440 Davis Court #1602
San Francisco, CA 94111-2496
415 781 5700

March 15, 2000

03 00050 61 00031501

Mr. John Hogden
Los Alamos National Laboratory; MS B265
Los Alamos, NM 87545

Subject:   Alliance for Development

Dear John,

Thanks for taking a look at POIMS, and for your constructive response to my earlier letter.

Last year I submitted a proposal to the National Science Foundation seeking support for research on a system of Communication Metrics. In December, NSF reported that the system would be very useful, but I lack the polish to perform the work. This overlooked the fact that the work has already been mostly done. What I need help on is creating a stronger implementation of subject management, actually a little broader than in the NSF proposal. Earlier on June 15 NSF staff anticipated that reviewers would not grasp Communication Metrics from a mere narrative, and so suggested providing a link to the web where people could experience the meaning of delivering intelligence on the web. So, we did that; but, NSF review panel did not have time to click on the link, and thereon concluded the system was only theoretical, and looking at Welch's credentials reasonably concluded taxpayer money would be better invested elsewhere. I would make the same decision under those circumstances, though I like to think I would have clicked on the link to check things out. But who knows.

In any event, you might take a crack at the NSF proposal. Can you see the lineaments of a system for computer aided thinking that significantly improves knowledge work, which I think is part of the objective your work is seeking? Then look at a letter from a computer vendor I received yesterday, and jumped it into an "intelligence" mode, as I did your letter today. Click on just a few links in the letter, and and try to get the feel that information is segmented in the underlying SDS record into chunks of related context, and that these are tied to chronologies showing traceability to original sources, all of which emulates, in significant degree, human intelligence, as explained in POIMS.

Please don't spend too much time on this. Usually, if you don't get into the feel of an "intelligent" environment right away, people come away frustrated. That is what the U.S. Army Corps of Engineers found when they evaluated this system in 1997.

Keep in mind that in looking at the online record, you are only seeing 25% of the system. The real power is with the person who uses SDS to create stuff, and ties it all together. Indeed, we call it computer aided thinking, because we are using the technology to enhance our thinking by adding the dimension of time (chronology) to the conventional practice of writing that produces information.

Where the lab comes in is as a possible development partner to "polish" the subject management feature, which was, again, proposed to NSF, although they never really focused on that point. Subject management is the "organization" part of intelligence. It is a very complicated matter, but my system allows people to assemble related information into chronologies of cause and effect. That is where "pattern recognition" comes into the picture, which is the subject of my first letter. People recognize consistent patterns in stories over time and thereby acquire knowledge, called experience. My system greatly aids this process, and I think it can likely be done even better with an enhanced design.

The reason the lab might support this is to acquire capability to perform its work that is not presently available by any other means, i.e., the ability to convert information into useful knowledge faster, better and cheaper. I am deliberately leaving the nuts and bolts of subject management a little vague, because for one thing it takes a long explanation, and seeing it used only takes a few seconds to recognize the power. Second, it is a proprietary capability that I would be willing to share with Lab in exchange for support to create the enhancement.

A second, more obtuse, development objective is to look for a way to further automate the process of connecting related chunks of information. I really don't think it can be done much better than I am doing it now, because at the end of the day I am not sure I want a machine doing the thinking. I want it to help me do my thinking and to synthesize the thinking of others. On the other hand, if we were to get going and do the subject management task which has big potential, we could look at this second part, and maybe get lucky and find a way to tie things together that saves a lot of time. That's really what it's all about. saving time, while being more accurate.

Sorry again for a long story. If you want, we can talk on the phone. Just let me know. I realize some of this sounds flaky, but a real system exists that works every day, and is working right now in thinking about how to communicate effectively on a difficult, complex, indeed "touchy," subject like "intelligence. This system works, and it saves real money, as reported by the U.S. Army Corps of Engineers. See what you think about joining in to enhance civilization. You may not have time for this, but here is a recent review by a member of the Colloquium at Stanford, who observed the system being used.



Rod Welch