CONCLUSIONS

As in the United States, many applications with a positive ROI have been deployed in Japan in which KBSs are used as a part of solving a problem. Generally speaking, KBS as a business has not been successful (there are no success stories comparable to ORACLE, SYBASE, and INGRES products in the database world). KBS applications tend to be complex and/or of narrow applicability, and to require continuous updating. Also, in contrast to conventional data processing, where economically justifiable applications are easy to find, selecting a KBS application is often a hit or miss process. For all these reasons, KBS technology has not done well as a business in either country.

Most successful new technologies grow initially at an exponential rate. While this was the case for KBS technology in Japan in the early days, we appear to be seeing slower growth there now. Even after a successful application, we often did not see the use of the technology multiplying within a company.

In order to have the potential for rapid growth, we heard in Japan that the technology needs to be improved in the following ways:

  1. Easier knowledge acquisition
  2. Reuse of the KB
  3. Ability to fully and directly express more of the model of the domain
  4. Need for a mature CASE-like methodology
  5. Easier development of the user interface

With these improvements, the cost and skill requirements of developing a new application could be brought down so that more attempts would be made to explore new areas of applicability (cutting the cost of wildcatting). This should realize the enormous potential of KBS technology.

Because so much of the technology has been mastered by and emanates from Japanese computer manufacturers, integration with conventional computing techniques is the norm and is handled easily. The JCMs' efforts, while downsized in some companies, continue to move ahead strongly -- witness the rewriting of shells from LISP to C, the attempt to develop task specific shells, and the move to workstations and PCs from mainframes. There is evidence that commercially the U.S. is investing at a lower rate compared to its previous rate -- owing to lack of venture capital and downsizing of computer company efforts, resulting in fewer new product announcements. This may repeat the familiar story of U.S. companies halting investment where the returns are poor, while the Japanese continue to invest. The Japanese may in this way develop commercial superiority.


Published: May 1993; WTEC Hyper-Librarian