Colloquium at Stanford
The Unfinished Revolution

Memorandum


Date: Thu, 2 Mar 2000 01:44:54 -0800 (PST)

From:   Eugene Kim
Reply-To: unrev-II@onelist.com

To:     unrev-II@onelist.com

Subject:   Use of Case Studies in a DKR

From:   Eugene Kim

On Sat, 26 Feb 2000, Eric Armstrong wrote:

We mare almost certainly decades away from having the capacity to automatically index a story under all the many headings that might be appropriate. The story about improving cars, for example, might be about improving designs, overcoming management resistance, achieving worker acceptance, getting customers involved, moving to standardized interfaces, none of the above, all of the above, or any combination of the above and many more.

How could we do that?

David Gelertner (who founded MirrorWorlds) wrote a wonderful essay in _Beyond Calculation: The Next Fifty Years of Computing_ (Copernicus 1997) where he describes an emotion function for indexing loosely associated concepts. The idea is simple; the implementation, perhaps intractable. A common association, for example, are women and flowers (excuse my male-centrism for a moment). Gelertner's explanation for why we associate these two seemingly disparate concepts is that they inspire similar emotions. This principle served as the foundation for one of Gelertner's research projects at Yale, the FGP Machine (Fetch, Generalize, and Project).

Incidentally, Gelertner, I think, falls in the opposite camp of Engelbart. A point that Peter made in his e-mail and in his CIM paper is that trying to automate human capabilities within a system is a bad strategy; it's better to augment human capabilities with tools. Gelertner is trying to automate human capabilities with the FGP machine. For the record, despite my fascination with Gelertner's work, I fall in the Engelbart camp. :-)

Sincerely,

-Eugene

Eugene Kim
eekim@eekim.com
http://www.eekim.com/