EnglishΕλληνικά
 
 


CMTool
The effort required to construct a task model and to maintain various instances of it is one of the main reasons discouraging the wider adoption of task modeling (Kieras, 1996). This finding inspired us to develop appropriate tool, the Cognitive Modeling Tool (CMT).

CMT has been developed to facilitate the task modeling process. Task models are structured in a hierarchical way, as shown in figure 2. CMT uses a direct manipulation approach for editing and modifying this task hierarchy. Any specific node represents a task relating to a user goal. The sub-tasks which serve to accomplish this goal are associated to the node. These subtasks can be related through a specific plan involving OR, AND logic operators. This plan is usually shown next to the specific high-level goal. Additional information relating to comments on the goal accomplishment can be attached to the node.

An individual user task model can be annotated with comments relating to task execution. A special notation has been defined, extending the HTA plan notation (Shepherd, 1989), with tokens referring to user deviation from expected task execution. So token (!) marks a non-destructive syntactic deviation, while (x!) marks a not-completed task execution. Also introduction of new unforeseen tasks in a plan by a user are marked as {task-id}, as shown in the figure below.



Both textual and graphical notations (e.g. GOMS and HTA notations) can be used for task representation. If the analysis is focused on time required to accomplish a task, time related information can also be stored on each task. Concurrent display of a keystroke log file and the corresponding task model structure is supported. Task models are stored in a relational database, grouping the various systems analyzed, with additional identification information (designer’s model or revised designer’s task model (DTM) or user’s task models (UTM)). Also quantitative analysis tools are supported to extract useful metrics related to the analyzed tasks, such as number of keystrokes required to achieve a specific goal or sub-goal, mean time required and interaction complexity of specific user model compared to primary designer’s expectations or to revised and adapted model.

The CMT can support communication across a design team (often consisting of designers from different disciplines and backgrounds) because of its ability of exporting task models in various possible representations: tree view, sequential view and structured report view, all consistent with each other thus supporting the design process. Novel functionality of the tool is its ability to automate synthesis of task structures already stored in the database. Using this feature, various sub-goal structures can be combined or not shown in a task model, according to the degree of detail required in the context of a particular study. Additionally, since task models are stored into a relational database schema, appropriate querying can lead to interesting comparisons relating to various aspects of modeling, such as execution times and frequency of usage patterns, both in absolute or relative values (e.g. compared to other task structures). In CMT the evaluator or the designer can select parts of a task structure representing a specific problem solving strategy, which can be stored for future reference or comparison with other users' strategies. Additionally CMT supports storage of various users’ characteristics and further analysis is supported, focused on user characteristics such as age, grades and gender.



A novel and time-saving functionality of the CMTool is its ability to automate synthesis of task structures already stored in the database. Using this feature, various sub-goal structures can be combined or temporarily hidden away in a task model, according to the degree of detail required in the context of a particular study. Additionally, since task models are stored in a relational database schema, interesting comparisons can be made through database querying, relating to various aspects of modelling, such as execution times and frequency of usage patterns, both in absolute or relative values (e.g. compared to other task structures). In CMTool the evaluator or the designer can select parts of a task structure representing a specific problem solving strategy, which can be stored for future reference or comparison with other users’ problem solving strategies. Additionally, CMTool supports storage of various users’ characteristics, such as age, gender etc, so that further analysis can be supported, focused on them. In CMTool the evaluator or the designer can select parts of a task structure representing a specific problem solving strategy, which can be stored for future reference or comparison with other users' strategies. In addition, the CMTool facilitates the storage of various users’ characteristics and further analysis is supported, focused on user characteristics. The possibility to analyse system’s usage in five dimensions is a contribution of the tool to the evaluation of open environments both in usability and learning terms.
(1) general problem solving strategies,
(2) users,
(3) specific activities to carry out a strategy,
(4) tools and
(5) usability problems detected in interaction.

This process is carried out through a visual stored information query environment, shown in the Figure above. Modelling experiments could be analysed across any possible element of the five different dimensions discussed (e.g. “Show all encountered problems related to tool X”, or “Identify all usability problems of users age 15 and above related with problem solving strategy Y“, etc.).Complex and in-depth analysis can be carried out according to any of these dimensions, supporting in-depth study and analysis of encountered problems. For example, a usability engineer could focus on a classification of problems detected concerning particular tools.
 



Computer-supported collaborative learning has been an active area of research since the beginning for the HCI group more>>



Web usability team of the HCI Group has been active in studying human-web interaction and ways to support the design of accessible, findable, usable and aesthetically appealing web sites. more>>


Mobile Technology Unit of the HCI Group has been studying design and evaluation of mobile applicationss more>>

Hci Group | Electrical and Computer Engineering | University of Patras