End-user Programming Environment :CLEPE

In practice, it is really hard to develop an automatic problem solving system which can cope with a variety of problems we expect to be solved by computer systems. The main reason is that the knowledge needed for solving the problems which varies considerably depending on the properties of the problems. This fact implies that we should realize the common knowledge, sometimes ignored, that users have more knowledge than computers. From this point of view, importance of user-centric system [DeBellis, 96] is now widely shared by many researchers.

End-user programming environment, as the incarnation of the philosophy, provides end-users with a variety of functional components which stand for the concepts appearing in the target task and allows them to build their own problem solving models in terms of those components. In such an environment, end-users can easily describe his/her knowledge by using the components.

To realize an end-user programming environment, it is necessary that the environment should

1. adaptively evolve according to the changes of requirements and the changes of the target world dealt by it,
2. have a framework for explicitly representing computational semantics of the components provided for end-users,
3. have a framework for end-users to easily and smoothly externalize problem solving knowledge in their mind in the computer-readable form, and
4. have the capability to interpret the description and generate the runnable problem solving model with both rigid computational semantics and high cognitive fidelity.

Our research on task ontology concerns all of the above requirements for end-user programming environments. In principle, task ontology is a systematic definition of the concepts appearing in end-users' understanding of problem solving.

We expect that, in terms of ontology, the environment would be able to capture the end-users' conceptual model of problem solving in right abstract level and provide them with useful programming guidance.

Our research project aims at developing a task-ontology (static user model) embedded in the environment which satisfies all the four requirements above.

Conceptual LEvel Programming Environment, named CLEPE, is an environment for end-users to describe problem solving knowledge based on the task ontology. The target tasks of CLEPE are rather routine tasks, such as scheduling, salary calculation, and so on.

In CLEPE, end-users describe their own problem solving knowledge in a diagrammatic representation with a constrained set of natural language sentences. And then, he/she can verify system's interpretation of the description using the conceptual level execution functionality of CLEPE. The continuity from the diagrammatic representation to conceptual computational semantics is one of the key features of CLEPE,

In this environment end-users can

1. describe their own problem solving processes using the concepts of which they are most conscious,
2. understand the execution result of the model in appropriate level of abstraction and
3. debug it at the conceptual level.

1. An Ontology for Building A Conceptual Problem Solving Model, Proc. of ECAI98 Workshop on Applications of ontologies and problem-solving model, pp. 126-133, (1998).(also submitted to Knowledge and Information Systems)
2. CLEPE: a Task Ontology Based Conceptual Level Programming Environment, (in Japanese)
3. Capturing a Conceptual Model for End-user Programming -Task Ontology as a Static User Model-, Proc of UM'97
4. Design of a Conceptual Level Programming Environment Based on Task Ontology, Proc of BKK'96