A sentient environment is one that possesses a sensor-supported, dynamic, and self-updating internal representation of its own context. Wireless sensors could be integrated as a network to create a Sensor-Based Aware Environment (SBAE). The integration between Building Information Modelling (BIM) and the SBAE has the potential of a solution to support the sensing and responding to activities within the built environment such as intelligent facilities management. Two of the reasons of creating an aware built environment are to keep human error minimal and speed up the process of collecting, sharing and visualizing the data of the context. The aim of this research is to establish a computational “self-organizing” model of an aware built environment that maintain a data representation of the ambient environment and building structure through the integration of models of machine awareness, pervasive computing systems (wireless sensor networks, embedded computing) and object repositories of building data.

Large design projects, such as those in the AEC domain, involve collaboration among a number of design disciplines, often in separate locations. With the increase in CAD usage in design offices, there has been an increase in the interest in collaboration using the electronic medium, both synchronously and asynchronously. The use of a single shared database representing a single model of a building has been widely put forward but this paper argues that this does not take into account the different representations required by each discipline. This paper puts forward an environment which provides real-time multi-user collaboration in a 3D virtual world for designers in different locations. Agent technology is used to manage the different views, creation and modifications of objects in the 3D virtual world and the necessary relationships with the database(s) belonging to each discipline.

Tabletop systems provide a platform for developing novel interaction systems, including tangible user interfaces (TUIs). This research presents a study of the effects of tangible user interfaces (TUIs) on designers' spatial cognition and design communication. In devising an experiment that can highlight the impact on spatial cognition while using TUIs, we compared designers using a tangible user interface (TUI) on a tabletop system to designers using a graphical user interface (GUI) on a typical desktop computer with mouse and keyboard. The designers were given a configuration design task in which they manipulated 3D objects to meet design specifications. Our preliminary findings are that designers using the tabletop system with TUIs reasoned about spatial relationship among 3D objects and discovered unexpected spatial relationships, while the designers using the traditional keyboard and mouse interfaces reasoned about individual 3D objects.

Legislation requires the construction industry to check building designs for compliance against numerous building codes. This task is complex and failure to correctly assess designs for compliance can result in high, long-term costs. This presentation describes an automated code checking system, DesignCheck, that enables the modelling of extended design information and encoding of a wide range of domain knowledge. It uses Industry Foundation Classes (IFCs) as an intermediate model for translating 3D object-based CAD models into the Code Check Internal Model. Building codes are interpreted and encoded into the object-based rules which interface to the Code Check Internal Model. Geometry engine and semantics interpretation are used in an object-based database to support design performance verification. The first code to be implemented is the disabled access code, since access provisions have a high risk or cost of failure.

Jonathan Borg (University of Malta)
Amaresh Charabarti (Indian Institute of Science)
Tom Kvan (University of Hong Kong)
Larry Leifer (Stanford University)
Ian Parmee (University of West England)

Impossible Geographies 01: Memory by Petra Gemeinboeck with Mary Agnes Krell
This new interactive installation dynamically traces visitor's actions, mixing them in unexpected ways with memories held and stolen by the physical space. Throughout the exhibition, those memories of visitors and actions seep into the environment, creating a virtually woven fabric of events that grows and evolves over time. The piece uses memory as a metaphor for the fluid boundaries between the physical and the virtual. It consists of a series of implied and shifting geographies, signified only by beams of light. When crossed, visitors interrupt the space and trigger a kind of evolutionary storytelling. As a result, the space becomes infused with the presence of past events, making slippery the relationship between the fictive and the “real” memory. Impossible Geographies is, in a way, a system that exposes, invents and mixes layers of reality beneath the surface of urban spaces.

Diffusion
Thomas Lorenz is an architect and a lecturer at the faculty of Architecture, Vienna University of Technology. Thomas will introduce the workshop theme he will teach at the University of Sydney and present a selection of student projects he has directed in cooperation with institutions and companies over the last five years. They have been shown at electronic art festivals, such as the Ars Electronica Festival in Linz, Austria and the Filmwinter in Stuttgart, Germany: ”Raumframes”, an international film competition on the subject of film and space, “Lspace” – video projections for the elevator floor of the Ars Electronica Center [www.double-happiness.co.uk/lspace] “Diffusion” - projects for the media facade of the Ars Electronica Center. [www.aec.at/en/archives/center_projekt_ausgabe.asp?iProjectID=12797]

When humans interact with other people in the surrounding environment we make use of implicit situational information. We naturally assume and interpret the context of the situation we are in and respond fittingly. Computers are not as good as humans in assuming situational information from their environment and in using it in interactions. They cannot easily take advantage of such information in a transparent way, and if they can they usually require that it is explicitly provided. In addition to being able to obtain the information, it must include some reasoning to process the information and to deduce the meaning. This is probably the most challenging problem, since reasoning is often indirect or deducible by combining different pieces of context information.The aim of this research is to develop a computational model for mobile computing with a focus on context awareness and situated reasoning. From the research I will build and test a computational model for the design and implementation of contextually aware situated reasoning system for a mobile computing environment. The model will demonstrate an application that can adapt to the quality of context information supplied by the location and spatial information systems, and then reason and represent this information so that it is useful to the user. Mobile computing devices that intelligently support user tasks by acting autonomously on behalf of users in these integrated environments will become increasingly useful and evident in the future.

Research in spatio-visual similarity in design is vital in advancing our understanding of the concept of similarity itself and the role it plays in designing so that design support tools can be developed to aid in reasoning. In the past 40 years, research within and outside the design domain has revealed the importance of salient shape and spatial features contained in 2D diagrams and, to a somewhat limited extent, the significance of contextual dependencies. However there is still little known about their role in design. To address this gap this research has undertaken both cognitive and computational investigations. The former investigations have provided empirical evidence that extends our current understanding of how designers classify design diagrams based on their perceived similarity. The later investigations have resulted in a computational model, Q-SOM, for measuring 2D diagram similarity using self-organising maps. The approach assumes similarity is related to the way information is processed and that the reporting of similarity judgements is a meta-cognitive process requiring explicit comparison of visual design information both prior and subsequent to processing by the system. A summary of this work will be presented to draw attention to diagrammatic reasoning in the design domain, where discussion will be aimed at exploring the different levels of spatio-visual information as well as the importance of relevance feedback. It is argued that both these aspects are required if formal comparisons are to produce meaningful (cognitively congruous) similarity assessments of complex 2D design diagrams.

Wayfinding is such a fundamental task both in physical and virtual worlds. However supporting wayfinding is difficult for the virtual worlds as many of the sensorial stimuli are absent in them compared to the physical environments. This task of wayfinding can be even more complex and challenging in the midst of changing environments. We have designed a conceptual wayfinding model for the swarm based wayfinding tool to support wayfinding in such dynamic 3D virtual worlds. The model comprises cognitive interface model and wayfinding swarm model. These are embodied in the virtual world and interact with the users to aid them to wayfind. In this seminar, various strategies for the swarm model and the development the cognitive interface agent are presented.

The empirical studies on visuo spatial working memory (VSWM) show that the capacity of the VSWM is limited when visuo-spatial tasks are done using imagery. Thus externalization is needed to off-load the visuo spatial working memory. For the same reason drawings and diagrams play an important role in designing. This paper presents the cognitive activity differences of three expert architects when they design in blindfolded and sketching conditions. It was observed that all participants’ overall cognitive activity in blindfolded condition dropped below their activity in sketching condition, approximately after 20 minutes during the timeline of the design sessions. This drop in performance can be explained by higher cognitive demands in blindfolded conditions. We concluded that sketching off-loads the VSWM.

Traditional approaches to agent design require that agents be pre-programmed with goals, rewards, consecutive learning episodes or examples of correct behaviour. This is undesirable in complex, dynamic environments where pre-programmed domain theories can only approximate the true state of the world and where pre-programmed goals can become obsolete. This research is drawing on theories from psychology and computer science to develop and evaluate a model for motivated agents that can develop competencies and adapt in a wide range of dynamic environments. Our approach introduces a model for motivated learning agents governed by a general, domain independent motivation process. Motivated learning agents can identify interesting events in their environment and use these as the basis for goals from which new
behaviours can be learned.

Design optimization is the selection of the best solution from alternative solutions to a given problem. Traditional forms of optimization require that all variables, constraints and functions be defined prior to the execution of the optimization process. Design, however, proceeds by a situated interaction between the design agency (human or computational) and the external environment, where the variables and goals keep transforming as a result of the interaction between the two, and where variables may be incompletely defined at the conceptual stages. This research is an attempt to explore whether a situated process may be developed for design optimization, where the design agent is able to reformulate the optimization problem depending on the interactions and new forms of structure and behavior which develop in the interactive process.
Emergence in design may be one way of acheiving the above objective. Emergent structures and behaviors, which are a result of the interactive process and are not put in intentionally by the agent or the environment, if interpreted and re-represented by the design agent may cause the introduction of new variables into the design process.
The main idea is that the design agent must learn, interpret and construct its knowledge base by interactive experiences, so that it may function in a robust manner in uncertain and incompletely informed design environments where everything is not defined a priori and goals and parameters develop as the designing proceeds.

A building can be viewed as a system which consists of a number of building constructs. When designers develop a building design project, different representations are used for different building constructs. The purpose for developing a building is for people usage and for goods storage. Compared with building constructs, which are static aspects of building, people and goods are dynamic aspects of buildings. The majority of current applications of computer aided building design are for the final stages of building design where precise data such as numerical data is needed.
We develop a qualitative and quantitative energy-based unified representation for building design called Archi Bond Graphs (ABGs) that can be applied for the static and dynamic aspects of buildings and can be applied in the conceptual and final building design stages. The development of ABGs is inspired by bond graphs, which are called regular bond graphs (RBGs) in our research. We develop bond graphs for multiple domains (MBGs) based on RBGs and focus on the domain of architecture. ABGs are developed for three major aspects of a building: as a collection of objects, as a collection for people and their goods, and as a container and transformer of processes.
ABGs are applied in representing disparate building constructs' topologies associated with their energy flow of a building. Currently, we are focusing on the development of qualitative Archi Bond Graphs (ABGs) for building design. QABGs draw on qualitative physics which utilize direct symbols to represent continuous properties of the system being represented. .

Complex design geometry is typically rationalised (made buildable) either before or after the actual form finding process. Next to pre-rationalisation and post-rationalisation, co-rationalisation has recently emerged as a new approach. This presentation discusses how co-rationalisation allows the integration of rationalisation and form finding into a single process. I use as a case study space frame structures resembling liquid foam similar to the structural system of the Beijing National Swimming Centre (PTW, ARUP and the China State Construction Engineering Corporation).

Bio: Thomas is a lecturer at the School of Design in Hong Kong and a PhD candidate at RMIT University. Most of his teaching and research is in the fields of CAAD, design systems, interaction and industrial design.

Design concepts which state commonalities of a design process are regarded to be formed as a consequence of the situatedness of designing. Concept formation, as a process of incremental unsupervised acquisition of categories and their intentional descriptions (or conceptual knowledge), is here depicted from a situated lens. This presentation also discusses how a situated agent model can wrap around a design optimization tool and construct concepts from interaction between the agent, the design problem and the use of the tool. We present our preliminary result of a prototype system.

Aware environments represent the next evolutionary development step in building and other industries. Such an environment supports wide range of applications like monitoring and evaluation of the building performance.
The facilities management domain requires new ways toward a comprehensive approach to facility state information that could make data available throughout the life cycle of a building. The advancement of wireless sensor technology and wide adaptation of the Building Information Modelling (BIM) could potentially suggest new type of applications.
Towards this end, wireless sensors could be integrated as a network into a built environment to create a Sensor-Based Aware Environment (SBAE). This supports the autonomous process of collecting, updating and visualizing the data of the context. The aim of this research is to establish a computational “self-configuring” model of a sensor-based aware built environment that maintains data representation of the ambient environment and building structure through the integration of models of machine awareness, pervasive computing systems (wireless sensor networks) and object repositories of building data.

Design teams' communication between team members witnesses the context of a team's activities in a project/product's design procedure. Mining documents generated by design teams and filtering the useful information would help to find out the problems under the table and assist manager to provide advices to improve teams' performance.The research aim is to develop a CL(Computational
Linguistics) system that can analyse how a design team appraise a design project through text based communication. This research will use state of the art incremental statistics machine learning methods and CL for analyzing documents generated by design teams in working context.

1. A FRAMEWORK FOR WEARABLE DISPLAYS IN TEAM-SPORT
Technology and visualization are constantly changing the face of team-sport. As these advance and become smaller, the application of wearable devices becomes more common. This research presents a framework for the implementation of wearable displays within team-sport, the design of which involves many design decisions within these two fields.

2. A.Q. MEASURING AESTHETIC QUALITY IN INFORMATION VISUALISATION
Considering aesthetics in the creation and assessment of information visualisations will improve overall user experience. This research proposes a model of aesthetics that is transformed into mathematical metrics which are applied to current information visualisation applications to objectively determine their aesthetic quality.

3. VISUALISING THE STATE OF THE WORLD: A CONCEPTUAL MODEL
The aim of this study is to form a conceptual model pertaining to the visualisation of data from The Sentient.

4. UNSUPERVISED LEARNING IN MULTI-AGENT SYSTEMS
Multi-agent systems comprised of identical interacting agents have many applications in complex distributed systems, but programming their behaviour in dynamic and sophisticated environments is difficult. Unsupervised learning techniques offer the potential for the directed exploration and discovery of possible behaviours without the need for explicit programming of every desired behaviour in advance.

As research on cognition and creativity has received increased attention worldwide through out the last two decades, there has been an emphasis on research that investigates how designers think and behave. These researches attempt to establish explanatory frameworks and empirical understanding about designing by studying the activities of designers. The belief is that by understanding and studying how designers behave while designing, an opportunity arises that allows us to understand design practice itself. The aim of this research is to adopt and adapt functional grammatical linguistic analysis technique to characterise how designers describe their design practice in their respective design text. By making use of formal, functional grammatical analysis, this research investigates specific linguistic codes and grammatical forms employed by designers and the way that they account for designing and the designed itself.

Design behavior is regarded as a “black-box” process. Some researchers attempt to understand this process in terms of cognitive science, such as visual perception, design knowledge, design strategy, and the process of sketching…etc. In terms of different cognitive behaviors, designers could be distinguished into three categories: expert, style, or creative designer. Some designers, who have been considered as expert designers, will transform some declarative design knowledge into a procedural one. By doing so, they could tackle design problem immediately and precisely when they front similar problems. On the other hand, some designers who could generate design solutions by using some distinguishable elements and features will be recognized as style designers. However, there are divergent ways to consider designers as creative ones in terms of personal design cognition. Simon purposed that before being recognized a creative designer, people should be considered as an expert designer; on the other hand, Gardner addressed that the creative behaviors could be found during the process of the development of expert designers.
In the aspect of cognitive studies, these designers, expert, style, and creative designers, have three sorts of divergent design characteristics, separately. These characteristics indicate that they utilized dissimilar design knowledge, design process, or ability of sketching to solve problems. Therefore, in terms of three categories of cognitive studies, this thesis is to investigate the development of creative designers, and to analyze the cognitive factors, which could be found in the design behavior of the creative designers.

Motivation and Adaptivity in Intelligent Environments

Recent developments in networked 3D virtual worlds and the proliferation of high bandwidth communications technology have the potential to transform the nature of distance collaboration in professional design.
There have been numerous developments in systems that support collaboration that have resulted in system architectures to support information sharing and remote communication. Whilst these initiatives have led to important advances in the enabling technologies required to support changes in global economic practices, there remains a gap in our understanding of the impact of the technologies on the working practices of the people who are the primary users of such systems. This presentation will highlight the characteristics of a high bandwidth environment that includes a 3D virtual world, sketching, video conference, and design object and relationship management. The environment has been used to study the different collaborative design behaviours in 2D sketching and 3D virtual worlds..

Hubscape is a spatial ambient visualization installation developed by a group of undergraduate students in a studio-led course unit. It uses standard home automation hardware and multiple multimedia projections to display real-time, abstract datasets in physical space. The system extracts electronic and sensorial data in real time to generate data-driven atmospheres in space through electrically controlled devices and visual projections. Hubscape is aesthetically integrated into a computer lab hub room’s architecture to unobtrusively reflect nearby electronic and human activities, such as network traffic, timetable information, and temperature, motion and sound measurements.

This presentation begins with an in-depth look at the ideas of situatedness and their origins. It brings together literature from many fields in an attempt to understand the motivations and significance of the situated view of cognition. The ideas are traced from the educational philsopher Dewey, through the ideas that were developed from his work, leading to the situated agents described by Gero and Smith. The treatment of concepts is recognised as an area that requires further analysis from a situated viewpoint, raising the idea of advancing from concepts as a means of categrisation to categorisation occurring as a result of construction. The concepts of symmetry and sustainability are used as examples of the complexities involved in treatment of concepts - how such concepts are used and how they are held. The ideas raised are then applied to the activity of search over data, developing a hypothesis for indexless search using constructions on the basis of concepts to search data where search terms are not explicit within the data. The presentation focusses on the background literature and forming a theoretical basis for future research directions.

This research proposes to study the effects of aesthetic value on user experience using information visualization as context. Through first categorically ranking different information visualization methods, an initial assessment of aesthetic value will be applied to individual elements that comprise the user experience. Focusing not on a result set based solely upon of time of task completion, but instead empirically investigating how factors of interpretation and recollection correlate with aesthetic value, as well as their effect on the greater perception of user experience. When a relationship between aesthetic and these defined elements of user experience are found, a framework of guidelines will result as to why aesthetic is an important quality within the field of information visualization..

Many design researchers have tried to characterize design through the different activities that the designers demonstrate, such as problem analysing, solution generation and evaluation of the solutions. Others have documented the way that various drawing tools support design. This study combines these approaches to understand design activities by exploring the interplay between designers' representations and their behaviour in a collaborative 3D virtual world. It explores case studies of how designers sketch and model collaboratively as they solve a conceptual design problem and how their representational activities are mapped with their design activities. The aim of the study is to develop an understanding of the role of different design environments in collaborative design, in particular, characterizing the differences while using digital environments and sketching in the conceptual design phase.

This seminar will look at a variety of evidence that gives us clues about the nature of design knowledge. It will explore whether there is something special and unusual about what Nigel Cross has called 'a designerly way of knowing'. The investigation will cover the characteristics of design problems on the one hand and of our knowledge memory systems on the other.
The seminar will rely on research currently being done at Sheffield and more general work going back over several decades. It will look at data from controlled experiments, from observations in the field and from interviews with outstanding and leading designers.

This seminar will look at the long history of design process maps. It will argue that while apparently logical hey are also often misleading and have not been generally helpful in the development of our understanding of design. It will ask why they have been so popular originally with researchers, clients and professional and industrial bodies. Alternatives ways of building design process models that may offer more future research possibilities will be introduced. The seminar will rely on both historical and contemporary work. Building on these arguments and some of the ideas discussed in seminar 1 it will try to arrive at a summary of current design process modeling.

coming soon...

coming soon...

Intuition plays a key role in designing. In contrast to reasoning in design, intuition is the process whereby design decisions are based on an 'instinctive knowing' for the right or preferred solution path.
In a typical design problem functional and aesthetic constraints must be considered. While numerous methods are available for computational reasoning about function, aesthetic decision-making is left largely to the designer's intuition. For complex problems, such as for spatial designing, aesthetic intuition must draw on an array of perceptual modalities e.g. visual, auditory, kinetic, tactual.
Computational models of aesthetic design processes typically consider spatial design a solely visual process. As a result, many CAAD tools place a large emphasis on supporting visual intuition but have limited capacity to support other modes of aesthetic intuition.
Initial experiments which explore the notion of multi-modal spatial designing will be presented.

Sonification is the process of representing non-auditory, non-speech visual and other abstract data in an auditory or bi-modal (audio-visual) format to enhance our understanding of the information. A great deal of the data in the scientific community is produced in the form of three dimensional vector fields. Although a great deal of work has gone into displaying these data sets, one of the fundamental problems is that they contain a vast amount of data and it is all too easy to overload the user visual system with too much information. The main goal of this project is to attempt to offload some of the data processing work onto the user auditory system in such a manner that the user can intuitively understand the data.
Understanding the way we listen leads to principles that facilitate intuitively understood sonification, and are helpful in workplace and ambient display situations.

This talk will introduce the ideas from a recently published book of the same title. It argues that our behaviour in space is part of a global language through which we communicate personality, intentions, possessions, attitudes, culture and values.
It deals with space not as an abstract formal and visual concept but as a human and social phenomenon. It represents the fruits of over thirty years of research into the psychology of how we interact with space in general and architectural space in particular.
It explores how we perceive space and how the particularities of our visual and perceptual systems constrain and control this process. It shows that many of the aspects of how we make sense of architectural space relate directly to the human cognitive system. It examines how space serves to provide for many of our most fundamental emotional needs.

In this seminar the development of a quantitative method to study team designing processes will be presented. Linkograhy - a graphical representation of the design moves - will be revisited; and a new way of interpreting the linkography using entropy from Shannon's information theory to compare design processes will be discussed. Hypothetical scenarios will be used to illustrate the feasibility. A case study (using some of the CRC project - Studying Designers in a Collaborative Environment - data) had been conducted and the results will be presented in this seminar.

Design optimization deals with the determination of the values of the variables, which describe the structure of a design, that maximize or minimize some set of expected behaviours. The structure variables themselves are all defined before the commencement of the optimization process as are the variables that describe the expected behaviours. The standard issues are how to represent a design through appropriate structure variables, how to represent the behaviours and their derivations from structure and how to produce an efficient and effective optimization procedure. In general, all the structure and behaviour variables have to be defined before the optimization procedure can commence.
Situated design computing deals with the interaction between computational design systems and their environments in such a way that the systems change and adapt through their interactions. Issues that situated design computing deal with include: how experience is developed and used, how representations are a function of interpretation rather than encoding, how where a system is when it does what it does affects its future behaviour.
This seminar examines how concepts from design computing can be used in design optimization and presents some early results of developing experience during design optimization. It concludes with a discussion of possible roles for this approach.

Sketching is essential in a design process, for providing a dialogue medium for designersthinking/externalizing process and for storing ideas and solutions. We questioned if a designer is able to start designing without sketching and end up with a reasonable design solution. Three architects are engaged in two separate design processes, one is the experiment condition where they were not allowed sketch, and the other, the control condition where they were allowed to sketch. In the experiment condition, architects were required to put on a blindfold and think aloud while designing, they were required to sketch what they held in their minds at the end of the session. Design processes are video-taped in both conditions. The sketch assessment results show that in both conditions the design outcomes fit in the given dimensions of the site, accommodate the space requirements and allow an effective use for the clients. Thus, when the participants were blindfolded, they were able to produce designs by using their cognitive resources to create and hold an internal representation of the design rather than by sketching. The results show that there is no significant difference between sketching and not sketching based on three levels of analysis; design outcome, cognitive activity and idea links.

While 3D virtual worlds draw on our cognitive abilities to navigate in 3D virtual worlds, the role of aural and visual cues in navigating 3D virtual worlds is not well understood. We have developed a study of people performing navigation tasks in a 3D virtual world to characterize the role of aural and visual cues, as well as to determine if these cues improve human navigation performance. The study draws on similar studies of navigation in the physical world to provide an empirical study methodology and a coding scheme for the data collected. An hypothesis of the study is that the combination of visual and aural cues supports navigation better than the use of only visual cues as navigation aids. Our results show that performance is improved, and that the addition of different kinds of aural cues can affect navigation behavior.

Landmarks play an inevitable role in human experience of space, spatial cognitive processes, and spatial (e.g., route) communication.
I will present formal models to identify features in databases that may be used as landmarks in computer generated route directions, and will also discuss their use in such route directions.

A building is a complex combination of elements and processes and can be viewed as a system. The building integrates a variety of constructs: spatial layout, circulation arrangement, and energy movement and distribution, including electricity, hydraulics, conditioned air, thermal energy, light energy, sound energy and communications. Each of these constructs is a subsystem of the building system. This research for developing a unified representation for building design is called Archi Bond Graphs (ABG). ABG, consisting of systematic graphical representation and mathematic equations, is a qualitative and quantitative energy-based representation which can represent a variety of building subsystems. It can be applied in the conceptual and final stages of building design to represent both static and dynamic aspects of buildings. In this seminar, the development of Archi Bond Graphs, from regular bond graphs (RBG) to bond graphs for multiple domains (MBG) and then ABG, will be briefly introduced. In addition, a preliminary ABG application will also be presented.

Developments in digital design workbenches that combine Augmented Reality (AR) systems and tangible user interfaces (TUIs) on a horizontal display surface provide a new kind of physical and digital environment for collaborative design. The combination of tangible interaction with AR display techniques change the dynamics of the collaboration and have an impact on the designers' perception of 3D models. We are studying the effects of TUIs on designers spatial cognition and design communication in order to identify how such tangible systems can be used to provide better support for collaborative design. Specifically, we are comparing tangible user interfaces (TUIs) with graphical user interfaces (GUIs) in a collaborative design task with a focus on characterising the impact these user interfaces have on spatial cognition.

A designer's cognitive activity doesn't necessarily stop when he or she stops designing. This research study considers problem solving outcomes that occur while designers are not intentionally designing, as well as when they are. Forty architects have been asked to describe how they design, their breakthroughs and unexpected discoveries, made throughout the conceptual design stage of a project. Their statements suggest that nearly all highly motivated skilled architects develop effective, individual methods of disengaging from currently unproductive designing when necessary to achieve conceptual breakthroughs.
Statements by the interviewees suggest that discoveries during, or just after times when they have not been actively designing, referred to as 'cold' discoveries, are more significant both in number and quality than is currently recognized. One interesting question is whether the experience and appreciation of 'cold' discoveries is related to how people design or to something else. The seminar will include discussion of factors that might influence variations in 'cold' discovery outcomes.

Wayfinding is a cognitive element of navigation that allows people to plan and to form strategies prior to making any movements. In 3D virtual worlds people still perform navigational tasks just as they do in the physical worlds. To support this task of wayfinding in 3D virtual worlds, many have employed uniquely recognisable landmarks, provided maps, included thumbnails of notable objects, and embed sounds. These attempts have succeed in varying degree in aiding people to wayfind. We propose a dynamic wayfinding tool using swarm creatures to provide more extensive set of options available for people to aid in their decision making. The tool hopes to provide dynamic trails leading to the desired locations, create rough map of the local area, and generate teleport/warp gates. The progress of the development of such a tool is presented.

Design optimization is concerned with identifying optimal solutions which meet design objectives while conforming to design constraints. This process involves a number of tasks that are both knowledge-intensive and error-prone. Most optimization tools focus on gathering a variety of mathematical programming algorithms and providing the means for the user to define design problems. Therefore, designers heavily rely on their experiences to select an appropriate optimization algorithm. The limitation of this manual process may result in sub-optimal design solutions and inefficient design. To improve the performance of these systems, knowledge-based design optimization systems have been studied with the aim of leverage design tasks which require human expertise. However, these knowledge-intensive programs or so called reflexive agents respond to environmental stimuli based solely on their pre-programmed responses, and keep repeating themselves irrespective of their interactions with the design environment. The knowledge and functions are encoded in a “hard-wired” manner during the development stage.
This seminar introduces an approach that makes a reflexive agent “situated”, so as to enable this agent to learn by its use. As a consequence, the above-mentioned knowledge is constructed and grounded from interactions. Via an implementation of the agent's experience, we explore the operational features of a constructive memory model in design optimization area. Some potential impacts in this field will also be discussed.

Gesture is physical communication that uses motions of the hands, fingers and arms, with the intent to convey meaning to a human being or to an entity. In the cognitive science domains, gesture has been demonstrated as an aid to visuospatial reasoning, information retrieval and as a mechanism for lightening cognitive load. This indicates that gesture might benefit the processes involved in designing, not as a translation device to depict thought in externalisations (e.g sketches), but directly through its proprioceptive and visuospatial qualities.
The proposed research involves experiments with designers in which gesture is investigated using a computational system of real time gesture capture and analysis, examining hand-gesture responses to sound. A deeper understanding of gestural characteristics, such as velocity, scope, motion and repetition leads to better understanding of spatio-temporal gestures used by designers in visual and conceptual activities. This knowledge can benefit software design (design tools), collaborative design, long-distance videoconference collaboration, augmented reality design environments and understanding the way in which designers think.

Sonification is an emerging field in Human Computer Interaction (HCI). An important goal of interface design is to provide an effective representation. Some dense sets of data benefit from auditory representation (sonification) because the time-based, linear representation of sound allows us to easily recognize trends, patterns, information clusters and recurrences of behavior. Sonification leads to an easier display of assimilation of the information, especially in visually overloaded workplaces. For sonification to be of optimal benefits, the rapid cognition, comprehension and assimilation of auditory information in a non-demanding, intuitive way are essential to its efficacy. Various types of information need to display clearly and unambiguously and some obstacles need to be addressed for a discriminate representation of a large data set.
It is a significant challenge to determine how to map data to an auditory display because of the lack of applied theories of mapping generation for a successful sonification.
This proposal investigates cognitive aspects of auditory perception to bridge the gap and challenge, and explores the sonification design guidelines based on this investigation. The guidelines intend to facilitate an intuitive and discriminated interpretation.

A large and increasing amount of research has been performed on agent-oriented approaches to design computation, much of which is described informally. Using a strong theoretical basis is vital for research and development in any field, and formalism is one such theoretical basis. We therefore present the beginnings of research to this end. The beginnings of work on formalising situated design agency will be presented. Our intentions regarding implementing and testing a corresponding design agent implementation platform will be described.

Recognition of emergent features during design heavily influences future design actions. Theoretical models of emergence show that the ability to perceive emergence can be facilitated by multiple representations. Re-representation is the process of creating alternate representations from existing representations. Psychologists have shown that humans can do this intuitively. The research has two aims: to explore ways that human abilities to re-represent could be better supported by computational design tools, and to understand the role of re-representation in self-learning. To achieve this, two models have been implemented.
The first looks at a means by which alternate representations can be computationally created from an initial representation. This is done through the organisation of sense data using neural networks and subsequent restructuring to produce multiple representations. An abstract computational framework for re-representation is presented as the basis for the model. The model is given an initial representation in the form of a design sketch and produces graphical output of multiple representations. This leads to discussion about the way that human re-representation could be supported in existing design tools through similar methods.
The second model explores the role of re-representation in a situated concept formation agent. This model is an analogy for the open-ended visual world that humans create for themselves through an ability to create concepts based upon recognised emergent features. The agent is given an initial design sketch and it goes through an iterative process of perception, conception and action, where action consists of drawing using the concepts it has access to. Re-representation occurs prior to perception, and in this way the agent is able to build up a hierarchy of concepts. Experiments with the agent demonstrate that an ability to self-learn using existing knowledge can be facilitated by the ability to re-represent.

"Influencing Discussion Participation Using an Ambient Display" (Shun Takahashi)
Discussion is one of the most known collaboration techniques in the world. In organizations, discussion is used for multiple purposes and group decision making is one of them. Group decision making is the process of arriving at a judgment based upon the feedback of multiple individuals and it is a key component to the functioning of an organization since organizational performance involves more than just individual action. One of the advantages in such collaboration is that the group can share different viewpoints. If group participants produce more ideas or viewpoints for the topic, there is a high chance that they come up with a solid decision. An equal participation is a key to assist producing more ideas. There are three participants in discussions: over-participants, participants and under-participants. Notifying participants about their contributions to the discussion could give them an idea of how they are doing in comparison to the others. This could motivate over-participants to speak more and under-participants to speak less. It is ideal to notify the current participation situation unobtrusively since the distraction could break a flow of discussion and it may direct them to a faulty decision. An ambient display is the ideal way to convey the information as it communicates on the periphery of human perception unobtrusively, requiring minimal attention and cognitive load.

"Plastic Surgery Techniques for Evolutionary Design Systems" (Nicholas Preema)
One problem with evolutionary design systems is that they often converge at very good solutions that are still somewhat short of the ideal. In many cases, continuing the evolutionary process may at best take an unreasonable amount of time to achieve some small improvements, or at worse result in no apparent increase in fitness. This research will develop a new genetic operator, one which will attempt to use domain knowledge to directly modify design genotypes. In keeping with the tradition in evolutionary computing of drawing analogies from the natural world, this process is likened to the practice of humans of performing plastic surgery. While the genes of people set the form from the inside and cannot be changed readily, plastic surgery can be used change a person's form directly from the outside, with the idea being to increase a person's visual "fitness".
Thus it is hoped that, given an encoded design genotype created through the evolutionary process, by using domain knowledge, a kind of guided mutation is applied directly to the design phenotype in order to make a small jump in the design space to another point with a slightly higher fitness. When to perform this plastic surgery, and what kind of domain knowledge should be used will be discussed. Finally, two design problems that will be implemented used a plastic surgery-enabled evolutionary design system will be outlined.

"Using formal computable linguistic analysis techniques to expose the way designers develop concepts" (Nora Shaheed)
In the last decade, research on the human behaviour of creativity has amplified. Creative thinking skills are essential to assist the process of problem solving, as generally, knowledge alone is not enough to reach an innovative solution. One of the traditional means to study creative thinking patterns is through verbal protocol analysis. How can language in design be used to portray designers’ process and design cognition? Language analysis techniques provide both a concrete platform in which designers’ process can be studied and also presenting oppurtunities to develop computational linguistic tools. The aim of this research is to apply formal computable linguistic analysis techniques to expose the semantic
(lexical) and grammatical (systemic functional) structures by which designers interrelate and develop concepts.

"Motivated Learning Agents as a Basis for Intelligent Environments" (Owen Macindoe)
Developing an Intelligent Environment (IE) that can adapt to changes in the way that people use them has been a stated goal of IE researchers from the outset of work in the field, yet current implementations of IEs lack this adaptivity. In addition, adding new capabilities to current IEs via the installation of new sensors and effectors requires extensive manual configuration. Recently agents with a motivated learning model have been used with great success to adapt to new environments and acquire new capabilities in a self-directed manner. We propose to bring the motivated learning agent model to bear on the problem of developing IEs that can adapt dynamically to changes in their configuration and in the way the people use them.