have in this thesis been created and analysed. However, the suggestionsneed to be in incorporated in the tool and tested through users tests with the ABB architects.With the user tests as basis, the best suggestion can be selected.</p> / Architectural Wizard project

Thesis (Ph. D.)--West Virginia University, 2006. / Title from document title page. Document formatted into pages; contains xiii, 166 p. : ill. (some col.). Includes abstract. Includes bibliographical references (p. 130-140).

Thesis (M.Sc.)--York University, 2004. Graduate Programme in Computer Science. / Typescript. Includes bibliographical references (leaves 130-134). Also available on the Internet. MODE OF ACCESS via web browser by entering the following URL:http://gateway.proquest.com/openurl?url%5Fver=Z39.88-2004&res%5Fdat=xri:pqdiss&rft%5Fval%5Ffmt=info:ofi/fmt:kev:mtx:dissertation&rft%5Fdat=xri:pqdiss:MQ99311

In this thesis we propose a formal framework for specifying and validating properties of software system architectures. The framework is founded on a model of software architecture description languages (ADLs) and uses a theorem-proving based approach to formally and mechanically establish properties of architectures. Our approach allows models defined using existing ADLs to be validated against properties that may not be expressible using the original notation and tool-set. The central component of the framework is a conceptual model of architecture description languages. The model formalizes a salient, shared set of design categories, relationships and constraints that are fundamental to these notations. An advantage of an approach based on a conceptual model is that it provides a uniform view of design information across a selection of languages. This allows us to construct alternate formal representations of design information specified using existing ADLs. These representations can then be mechanically validated to ensure they meet their specific formal requirements. After defining the model we embed it in the logic of the PVS theorem-proving environment and illustrate its utility with a case study. We first demonstrate how the elements of a design are specified using the model, and then show how this representation is validated using machine-assisted proof. Our approach allows the correctness of a design to be established against a wide range of properties. We illustrate with structural properties, behavioural properties, relationships between the structural and behavioural specification, and dynamic, or evolving aspects of a system's topology.

Open source software is designed, developed and maintained by means of electronic media. These media include discussions on a variety of issues reflecting the evolution of a software system, such as reports on bugs and their fixes, new feature requests, design change, refactoring tasks, test plans, etc. Often this valuable information is simply buried as plain text in the mailing archives. We believe that email interactions collected prior to a product release are related to its source code modifications, or if they do not immediately correlate to change events of the current release, they might affect changes happening in future revisions. In this work, we propose a method to reason about the nature of software changes by mining and correlating electronic mailing list archives. Our approach is based on the assumption that developers use meaningful names and their domain knowledge in defining source code identifiers, such as classes and methods. We employ natural language processing techniques to find similarity between source code change history and history of public interactions surrounding these changes. Exact string matching is applied to find a set of common concepts between discussion vocabulary and changed code vocabulary. We apply our correlation method on two software systems, LSEdit and Apache Ant. The results of these exploratory case studies demonstrate the evidence of similarity between the content of free-form text emails among developers and the actual modifications in the code. We identify a set of correlation patterns between discussion and changed code vocabularies and discover that some releases referred to as minor should instead fall under the major category. These patterns can be used to give estimations about the type of a change and time needed to implement it.

In this thesis we propose a formal framework for specifying and validating properties of software system architectures. The framework is founded on a model of software architecture description languages (ADLs) and uses a theorem-proving based approach to formally and mechanically establish properties of architectures. Our approach allows models defined using existing ADLs to be validated against properties that may not be expressible using the original notation and tool-set. The central component of the framework is a conceptual model of architecture description languages. The model formalizes a salient, shared set of design categories, relationships and constraints that are fundamental to these notations. An advantage of an approach based on a conceptual model is that it provides a uniform view of design information across a selection of languages. This allows us to construct alternate formal representations of design information specified using existing ADLs. These representations can then be mechanically validated to ensure they meet their specific formal requirements. After defining the model we embed it in the logic of the PVS theorem-proving environment and illustrate its utility with a case study. We first demonstrate how the elements of a design are specified using the model, and then show how this representation is validated using machine-assisted proof. Our approach allows the correctness of a design to be established against a wide range of properties. We illustrate with structural properties, behavioural properties, relationships between the structural and behavioural specification, and dynamic, or evolving aspects of a system's topology.

Open source software is designed, developed and maintained by means of electronic media. These media include discussions on a variety of issues reflecting the evolution of a software system, such as reports on bugs and their fixes, new feature requests, design change, refactoring tasks, test plans, etc. Often this valuable information is simply buried as plain text in the mailing archives. We believe that email interactions collected prior to a product release are related to its source code modifications, or if they do not immediately correlate to change events of the current release, they might affect changes happening in future revisions. In this work, we propose a method to reason about the nature of software changes by mining and correlating electronic mailing list archives. Our approach is based on the assumption that developers use meaningful names and their domain knowledge in defining source code identifiers, such as classes and methods. We employ natural language processing techniques to find similarity between source code change history and history of public interactions surrounding these changes. Exact string matching is applied to find a set of common concepts between discussion vocabulary and changed code vocabulary. We apply our correlation method on two software systems, LSEdit and Apache Ant. The results of these exploratory case studies demonstrate the evidence of similarity between the content of free-form text emails among developers and the actual modifications in the code. We identify a set of correlation patterns between discussion and changed code vocabularies and discover that some releases referred to as minor should instead fall under the major category. These patterns can be used to give estimations about the type of a change and time needed to implement it.

Web-based user interfaces are used widely. They are replacing conventional desktop-based user interfaces in many domains and are emerging as front-ends for online businesses. The technologies for web user interfaces have advanced considerably to support high-quality user interfaces. However, the usability of web interfaces continues to be an issue. We still encounter web forms where basic interactive features are missing or work unexpectedly. User interface is a costly and error-prone area of software construction. This is particularly true for web user interfaces. They are typically implemented with fewer reusable components on programmers' toolboxes than conventional user interfaces built using user interface frameworks such as Windows Forms, Cocoa, and Qt. Consequently, web interface programmers tend to struggle with low productivity, or low quality and high defect rates. This thesis focuses on property models, a declarative approach to programming user interfaces. In this approach, common user interface behaviors are automatically derived from the specifications of the data manipulated by user interfaces. The approach aims to reuse user interface algorithms that are common across interfaces and allow the programmers to focus on application-specific concerns. This thesis work is a part of project "hotdrink," a JavaScript implementation of the property model system, which has the goal of providing the benefits of property models for web interfaces. This thesis builds on previous work on property models, and adds to it three reusable help and convenience features, which can be especially useful for web forms. In particular, this thesis describes the generic mechanisms of the following user interface features: (1) validating data coming from a user and presenting useful messages that help the user to fix errors, (2) controlling the flow of data through "pinning," and (3) canceling the user's previous actions through undoing. The main contributions of the thesis are the mechanisms and the software architecture that enable implementing these behaviors in a reusable manner. This thesis also presents several examples to illustrate the benefits of the proposed mechanisms.

Implementing information system within an enterprise is supposed to integrate the business vision and goal sufficiently. Failing to do so shall waste a lot of information system resources. In this study, we utilize an Architectural Integrated Business Process (AIBP) approach that makes use of software architecture on both business process and information system. This approach refers business process when doing analysis and design of information system. AIBP approach is likely favored by corporate executives, thenceforth help gain the most profit for the information system within an enterprise. AIBP approach uses software architecture to describe business process, and information system as well. There are two major factors in software architecture. The first one is the structure and the second one is the behavior. Structure is used to describe the elements of a system, while behavior is used to describe the interaction among those elements. Both structure and behavior will be formed in the same hierarchy, also called the architecture hierarchy. Once an architectural integrated business process is specified, the following implementation of information system becomes an easy task. The architecture hierarchy will guide the decomposition of information system correctly. AIBP also offers great amount of knowledge on the development of software components.

Precise specification of the architecture and design of software is a good practice. Such specifications contain a lot of information about the software that can potentially be exploited by tools, to reduce redundancy in software writing by automating routine tasks, as well as giving valuable feedback on the software. We first broadly classify such tools based on their application, and look at some of the issues involved in writing them. Most of these tools are built using conventional programming languages that do not provide primitives that are routinely required by such tools. We identify some patterns that are seen while writing such tools and abstract them out into a language specially designed for the purpose. Our effort is a language called LEADS, where we attempt to make writing such tools a lot easier, modular, organized and portable. LEADS is based on the Pattern-Action approach, where one specifies the pattern of information of interest and the actions to be taken when it is found. The patterns are specified using a custom declarative language that is enriched with special primitives for routinely occurring constructs. LEADS also decouples itself from the specification environments and formats thereby ensuring wide applicability. This enables LEADS scripts to be ported across design environments with minimal effort.