1. What is the problem posed by the paper ?
Ans.1: The paper establishes the review of characteristics associated with the global knowledge based economy and makes strong suggestions with respect to the increasing need and demand to change in novelty adaption by workers, managers and organizations. The researcher emphasizes on the focus to design computer tools that foster continuous learning as the need to support adaptation and continuous learning will only increase with time. The potential programmatic approach has been defined as the cognitive work analysis framework as a solution to this practical design challenge.
The basic problem posed by the paper is the quest to design a framework as a computer based tool to help workers perform open ended intellectual tasks that require discretionary decision making skills (Vicente, 2000). The discretionary decision making tasks as researched by Vicente, 2000, require continuous learning and adaptation to novelty with respect to organisation and tasks. The problem posed by the paper revolves around the development and design of systems to support workers in complex sociotechnical systems. The problem discussed in the paper lays emphasis on the fact that complex sociotechnical systems tend to have many if not all of the below mentioned characteristics:
* Large problem spaces
* Social
* Heterogeneous perspectives
* Distributed
* Dynamic
* Potentially high hazards
* Many coupled subsystems
* Significant use of automation
* Uncertain data
* Mediated interaction via computers
* Disturbances
The problem posed by the research paper emphasizes over the fact that the HCI researchers and designer toolkits need to be improvised as per the work analysis techniques, suitable for complex sociotechnical systems because different type of problems require to be dealt with different type of solutions. CWA tries to support these demands by adopting a flexible constraint-based approach, by identifying five categories of goal-relevant constraints, by providing modeling tools that can be used to analyze each layer of constraint, and by linking each layer of constraint to a particular category of systems design interventions. The main problem posed by the paper is the development and design of a computer based tool that equips open ended intellectual task workers with flexible decision making capabilities.
2. What approach is proposed to solve the problem?
Ans. 2: The Cognitive work analysis (CWA [Rasmussen et al. 1994; Vicente 1999]) approach has been the central approach that has been proposed in the paper to solve the problem of designing a computer based tool that equips open ended intellectual task workers with flexible decision making capabilities. The CWA has been defined as a work analysis framework which can be used to support worker adaptation and continuous learning by creating computer based tools to support the same. CWA was developed since the 1960s by researchers at Ris� National Laboratory in Roskilde, Denmark (see Vicente [1998; 2000]. Following are the characteristics that make CWA framework, support the continuous learning objective.
* A Constraint-Based Approach
The primary reason to support the suitability of CWA as a work analysis framework that supports continuous learning and adaptation to novelty and change, is the fact that it is flexible enough to support variability in action that has sensitivity to local, contextual details that cannot be anticipated. Adopting a constraint based approach , CWA accomplishes the aforesaid objective. The generic illustration of the basic idea is given by the following figure:
The set of action possibilities is defined by the resulting constraint space. These possibilities represent the relevant degrees of freedom for productive action. However, the constraint boundaries, and thus the constraint space, will change as a function of the context. The designers can have successful generation of a set of information implied to design a computer based tool, by the identification of these constraints of goal-relevance. The designed and developed tool can be a helpful can be a helpful medium that provides knowledge workers with feedback and decision support as to the constraints that need to be respected, but would be unsuccessful in identifying a path or trajectory through the constraint space. It would be up to workers to decide which trajectory (i.e., set of actions) to take for a particular set of circumstances. This constraint-based approach thereby provides workers with continual opportunities for learning and the flexibility to adapt within the space of relevant action possibilities.
Thus, the constraint space provides the flexibility required to support situated action [Suchman 1987]. In addition, the same worker can choose different trajectories, even when the circumstances remain the same. As a result, the constraint space is also flexible enough to support the intrinsic variability frequently observed in human action. Finally, different workers can choose different trajectories to achieve the same outcome in different ways. Therefore, the constraint space is also flexible enough to support individual differences between workers.
Five Layers of Constraint
the constraint based approach of which the CWA framework is an example, comprises of the five layers of constraints. The first layer can be visualized as the work domain which is a map of the environment that needs to be acted upon. The set of control tasks representing the as to what is needed to be done to the work domain, forms the second layer of constraint. The set of strategies representing the different processes with the help of which actions can be carried out effectively, represents the third layer of constraint. The socio-organizational structure, representing the preceding set of demand allocation among actors along with the ideology behind productive organization and coordination of the actors among themselves, makes the fourth layer of constraint. The fifth and the final layer of constraint represents the capabilities required for success through the set of worker competencies. The following figure represents the approach diagrammatically:
The second phase, control task analysis, inherits the constraints of the first phase but adds additional constraints as well. Although the work domain provides a large number of total degrees of freedom, when actors are pursuing a particular control task, only a subset of those degrees of freedom are usually relevant. Furthermore, there are new constraints that must be respected above and beyond those imposed by the work domain. For some control tasks, it is important that certain actions be performed before others. This constraint is a property of the control task, not the work domain, the net result is a reduction in the relevant degrees of freedom. When workers are pursuing a particular control task, only certain actions are meaningful and require consideration.
The third phase of CWA is depicted by two hypothetical strategies, A and B, for control task 2 (strategies for control task 1 are not shown for the sake of clarity). The strategy phase also inherits the constraints associated with previous phases of analysis. After all, a strategy cannot make a work domain do something that it is not capable of doing. This is why the two sets for strategies A and B are subsets of the work domain set. In addition, a strategy must also work within the constraints associated with its corresponding control task; otherwise it will not reliably achieve the required task goals. It is for this reason that strategies A and B are nested inside the set for control task 2. The strategies phase also introduces new constraints of its own, however. The control task level merely identifies the degrees of freedom associated with achieving a particular goal.
The fourth phase of CWA, social-organizational analysis, follows a similar pattern. It too inherits the constraints imposed by previous phases of work analysis, and it too adds a new layer of constraint. There are multiple organizational structures that could conceivably be adopted for any one strategy. To take a very simple example, a strategy may be performed by one worker alone, by two workers in a collaborative manner, by one worker supervising another, or by a worker supervising automation. In each of these cases the same strategy is being adopted; the same control task is being pursued; and the same work domain is being acted upon. Nevertheless, these different organizational architectures have different constraints associated with them. A worker executing the strategy alone will likely draw on a different, although probably overlapping, set of relevant actions than two workers executing the strategy in a cooperative fashion.
Fifth phase of worker competencies reduces the degrees of freedom even further. There are certain things that people are simply not capable of doing. Consequently, particular ways of working are not feasible. For example, some activities require too much working memory load, too much time, too much knowledge, or too much computational effort for people to perform. These constraints are specifically associated with workers’ competencies, not with any of the other preceding phases of analysis alone. This final narrowing down of degrees of freedom by the worker competency set, which is nested within the social-organizational set for strategy A.
3. How convinced were you that the proposed solution would in fact help to cure the problem?
Ans.3: The proposed solution in the paper to solve dynamic intellectual decision making problems and issues is based on the flexibility of HCI to solve the problems. The unpredictability of future can be utilised potentially with the analysis presented in the article with a few things that appear to be indisputable. The paper points at the situation of near future when complex socio-technical systems become more open, change will become a norm, not an exception. Hence, to maintain the competitive edge in the global knowledge based economy, one will have to cater to the increasing demand for workers, managers and organisations regarding flexibility and adaptability, accompanying the need for learning to learn. This calls for the deliberate and systematic development and design of computer based tools to help and increase worker efficiency for the fulfilment of these challenging needs. With the identification of the valid goal relevant constraints, designers can succeed in the generation of sets of information requirements that can be used for the design and development of the required computer based tool. This constraint-based approach thereby provides workers with continual opportunities for learning and the flexibility to adapt within the space of relevant action possibilities. Workers may select one trajectory, whereas in another circumstance, they may have to select a different trajectory to achieve the same task goals. The constraint space is also flexible enough to support the intrinsic variability frequently observed in human action. Finally, different workers can choose different trajectories to achieve the same outcome in different ways. Therefore, the constraint space is also flexible enough to support individual differences between workers.
The intended CWA framework is a programmatic approach to this practical problem. However, regardless of whether or not CWA itself succeeds in achieving its objectives, HCI has to be concerned increasingly with designing systematically for effective worker adaptation. The CWA framework is intended to be a programmatic approach to one large class of HCI problems: how to design computer-based tools that help workers perform open-ended intellectual tasks that require discretionary decision making. These tasks require continuous learning and adaptation to novelty. The demands have been tried to be supported by CWA, by adopting a flexible constraint based approach that works by identification of five categories of constraints with respect to goal-relevance and also providing tools(modelling) that can be used to analyse each layer of constraints, by linking each layer of constraints to a particular category of systems design interventions. The CWA framework’s origin can be traced back to 1960’s when its research started at Ris� National Laboratory in Denmark. Moreover,CWA has been utilized for a wide variety of issues like system modelling, system design, training needs analysis, training program evaluation and design, interface design and evaluation, tender evaluation, team design and error management strategy design.
CWA’s applicability can also be gauged through its application in a number of important fields like aviation and air force, nuclear power, navy, military command and control, road transport, health care, air traffic control and manufacturing.
The above literature and evidence on the long standing research in CWA, its evolution over time, its utility and finally application in different fields illicit its wide spread usefulness. The above mentioned facts, conclusions and arguments convince me that CWA as a framework would provide efficiency base to HCI systems for workers to have a self-learning and adaptive decision making skill set and environment.
4. What did you understand by the concepts of task and strategies as described in the paper ?
Ans.4: The understanding of the concepts of tasks and strategies have been clearly defined, explained and illustrated in the second and third phase of the constraint based approach of the recommended CWA framework. The control task analysis phase is seen to inherit the constraints of the first phase but has added additional constraints as well. Two hypothetical control tasks have been illustrated by an example in the following figure:
However the work domain equips the framework with a large number of degrees of freedom, only a subset of those degrees of freedom are of usual reference, when actors pursue a control task in particular. The researcher clarifies this concept by giving the example of pilots while they navigate at a cruising altitude, constraints like landing gear and brakes are irrelevant. Also, the researcher highlights that new constraints become visible and must be respected beyond the ones imposed by the domain of work. For some control tasks, it is important that certain actions be performed before others. This constraint is a property of the control task, not the work domain.
The third phase of CWA emphasizes and depicts two hypothetical strategies A and B, for control task 2 (strategies for control task 1 are not shown for the sake of clarity).The strategy phase inherits the constraints associated with previous phases of analysis. This can be defined as the reason why the two sets for strategies A and B in Figure 2 are subsets of the work domain set.
The strategies phase also introduces new constraints of its own, however. The control task level merely identifies the degrees of freedom associated with achieving a particular goal. When a particular strategy for performing the control task is identified, some degrees of freedom are usually not relevant because they are only required for other strategies. A specific strategy imposes a certain flow or process that adds constraints on top of those that are imposed by merely achieving the desired outcome. It is for this reason that strategies A and B are nested within the set for control task 2 in Figure 2.
5. How valid are the paper’s conclusions?
Ans.5: The papers conclusion stands valid. The proposal of the CWA framework as a solution to a large class of HCI problems with the help of computer based tool design, would definitely help workers to perform open ended intellectual tasks that involve dynamic decision making. The case study of Hedge Funds, which are defined to be based on very complex computer-aided trading strategies were used as a fascinating example of HCI in complex socio-technical systems, can be restructured to explain the validity of the conclusion. To support HCI system CWA framework is programmed for one major class of HCI problems which helps in designing the tools which is computer originated to help workers, to perform open-ended intellectual tasks to make flexible decision making. These tasks need innovation in continuous learning and adaptation. CWA supported these requirements accepting a flexible constraint-based approach, by identifying five categories of goal-relevant constraints, by providing modelling tools that can be used to analyze each layer of constraint, and by linking each layer of constraint to a particular category of systems design interventions.
The researcher has raised a valid point that the worker, manager and organisation, all require continuous learning and adaptation to novelty. The proposal of the CWA framework as a potential solution to solve the problem a hand and has been supported by the validity of the argument that computerisation will continue to change the nature of work on a daily and strategic basis. “Technological change means change in job requirement, the ability to satisfy changing and not entirely predictable, job requirements in a complex, culturally diverse and constantly evolving environment will require a literate work force that has good problem solving skills and learning skills”. (Nickerson, 1995, p. 42).Adaptation is not only relevant to workers but to managers as well, the markets are particularly affected by the symptoms accompanying the trend toward a global knowledge-based economy, exhibiting a high degree of volatility and uncertainty. Successful managers adapted to all of this turbulent change to keep their companies competitive. However, as the trends identified by Brzustowski [1998] affect different sectors of the economy, we can expect to see the same strong need for managers to adapt to uncertainty and novelty in other industries as well. Even for the organisation it is very important to accept, embrace, and seek changes. This, in turn, would require a decentralized organizational structure whose individuals are committed to continuous learning and adaptation to novelty. Vicente (2000) argued on the need of adaptation as it may increase in future due to the trend towards global knowledge- based economy. Associated workers, managers, and organizations need to be much supple and adaptive than before due to changing and unpredictable situations of the global knowledge-based economy which will continue in transforming the background of new effort.
CWA tries to support these demands by adopting a flexible constraint based approach and by identifying five categories of goal relevant constraints, by providing modelling tools that can be used to analyze each layer of constraint and by linking each layer of constraint to a particular category of system design interventions.
6. What sort of evidence would improve the validity of the conclusions ?
Ans.6: The concluding findings of the paper prompt the organisations and designers towards the situation where they would have to manage the increasing demand for workers, managers and organisational spaces to be flexible and adaptive in order to maintain competitive edge in the market. Focussing on the case study to infer to the validity of the conclusion we found out that in 1998 , when the world was facing experiencing a financial setback, hedge funds were experiencing unprecedented losses. Hedge funds can be defined as a high tech form of financial investments which rely heavily, on computer models, to arrive at trading decisions. Historical patterns based on very large amount of data over many years of market behaviour are used to construct the aforesaid computer models.
When prices move outside of the normal relationship defined by referent, a signal is sent to make the trade, it is expected that prices will revert back to historical pattern. As an evidence to be provided against the validity of the conclusions stated by the paper, focussing on the quantitative statistical images , it is obvious that these investment strategies usually take advantage of very small price discrepancies, as a result enormous investments are required to generate significant returns. During the financial setback in 1998, hedge funds lost their value. One hedge fund lost 44% of its value (US $ 2 billion) in one month, Smith Soloman Barney Holdings lost US $ 300 million and Merrill Lynch lost US $ 135 million.
Main reason for the failure of the hedge funds were,
� The assumptions that were embedded in the computer models were violated.
� There was a breakdown in the historical patterns.
� The computer models were “black box” models that were making automated decisions without the real-time input of seasoned traders;
Hence the focal characteristics of complex sociotechnical systems which work as open systems meant that they are subject to disturbances that are of unfamiliar nature to workers and that were not and could not have been anticipated by designers. These unanticipated events can range from the catastrophic such as multiple financial disturbances during August of 1998. Norros [1996] conducted a field study of flexible manufacturing systems, and found that workers had to cope with an average of three disturbances per hour (i.e., events during which the system was functioning in ways that were not anticipated by designers). Because these events are unanticipated, the procedures or automation that designers provide will�by definition�not be directly applicable in these cases. Therefore, to deal with these disturbances effectively, workers must use their expertise and ingenuity to improvise a novel solution. In complex sociotechnical systems, the primary value of having people in the system is precisely to play this adaptive role. The aforesaid facts provide actual evidence towards the validity of the conclusion.
7. Describe in outline the types of studies (and methods) that would produce the necessary evidence?
Ans.7: The type of study and its associated methods that produce necessary evidence with respect to the problem posed by the paper is Empirical study. The object of study in the paper belongs to the tangible world of people, objects and events and hence it is evident that the study is called “empirical” or “factual” as a contrast to formal sciences like mathematics and logic which have no association with imperia. The latter deal with theory only and aim at the clarification of theoretical structures i.e. the forms of thinking, such as the processes of logical or mathematical analysis. The features of empirical study that include evidence of its usage with respect to the paper include quantitative and qualitative data collection. The empirical study is primarily used to identify and analyse information technology products and systems and how they are used to solve problems or to produce valid recommendations to deal with structural decision making, organisational and individual problem solving and development. The data collection methods mainly include detailed extensive questionnaire surveys, interviews, observations and recording, analysis and review of conversations. In the case of empirical study techniques the experimental design varies and may have more exploratory or more focused questions. The introduction of design rationale diagrams proved helpful for software development performance by providing measurement of project completion times, evidence of use and improved comprehension from interviews and analysis of protocols demonstrating improved understanding of issues. The empirical investigations use a mixture of data collection techniques. Those investigations have further been utilised t discover patterns in the data trends over the period of time, differences in use between technologies and different group of users. The empirical investigations in the paper are more evident as they are used to analyse processes and models based on how people make decisions, the problems they face and adaptability of user groups to new technologies.
THE ACTION RESEARCH (method) PARADIGM:
From its origin, this form of research has been used to obtain evidence by professionals who want to investigate and improve their own working practices. Academics carrying out action research into a situation are expected to collaborate with people in their natural settings. Their roles might change so that rather than coming in as experts to analyse, diagnose and suggest changes, they are facilitators who enable practitioners to do their own research. The researcher is actively involved with expected benefits for both the researcher and industry. Action research is characterised by concentration on real life and practical issues. It addresses the concerns and complex problems expressed by people living, working and acting in the real world. Researchers here do not simply observe and describe, they are concerned with doing things that make a difference, and learning about how the change was effected. This involves actively collaborating with people who live and work in the situation.
There are no restrictions on the type of data appropriate for action research; both quantitative and qualitative data can be used. Action research can relate to both action and research. Ideally, both types of outcomes should be achieved and that is what the paper aims at providing evidence to the usage of the action research methodology.