ORB Visualization
(soon)
National Knowledge Project
comments
Note from someone
Paul,
We can enumerate some well-posed problems that require solution. People can then respond with coherent arguments.
So, here's an example breakdown:
· Didn't have relevant information in system: go get it
· Couldn't find the information: make better retrieval systems
· Didn't recognize what was relevant: envision better
· Couldn't assemble and integrate: build organizational processes and systems to support integration
· Found and assembled information but it was suppressed: work on org structure & personnel
· Couldn't produce the analysis: get better people, develop better theories, improve education
If you think that the combination of human and computer knowledge processes didn't work, well then perhaps it is time to bring in the A-team to design systems of systems with the superior properties.
After developing one of the first demonstrations of argument-structured wide-area collaboration we found the follow up by the B and C teams to be rather uninspiring.
Historically, the government does best on sophisticated projects when it can mobilize the best talent the country has, typically with some sense of urgency.
Personally, I believe that we are poised for some major advances in understanding unrestricted text (based on progress in the computational linguistics community), but superior retrieval tools and even power analysis tools may not make a big impact unless they are embedded in the proper organizational contexts and used by capable and motivated people.
So, chances are that the major improvements can be had by working on superior organizational designs and processes, rather than by introducing technical enhancements for specific functions.
The caveat are:
· If people are buried under redundant and extraneous information, then improving the precision of retrieval and routing (while ensuring completeness) can help;
· If failure to integrate information across administrative boundaries or compartments is causing difficulties, then organizational adjustments and secured knowledge-based collaboration systems can help.
This is a real problem in computer security and access control. Software productivity and expenses for military and intel systems is a disaster -- many major systems exhibit software costs in the 50-80% range and require highly-paid US programmers.
Those who complain that we cannot swap out the entire infrastructure fail to realize that we recreate the entire infrastructure every 5-10 years, so we only need better designs that we phase in during the subsequent cycles.
Since some of you are systems theorists, consider the task as moving toward ultra stable (autopoeitic) organizations that continue to achieve their mission through major evolutions in the environment and under attrition from attacks by adversaries. If this is what we want, then question is:
How do you design organizations that can learn and continuously but incrementally restructure themselves to take advantage of technical (or initiate) change while confirming that restructurings are adaptive (preferably without empirical counter examples)?