Supporting Learning as an Iterative Process in a Social Context

This paper deals with today’s shortcomings of learning in the working environment, it discusses the state of the art in the literature, and introduces our ideas of supporting the learning on demand process in a social environment. The support provides opportunities for creating, editing and using memos or registered screen actions and exchanging them within a group of domain workers. We are currently developing a conceptual framework that will be implemented and evaluated in a realistic work setting. Current learning challenges in co-operative task accomplishment employing information technology in the working environment is not sufficiently supported neither by system introduction procedures (training) nor by technical system features (help facilities). With the system LEAR (Learner’s Living Repository), we propose a solution to support users in exploiting and exchanging learning and consultation episodes on group level: Users can describe problems they encountered or solutions they found when using the tool and accomplishing tasks by memos or by registered screen actions, comment on them, and store them in a personal ‘demotheque’. Users can send memos and clips that describe questions, problems with the tool, or breakdowns when using the tool as a request for off-line help to a consultant. Memos and clips that describe representative solutions can be made available to a group of users in a ‘purse for demos’. The Problem and the Aim Learning becomes an integrated part of life and an integrated part of work, too. Learning happens planned and unplanned, controlled and uncontrolled, consciously and unconsciously, single and collectively. Today’s working life and its widespread use of technology require more than ever to acquire permanently new domain and tool knowledge. System design also requires the feedback from users and their task performance during the design phase (i.e., by user participation see [Mambrey, Oppermann & Tepper, 1986]) and after the design phase (i.e., by usability evaluation see [Reiterer & Oppermann, 1994]). The optimal approach contains an analysis-design-evaluation-redesign process (see [Oppermann, 1994]). As the user’s task competence can dynamically be increased by a flexible work organisation and task support, the user’s tool competence should dynamically be increased by systems suitable for learning, exploration facilities tolerant for correction, and support environments reinforcing recapitulation and re-evaluation of problems and solutions. In this paper, we focus our view on the latter: to increase the tool competence of the user by strengthen the learning process and the reuse of already acquired knowledge in further working situations in a social context. Four aspects of learning will form the main focus of our work: • learning is ubiquitous, it has to be supported in every working situation, not only in particular learning phases or environments, • learning is a combination of exploration and instruction: people learn by trying things out and by consulting technical or human help facilities, • learning is an iterative phenomenon; it evolves step by step using early knowledge for later understanding • learning is an individual and a social activity: people learn on their own but they also appreciate the support and knowledge exchange in social interactions. Solutions and Deficiencies Training and learning on the job. We assume the learning process as being integrated into the task accomplishment [Dutke & Schönpflug, 1987, pp. 295f.; Paul, 1995, p. 168]. A substantial part of learning does not happen during the training but during task performance. Users explore the system in use and try out functions to reach their goals. “...people learn best when engrossed in the topic, motivated to seek out new knowledge and skills because they need them in order to solve the problem at hand” [Norman & Spohrer, 1996, 26]. “...information that is accessed but never put to use during the learning process may be difficult to retrieve and use when the need arises in the real world” [Schank & Kass, 1996, 28]. A ‘guided exploration’ facility was proposed to support this kind of learning [Carroll et al., 1987-1988; Carroll, 1990]. Guided exploration owes its origins in the concept of ‘discovery learning’ out of the late ‘60s and early ‘70s [Williams, 1992, p. 41]. Not any breakdown or new situation creates the need for acquiring new knowledge, i.e., to learn. Users in contrast do avoid learning. As Carroll and Rosson cite: “I want to do something, not learn to do everything” [Carroll et al., 1987-1988, p. 83]; they resume: “adults resist explicitly addressing themselves to new learning” [Carroll et al., 19871988, p. 101]; see also [Knowles, 1973; Kidd, 1977]. In particular, if the critical situation is supposed to occur only once the user is not motivated to learn a solution. It is sufficient if he or she is enabled to create the solution, for instance by the help of step by step instructions not meant to induce a knowledge acquisition with the user. Williams and Farkas give an example where a user who has exceptionally to produce a footnote instead of known endnotes for a particular journal will not accept the “compel (...) to ‘learn’ or ‘remember’ the procedures that he or she explicitly needs now in order to create the footnotes” [Williams, 1992, p. 44]. Only for recurrent problems and tasks new knowledge will be acquired. Support from on-line help. When problems arise, breakdowns occur or solutions are unknown, addressing the on-line-help is often insufficient for the user. The support users get from on-line help systems is restricted to the information that experts have brought into the system. Help from the system is restricted to information about system functionality and to well-known notorious problem situations [Fox, Grunst & Quast, 1994, pp. 186f.]. We only know one example that provides growing support based on questions of users and answers of consultants: ‘Answer Garden’, see [Ackerman & Malone, 1990]. On-line help support should be extendible to the user’s individual results of exploration based learning and to cooperative learning with consultants so as to integrate the learning results into technical support facilities (individualised help system). This individual help environment can be perceived as a user own created guiding solution in contrast to the ‘guided exploration’ manuals proposed by [Carroll, 1990] that was critiqued by [Williams, 1992, p. 49] for its inefficiency and ineffectiveness and its authoritarian nature. Consulting local or central experts. The learning process may occur individually where the user helps himor herself by exploration (trying things out) but often the user asks for help by consulting a competent colleague (‘poweruser’) in face-to-face interaction or by consulting an expert by telephone or remote diagnose. Learning supported by computer help and documentation without social support is not appreciated by many users. Users tend to prefer to “consult the ‘local expert’ or other users ... to translate their intentions into specific questions” [O’Malley, 1986, pp. 378f.; see also Brockmann, 1990 and Horton, 1990]. This consultation includes a constructive and co-operative communication between humans with complementary types of knowledge and expertise but being familiar with the same tasks and the same working environment, speaking the same jargon. Users are sometimes specialists themselves “assigned topics to master, and other users are made aware of when and whom to consult” [Carroll et al., 1987-1988, p. 85]. It is an illusion that users work alone with a system. “End users make good use of other people in their social environments to help them solve their computing problems and to compensate for gaps in their own knowledge of computers” [Nardi, 1993, pp. 104, 186]. Local experts can be enlarged by professionals with technical knowledge about the system in use but with less connection to the user community and the task at hand. The latter are less accessible for and less accepted by the users [Bannon, 1986, p. 406]. Computer experts or skilled domain workers cannot be strictly differentiated. Computer experts dispose of profound knowledge about information technology but only a thin spread of application or domain knowledge. Skilled domain workers dispose of profound knowledge about their technical domain but only of limited knowledge of information technology. Computer experts and domain experts (‘users’) are no homogeneous entities. Users are widely differentiated by novice and expert users. This distinction is insufficient in supposing a sudden leap from a novice to an expert. Most users will be positions in between as they have knowledge and experience in a limited area of an application and no or only little knowledge in the others. There will be a process of learning different areas of the application’s functionality, in particular with occasional or ‘discretionary users’ [Santhanam, 1993]. Communities of system users will emerge, in which individuals have different backgrounds of knowledge: substantial computer and substantial domain expertise distributed among different members of the community. The competence of the user groups together with the competence of professional system experts is the basis for their constructive interaction in problem solving. User support by personal interaction is limited by the capacity and availability of human experts. In particular in repeated situations of the same or a similar problem the consultation of a human expert confronts with restrictions: the user is ashamed to ask for the same help again and again and the expert pulls a long face over the same support demand. Personal interaction is also limited by the access of the consultant to the critical action episode of the user (the 1 ‘TeamInfo’ was developed as a shared repository for informal group-relevant information by [Berlin et al. 1993]. For producers of software, a ‘Living design Memory’ was proposed by [Terveen, Selfrigde & Long 1993]. problem or error situation). The error occurred before the consultant appears. The error or the problem cannot adequately be reconstructed by the user for the local expert and additionally not adequately be described for remote diagnoses. Exploratively acquired knowledge and solutions developed in consulting local experts or professionals are not reusable for the learner to exploit the substance when needed to solve a similar problem. In particular the way and the pitfalls of a solution are not available. Empirical studies show that users have problems with consultants and consultants have problems with their clients [Brezizinski, 1987; Liechti, 1988; Moning, 1993]. Consultants are overloaded; their increasing number is overcompensated by a yet increasing number of clients; members of the user service units show limited availability; they are often not interested in the needs of users; they ‘forget’ promises of problem solving that can‘t be executed immediately. Consultants have to solve (in their eyes) trivial problems and are therefore not motivated. User support is often organised on several levels [Brancheau, Vogel & Wetherbee, 1985] where the communication requires an exchange of problem and solution representations where verbal or written descriptions are expensive and misunderstandable. Our Approach: Learning as an Iterative Process Requires Support for the ReUse of Competence Iterative Learning: The learning process is iterative, i.e., the learner proceeds in his or her competence by several trials of acquisition and application of qualifications. The first trial to acquire knowledge may be (a) exploratory, (b) supported by technical or human consultants, (c) error prone, (d) with indirect solutions, and (e) with dead ends. The first step of learning provides the user with rudimentary knowledge about errors, risks, and solutions. Making only one experience is not sufficient for full an understanding and it is not robust to forgetting. It has to be reinforced and extended by re-use in later similar situations. The following figure shows the idea of the approach. facing a problem or an error exploring solutions finding a solution storing a solution annotating a solution t1 other business forget solution facing a similar problem remembering an earlier solution retrieving the solution applying the solution refining the solution