The objective of the tutorial is to introduce attendees to Activity Theory, a general theoretical framework for the analysis of human and communal action in the world. After an overview of the theory, focus shifts to how this framework can be utilized in practice. Some examples are shown of how this framework can provide a fresh perspective on certain extant problems in the fields of HCI and CSCW. Hopefully, participants become more aware of the nature and complexity of current controversies concerning the role of theory in the design of computer artifacts. By the end of the tutorial, participants should be able to understand the basic principles of the approach, and to describe their work activities in ways illuminated by this framework
Application units (AUs) are defined as the smallest self-contained units to be useful in the design and implementation of end-user tailorable applications, such as word processors, drawing programs, and e-mail systems. The notion of application units has been inspired by similar basic building blocks: Lego bricks, HyperCard buttons+scripts, MVC triads, and architectural patterns. It is proposed as a solution to problems in maintenance, reuse, and further development of generic software applications. An architecture for tailorable, generic applications is presented, application units are introduced, and an example of a generic application built out of AUs is given.
The Objectworks/Smalltalk programming environment offers a variety of interaction objects which allows programmers to design user interfaces for application. Programmers often lack the knowledge required to design good user interfaces. Beside the problem of consistency there are a lot of low level sources of errors, such as the size of buttons, the background and foreground colors, etc. User interfaces can be improved by providing this knowledge about design guidelines at this low level, thus giving programmers the opportunity to learn about the design of user interfaces CritiGUI is a knowledge-based critiquing system which supports programmers during the design process by providing this knowledge. It is an adaptable, after-task critiquing system which follows the analytical critiquing strategy.CritiGUI has knowledge about the design of menus, buttons, windows and especially about the use of colors, which is often based on personal preferences and false intuition rather than on ergonomic considerations. It can criticize single interaction objects as well as groups or the whole screen. In this paper the architecture of CritiGUI, its features and its use in the Objectworks/Smalltalk programming environment will be presented.
Transparency has been seen as a necessary aspect of successful human-computer interaction. In this paper we investigate this concept by looking at it from the point of view of activity theory. We show that transparency cannot be understood as a static feature of the interface, but that the crucial point in achieving transparent interaction is the ongoing development of unconscious operations, embedded in the process of use. We suggest that the process of deliberate formation and refinement of operations during the course of interaction, is supported by setting conditions for the creation of a zone of proximal development in the interface.
In this paper the authors make an attempt to describe some general approaches to generating meta-information structures and using them in hypermedia. They also present a solution of some analysis problems in hypermedia space by recursive methods of preliminary reduction. The reduction goal is a generation of macrostructure by typical mathematical approaches known as clasterization and optimization. The described approach is realised.
This paper addresses a new GUI architecture, called the shared interaction objects architecture, in which a GUI application consists of a storage of interaction objects implemented with a structured distributed shared memory, and several tasks sharing the storage. The main benefit of this architecture is that it offers an integrated framework of the implementation of several sophisticated interaction techniques such as multi-user interfaces, scripting, distributed GUI applications, gestures, and assistive technologies. Moreover, this paper reports that this architecture can be implemented with moderate performance by employing several implementation techniques such as R-trees, fully replicated coherence algorithm, multi-threading, and real-time scheduling.
The authors consider the situation in which a hypertext network grows due to a stream of new nodes. The network's evolution includes not only the growth of its node-link microstructure, but also the processes of becoming, maintaining, and collapsing unities of meaning in the population of nodes as well. The notion of hypertext macrodynamics is introduced to refer to these processes. In the authors' hypertext system (SMIsC), the unities, or coherence clusters, are system-detectable network regions, and the system provides users with nagivational assistance in building discourse-like trails. As the authors report, this system enabled them to imitate the microstructure growth caused by a long sequence of incoming nodes and to observe a real network's macrodynamics. Observed values included a node's “thematic potential” and a node's “Ioad/activity”, two indicators of a node's macrostate. Experimental data include two basic forms: time profiles of the network's “spatial sections” and spatial profiles of the network's “time sections”. The former are the two indicators' time series, while the latter are distributions of the two indicators' responses over populations of nodes. The data show that most nodes get individualized profiles in macrodynamics. The “macrohistory” trajectories of nodes range from very simple to very complicated. Many nodes tend to behave n a random, unpredictable way. In this connection, the authors refer to modern studies on chaotic systems. The authors believe that the possibility for users to control chaos in macrodynamics, i.e., in the creation of meaning, reveal a new kind of hypertext interactivity.
This paper describes an object-oriented approach to semantic modeling in hypermedia. A new object-oriented hypermedia data model, the HM Data Model, is used to illustrate the discussion. At the logical level, the HM Data Model defines data structures, operations on instances of these structures, and an additional dimension of navigation orthogonal to the traditional plane of link-based browsing. At the semantic level, the HM Data Model defines domain-specific categories, which form a conceptual schema for hypermedia data. Categories are purpose-oriented behavioural abstractions rather than purely structural abstractions as in other models. A category defines intelligent, domain-specific behaviour for instances of the category.
From a user-centered point of view, learning with hypermedia educational software is mainly a matter of navigation in a corpus of knowledge. However, although navigating obviously requires minimal competences in both computer interaction skills and domain knowledge, yet most current models of navigational support rely on a presupposed degree of proficiency in both domains. We suggest that every learner using a new navigational based learning system needs to be supported by an initial phase of orientation and initiation in both spaces of interface and domain contents. Activity metaphors seem to be a convenient way of supporting learner's cognitive transfer from familiar to unfamiliar domain in particular to help them structure their dynamic progress through the different phases of learning. We will refer to activity theory as a framework to structure both the support of the user navigation and the description of the learning process. As an overall navigational metaphor we will draw on a travel metaphor interpreted as a quest of objects by subjects (Greimas' semio-narrative theory). Considering that navigational issues in learning must evolve from a focus on navigation in the interface to navigation in the domain contents, we propose a three-phases navigational model specific for educational software. Each of the three phases call for different navigational support with Adaptive Hypermedia Systems (AHS) as one of them.
Diagrams (e.g., trees for hierarchical structures, or graphs for finite state machines) are often needed as part of advanceduser interfaces, and are frequently specific to a user's application. The implementation of editors for diagrams should be supported by a tool and based on a formal model. This paper gives an overview of DiaGen, our generator for diagram editors. An editor for a certain kind of diagram is generated from a specification, which includes a grammar to describe the structure of diagrams. The user of a diagram editor, however, does not have to be concerned with the grammar, but can manipulate diagrams very conveniently by direct manipulation.
This paper attempts to both describe and resolve some of the fundamental differences between the fields of industrial design, activity theory, and human-computer interaction. In particular, the role that social relationships play in using and learning to use artifacts to mediate activities is examined in detail. This examination leads to a unification of theory and practice in the three areas, providing a new perspective for the development of computer-based artifacts (particularly embedded computing systems). Two examples are presented to illustrate how this new perspective can be used both to understand and to guide the process of developing an artifact.
Current graphical user interface toolkits, by and large, provide widgets that have a static appearance. The focus of this paper is to outline some additions that we have made to a toolkit to support animated widgets, such as menus that open smoothly, buttons that bounce and slide when pushed, and animated labels.The animations that underly these new widgets rely on straightforward applications of well-known techniques for cartoon and computer animation. Our concern in this paper is to describe how we packaged the techniques into a form that application programmers could readily use. In particular, we describe extensions we made to the InterViews 3.1 toolkit.
We present a two-stage system for information exploration and retrieval. The first stage, named GALOIS, organizes the information contained into a database into a particular lattice structure; the second stage, named ULYSSES, is a visual interface to access the structure built in the earlier stage. In this paper we focus on the latter. ULYSSES is based on a tight integration of traditional and novel user interaction paradigms that can be seen as a search&bound approach to information retrieval. The user may search the retrieval space by browsing or querying, but he or she may also bound the retrieval space by specifying constraints that the information contained in it has to satisfy. These interaction modes can be naturally combined to produce a hybrid retrieval strategy that best reflects the user goals and his/her domain knowledge. The retrieval effectiveness of our system has been tested in an experiment on subject searching where it compared favourably with respect to a Boolean retrieval system.
The research is devoted to the problems of visual interfaces and hierarchical multilanguage technologies in HCI. The PYTHAGORAS software is based on graph representations and a hierarchical approach. The software may be used for the development of multienvironment software and as “hyper world” for students. Characteristic features of the PYTHAGORAS software include a preference for graphic descriptions with animation and the use of hierarchical multilanguage technology.
Programming by Demonstration, or PBD, is an exciting and developing branch of HCI research. With PBD techniques, end-users can add functionality to their environments without programming in the conventional sense. Virtually all research into PBD, however, presumes that the event history is a linear sequence of user actions. This paper challenges that notion by introducing Hierarchical Event Histories, a new approach which represents some of the end-user's task structure directly in the event history. PBD systems can then take advantage of this structure to operate more correctly and in more situations. To assist programmers in generating structured histories, we also present Hieractors, a new model that provides a simple and clear syntax for describing arbitrary, high-level application behaviors.
This paper describes the Retail User Assistant (RUA), a prototype system that provides continual adaptive help for the use of an existing retail point-of-sale device. A preliminary study compared performance of three groups: one using the retail device alone, one using the device with a nonadaptive version of the RUA, and one with the adaptive RUA. Trends in the results imply that performance support, both adaptive and nonadaptive, improved accuracy but slowed performance. The RUA's adaptation, which gave less detailed help as users learned the task, was readily accepted by users and seems to have improved users' performance times. This result is discussed in the context of previous research on adaptive help and adaptive interfaces.
Explanatory program visualization is a name for program visualization extended with natural language explanations. Explanatory visualization can seriously increase students' understanding of program behavior. This paper gives the rationale and background for explanatory visualization and introduces our work on using explanatory visualization in educational programming environments. In particular, we present first experimental results on using explanatory visualization and provide a fine-grained description of the implementation of adaptive explanatory visualization in our ITEM/IP-II system. This system employs student model to adapt the visualization to the student knowledge level.
An object oriented architecture for building user interfaces is presented. It is based on a finite state machine with a stack for translating system input events into application dependent events and on a “best fit” method of looking for an object to process an event. As a result, the user can more easily understand the implemented interface environment and the programmer can more simply code it.
The use of hypermedia in educational systems requires a new type of literacy from learners and authors. In this paper some issues of hypermediabased educational systems such as learner control, the database structuring and browsing tools will be discussed, which results in an architecture of an educational system. To build such an educational systems, the architecture of the authoring system Cadys is presented. For learning benefits an author who builds a hypermedia-based educational system has to observe other design criteria than in tutorial or instructional systems. Since authors usually are no design experts for hypermedia systems, the authoring system supports the author during the design process. A critic supervises this process in order to support the author.
This paper describes a method for designing a user interface from a variety of presentation elements (e.g., icon, menu, flowchart, bar chart, etc.). The interface is composed hierarchically, and at each step the most appropriate design alternatives (hereafter denoted by “As”) are selected, based on “hierarchical morphological multicriteria design” (HMMD). We will present a detailed example for data processing in a decision support system (DSS) for multicriteria ranking.
