Sophisticated development environments, such as PROTÉGÉ-II, can be difficult to use, because the models and tools are complex. Even developers with adequate training make mistakes, and introduce problems that could have avoided with hindsight. Besides the traditional approach of providing help systems for the developer, and of preventing the developer from making mistakes (e.g., by disabling irrelevant operations in the user interfaces of the tools), development environments benefit from critiquing systems. Such critiquing systems can make a thorough analysis in retrospect of ontologies and other models developed (i.e., when the developers have a stable version), and can report potential and serious problems.
The current prototype implementation of CT is sufficient for detecting rudimentary problems in knowledge-acquisition tools generated by DASH. The developer can take advantage of the reports generated by CT, and use this information to correct flaws and improve the knowledge-acquisition tools. However, there is much room for improvement of CT in terms of the problems it can detect, and in terms of the critique reporting. Despite these limitations of the CT implementation, we believe that critiquing systems can help developers avoid many problems in complex development environments.