Thanks to miniaturization of display technologies, the recent years have seen the emergence of a new generation of mobile devices containing multiple displays. They are identified as Mobile Multi-Display Environments (MMDEs), building on previous work in the field of Multi-Display Environments. This doctoral work presents the first exploration and classification of this research space. In particular, I identify the case of projection-screen MMDEs, mobile devices containing both projection and screen technologies. The dissertation address the following thesis: Providing re-configurability of displays' relative placements in the heterogeneous MMDE and providing interaction using kinaesthetic cues and spatial memory, users can manage complex and highly cognitively charged tasks as well as complex information management across multiple displays. To support this thesis, the dissertation answers research questions around the possibility of synchronous use of the displays given their inherent technological and physical disparities; the optimal relative positioning of the displays; the use of the mobile-projection unit as a secondary display and the projection spaces available around the user. The contributions of this work are multiple, the main contributions are: case studies evaluating and demonstrating the usefulness of synchronous use of the multiple displays; design guidelines for MMDEs; novel interaction techniques and scenario of use; a mathematical model of perceived depth in the mobile environment; and a series of prototypes and experiments that have been designed to support this work. The dissertation shows that multiple displays can be used synchronously to improve the device's capabilities in heterogeneous projection-screen MMDEs so users can perform complex tasks across both displays. This dissertation's vision is that mobile devices with multiple displays can become as useful and as widespread as their fixed MDEs. The results presented in this dissertation further the knowledge of MMDEs.

In the past decade, there has been significant interest in display technologies suitable for ubiquitous and ambient computing applications. Computer-augmented LED devices termed smart pixels have been gaining popularity in this research area as well as for commercial displays or architectural lighting. Compared with conventional display devices, such as LCD screens, smart pixels are flexible and unobtrusive. They may be deployed in locations where conventional displays would be infeasible and in irregular geometric arrangements, including 3D. However, up until now, smart pixels have been high complexity or limited in scalability or usability. This thesis argues for suitable requirements for a smart pixel technology which is scalable, usable, and low-complexity. A general architecture fulfilling these requirements is proposed. This architecture includes a wired network hierarchy for distributing power and data to pixels using power-line communication techniques, a location algorithm using visible-light communications and multi-view reconstruction to create an accurate 3D model of pixel positions, and a display representation suitable for allowing 2D or 3D, bitmap or vector graphic content to be rendered to a smart pixel display. An industry sponsored implementation of the architecture, Firefly, is described and used to evaluate the suitability of the proposed architecture. The major contributions of the work are the architecture itself, two novel power-line communication protocols which allow for dynamic addressing of pixels and are intended for use with high-power digital applications, and a 3D location algorithm using LEDs and cameras. The thesis concludes with a detailed summary of these contributions and proposes future work.

The current Internet architecture is facing serious scaling problems notably in its routing and addressing system. The TCPIIP stack uses the IP address to identify an endpoint host as well as to serve as a network topological locator. Due to the overloaded semantic of the IP address, mobility is not handled efficiently. Several Locator Identity Split proposals have been developed to decouple the actual semantic of the IP address, separating the endpoint identification and locator functions. For all Locator Identity Split protocols to work a mapping infrastructure is required. However most of the defined solutions tend to focus only on part of the problem space, and usually place end-host mobility aside. We believe that the mapping infrastructure would not scale efficiently with a large number of mobile nodes and would contribute to packet losses while the mobile nodes are moving. This thesis proposes a tiered architecture, which divides the mapping infrastructure into a core mapper and an internal mapper, such that it would scale under is a large number of mobile nodes, store and process their mapping records and provide an up-to-date mapping record. The tiered architecture is implemented using the GSE/8+8 rewriting approach. The core mapper uses a Chord DHT and the internal mapper is a hierarchy of routers with storing and rewriting capability. The tiered architecture is simulated in OMNeT++. The performance of the tiered architecture is evaluated by examining the core mapper and the internal mapper separately by simulating two real network scenarios based on the JANET topology and a University wireless network. The core mapper performance is evaluated against the Domain Name Server (DNS), and the internal mapper is evaluated using the end-host mobility provided by the Host based Identity Protocol (HIP).

This dissertation explores the relevance of computer science to physics. Beginning with a thorough technical analysis of the concept of information and theoretical computer science I distinguish between computation simpliciter and the narrower notion of digital computation which I define as symbol manipulation. I develop a detailed account of what it means to say a physical system implements, or carries out, a computation. I discuss the difference between analogue and digital computation and conclude it is a false dichotomy. A new category of device known as an experimental computer is proposed and distinguished from an analogue computer. I critique Geroch and Hartle's desideratum that all scientific theories be computable and I finish by looking at several attempts to define 'complexity' in computational terms.

In order to address anticipated high failure rates, reliability, availability and serviceability have become an urgent priority for next-generation high-performance computing (HPC) systems. This thesis aims to pave the way for highly available HPC systems by focusing on their most critical components and by reinforcing them with appropriate high availability solutions. Service components, such as head and service nodes, are the "Achilles heel" of a HPC system. A failure typically results in a complete system-wide outage. This thesis targets efficient software state replication mechanisms for service component redundancy to achieve high availability as well as high performance. Its methodology relies on defining a modern theoretical foundation for providing service- level high availability, identifying availability deficiencies of HPC systems, and comparing various service-level high availability methods. This thesis showcases several developed proof-of-concept prototypes providing high availability for services running on HPC head and service nodes using the symmetric active/ active replication method, i.e., state- machine replication, to complement prior work in this area using active/standby and asymmetric active/active configurations. Presented contributions include a generic taxonomy for service high availability, an insight into availability deficiencies of HPC systems, and a unified definition of service-level high availability methods. Further contributions encompass a fully functional symmetric active/active high availability prototype for a HPC job and resource management service that does not require modification of service, a fully functional symmetric active/active high availability prototype for a HPC parallel file system metadata service that offers high performance, and two preliminary prototypes for a transparent symmetric active/active replication software framework for client-service and dependent service scenarios that hide the replication infrastructure from clients and services. Assuming a mean-time to failure of 5,000 hours for a head or service node, all presented prototypes improve service availability from 99.285% to 99.995% in a two-node system, and to 99.99996% with three nodes.

I he flow and exchange of information is characteristic of the digital age. Information is increasingly consumed and produced. It has to be stored, channelled and processed. It also has to be reproduced in new forms again and again. The exponential growth in the volume of digital data has led to new challenges to visualize this data, such that it makes a significant difference to ones understanding of this data. This gain is known as information, which ultimately transforms into knowledge. Information Visualization attempts to create tools and processes that are an aid to cognition. Although this is a relatively new but established discipline, its roots can be traced to developments in the early 17th century, to what are now termed as Data Graphics. As a precursor to modern, dynamic, computer-based visualizations, Data Graphics underpin the science of visualization. This thesis looks at the design principles that govern the construction of historic and contemporary data displays. Although the medium on which Data Graphics have historically been constructed and presented has changed from paper to the computer, design principles have remained the same regardless of the medium or the source of content. The thesis then applies these principles to construct two applications based on large, complex and multivariate data-sets. The first is a proposal for a three-dimensional radar display to visualize Air Traffic Control data. The second application is a visual navigation tool to search within a hypertext document. Both case studies apply design principles inherent in data graphics and utilise human perceptual and cognitive abilities to extract information inherent in data. Finally, both applications are tested by conducting user studies. The contributions of the thesis lie in the construction of the two novel information visualization applications stated above, and by demonstrating that data graphic design principles transcend the medium in which they are produced and presented and can greatly enhance the beauty, efficacy and effectiveness of data displays.

Monitoring systems represent an important tool to support efforts aimed at improving productivity and quality, reducing waste and enhancing safety in manufacturing. Modern technologies including electronic devices, communication technology, the Internet, database systems and modern computer technology represent resources that can provide flexible and cost accessible attractive and efficient solutions for the implementation of distributed and intelligent monitoring systems. A new generation of microcontrollers offer a high level of integrated devices and operate at low power, making them the ideal choice for many embedded industrial applications. However, the development of application software for microcontroller- based implementations has normally been a restrictive factor. Before this work this has resulted in most process and condition monitoring systems being PC based. This research presents an intelligent and distributed monitoring system based on microcontroller technology, specifically the PIC18C452. The system uses a flexible architecture that can be adapted to the necessities of different monitoring applications. "Monitoring Modules" that can be deployed according to the application requirements were developed. Industrial networks and Internet technologies are employed to enhance communication, therefore allowing monitoring records to be made available in a remote database. The Petri-net concept is used to represent the monitoring task in such a way as to provide independence from the system's hardware and software. Extensions to the original Petri-net theory and new modelling elements, including the acquisition of analogue signals, required to support the use of this method in a microcontroller-based environment, are presented. These enhancements represent a major contribution of this research. Finally, the benefits of the system are considered by means of three application examples a simple Press Rig to illustrate the general features and use of the system, a more complicated Assembly Process Rig to show the flexibility of the modelling approach, and finally a CNC Milling Machine tool changer is used to demonstrate the system in a real manufacturing application.

The aim of this thesis is to compare the implementation of parameterisable LNS (logarithmic number system) and floating-point high dynamic range number systems on FPGA. The Virtex/Virtex-II range of FPGAs from Xilinx, which are the most popular FPGA technology, are used to implement the designs. The study focuses on using the low level primitives of the technology in an efficient way and so initially the design issues in implementing fixed-point operators are considered. The four basic operations of addition, multiplication, division and square root are considered. Carry- free adders, ripple-carry adders, parallel multipliers and digit recurrence division and square root are discussed. The floating-point operators use the word format and exceptions as described by the IEEE std-754. A dual-path adder implementation is described in detail, as are floating-point multiplier, divider and square root components. Results and comparisons with other works are given. The efficient implementation of function evaluation methods is considered next. An overview of current FPGA methods is given and a new piecewise polynomial implementation using the Taylor series is presented and compared with other designs in the literature. In the next section the LNS word format, accuracy and exceptions are described and two new LNS addition/subtraction function approximations are described. The algorithms for performing multiplication, division and powering in the LNS domain are also described and are compared with other designs in the open literature. Parameterisable conversion algorithms to convert to/from the fixed-point domain from/to the LNS and floating-point domain are described and implementation results given. In the next chapter MATLAB bit-true software models are given that have the exact functionality as the hardware models. The interfaces of the models are given and a serial communication system to perform low speed system tests is described. A comparison of the LNS and floating-point number systems in terms of area and delay is given. Different functions implemented in LNS and floating-point arithmetic are also compared and conclusions are drawn. The results show that when the LNS is implemented with a 6-bit or less characteristic it is superior to floating-point. However, for larger characteristic lengths the floating-point system is more efficient due to the delay and exponential area increase of the LNS addition operator. The LNS is beneficial for larger characteristics than 6-bits only for specialist applications that require a high portion of division, multiplication, square root, powering operations and few additions.

The work reported in this thesis is an attempt to apply integrated knowledge-based and adaptive hypermedia technologies in the area of electronic performance support. Moreover, this work is a contribution in the direction of "structured" hypermedia authoring of technical documentation. It tackles the main challenges associated with the systematic development of Web-based technical documentation which include the design, authoring, and implementation, and the creation of supporting CASE tools. The main contribution of this research is a systematic methodology for the development of hypermedia-based Performance Support Systems (PSSs) for the Web which adheres to the main characteristics of advanced PSSs. These characteristics are outlined in a conceptual model that complies with state-of-the-art technologies and current practices in the field of user performance support. First, the thesis suggests a conceptual model for advanced PSSs. These are characterised as mainly consisting of two loosely coupled components that are designed and accessed in a task-based and user-centred manner. The first component is a freely browsed technical documentation of the application domain. The second component is the expert advisor that provides assistance for more specific, complex, and difficult to learn tasks. The integrated technologies utilised in advanced PSSs include Web-based hypermedia and knowledge-based systems. Second, the thesis concentrates on the first component of advanced PSSs i.e. technical documentation. It suggests a usage-based data model for the design of technical documentation. The proposed model abstracts the intended purpose of the documentation, the tasks supported by the documentation, and the functional characteristics of documents. These abstractions are integrated in a usage-based semantic network where rules and valid relationships are identified. This design framework can then be used by authors in order to organise, generate, and maintain the technical documentation i.e. authoring. In addition, this model is also used to support a strategy for the adaptive retrieval of hypermedia documents. Third, the thesis suggests a model-driven hypermedia authoring approach for Web- based technical documentation. This approach utilises the usage-based data model for the design of technical documentation (described above). In addition, it complies with the principled guidelines of structured authoring. Finally, the thesis focuses on "intelligent" PSSs. It promotes the provision of intelligent performance support through the utilisation and integration of technologies used in developing knowledge-based diagnostic Expert Systems (ES) and adaptive hypermedia systems. This integration is implemented through the use of hypermedia which allows supporting content to be synchronized with the diagnostic ES inference process. The integrated adaptive diagnostic ES supports the user by providing what-to-do and how-to-do type of information tailored (adapted) to the user's knowledge of the subject domain. The special organisation of displays in an HTML-based user interface allows users, while employing the ES for fault diagnosis, to request detailed information about a certain diagnosis procedure, and then return to the ES to continue from where they left off. The solutions proposed in this thesis are demonstrated through the development of a prototype PSS for an all-terrain fork-lift truck. The performance support is provided through (i) a technical manual, (ii) a diagnostic ES for locating and correcting braking system faults, and (iii) an adaptive information retrieval utility.

A shared visual workspace and video in addition to voice are two functionalities or technologies which this thesis focuses on. What is clarified in this work is how these influence remote collaboration and conversational grounding in particular — where grounding refers to the pro-active process of seeking, creating and maintaining the shared meanings needed for conversational partners to communicate effectively. Additionally, this thesis clarifies how to support non-collocated synchronous mediated-collaboration around intelligence analytic tasks — away from traditional tasks that involve the identification or manipulation of physical objects which previous studies appear to favour. This research is guided by these three primary research questions: —RQ1) How can we expose aspects of conversational grounding in mediated communication involving different combinations of a video (showing a remote participant’s head and shoulder, and hands and work-area) and a fully shared visual workspace in addition to voice? —RQ2) In relation to the negotiated process of grounding, how can we explain what is happening when parties are collaborating on an intelligence task using a fully shared visual workspace? —RQ3) How can we design better fully shared visual workspace systems to support remote collaborative intelligence analysis tasks? Study1 — reported in Chapter 5, is an exploratory research which also serves as a groundwork for Study2. The findings there led to the formulation of more focused hypotheses later investigated in Study2. Further, the most significant contribution of the Study1 was the coding schema constructed for analysing the negotiation of common ground. Chapter 6, 7, 8 make up Study2. A human-participant experiment was conducted using a 2 x 2 factorial between-subjects design with 2-person teams and four media manipulations namely: video, no video, shared visual workspace and no shared visual workspace. Conversational grounding effort is operationalized as the number of repair-episodes per min (that is repair rate). Results here indicate that teams using shared visual workspace have a lower repair rate than those teams with no access to shared visual workspace. This result is statistically significant. Although teams using video equally had a lower repair rate than those teams not using video, this result was not statistically significant. This is consistent with prior research which found that a video showing a person’s face and shoulders is not terribly important in collaborative context. Results of another investigation demonstrate that regardless of the media condition, teams generally have a lower repair rate over time as the task progressed — this result was statistically significantly positive. Additionally, assessments of a questionnaire item measuring improvements of mutual agreements and shared understanding over time, showed a statistically significantly difference between the shared visual workspace group and the no shared visual workspace group, as was the participant’s rating of the effectiveness of the medium for information sharing. Results of a qualitative thematic analysis in Chapter 7 helps explain these statistical results and more. A conceptual process model of conversational grounding in shared visual workspace-mediated interaction is presented in Chapter 8. The model also summarises the research findings. The discourse there offer useful implications and guidelines for moving beyond current theories and models of the negotiation of common ground. Equally, practical design recommendations for the design of shared visual workspaces are also discussed there. Chapter 9, 10 reviews the research questions and considers how the research that has been presented addresses them, followed by a discussion of the contributions of the thesis, future work and conclusion. Overall, this thesis delivers the following contributions: —1) It advances existing knowledge silos and studies on media effects on conversational grounding — one of the ways it achieves that is by delivering a conceptual model framework for understanding conversational grounding processes in real-time remote collaborative intelligence analysis. —2) It delivers a new coding schema for the analysis of the negotiation of conversational grounding in remote work. —3) It offers four data-driven design recommendations for good practical design of shared visual workspace groupware that better support more natural communicative nuances.