Telos is a language for developing information systems. The basic idea behind Telos is to create a semantically and syntactically rich environment for transparently and efficiently encoding the knowledge needed to build metamodels for software engineering and database access.
While the attempt to delineate the knowledge necessary for creating and maintaining information systems--the authors distinguish subject, usage, system, and development “worlds”--is certainly worthwhile, the paper is a continuation of such orthodox ideas as semantic networks and frames. Consequently, Telos has inherited the eclectic nature of arbitrary relations of the nodes within semantic networks. It has kept the flavor of a system whose features are determined by clever syntax, but which lacks a rich enough semantics.
Moreover, Telos is similar to Chandrasekaran’s work on creating metalanguages for specific areas of problem solving (such as diagnosis and manufacturing) [1]. Chandrasekaran’s work originated in the medical domain. His ideas were based on the assumption that every problem-solving activity warrants a metalanguage capturing the idiosyncrasies of its domain.
Unfortunately, most approaches to knowledge representation, be they domain-specific, problem-solving specific, or multipurpose, have turned out to be successful theoretical undertakings rather than breakthroughs in implementation. Granted, the task of creating, maintaining, and especially using knowledge bases is gigantic. The authors have laid the groundwork for an interesting way of dealing with the issue related to information systems. What is needed now is a scalable system that proves their points.