[Stein, 1994] has suggested that ``agency'' (i.e., the property of being an agent) is determined by an observer's intentional stance; what a person views as an agent is an agent. While this may well be a tenable philosophical position, it is not clear what benefit it provides. Rather, in term of directing research efforts, particularly with the growing popularity of ``agents,'' it might be preferable to narrow the scope of the term. Even in research communities this designation is perhaps being abused. Work such as [Lansky, 1994] was once called an expert system. It was quite surprising to hear such a classic example of that paradigm being presented at the AAAI 1994 Spring Symposium on Software Agents.
The final system we discuss is Telescript ([Wayner, 1994]). Although few details of this proprietary system have been disclosed, enough information has been released to permit a tentative comparison. Telescript is a very sophisticated computational environment in which machine-independent programs move freely around a network. Telescript programs are interpreted, and interpreters exist for all standard platforms. Essentially, in terms of portability, it is the algorithmic equivalent of ``postscript.''
High Telescript, the system's programming language, is reportedly very similar to Smalltalk and Modula-3. Thus, it does not provide the right level of abstraction for writing agent applications. Furthermore, Telescript programs have fixed meanings, i.e. primitives are not interpreted with respect to their context. However, it would seem that Telescript might be an ideal system for reimplementing SodaBot. Regardless, we look forward to seeing what comes out of this very promising endeavor. Telescript seems like a step in the right direction.