2-D diagrams LO4503 and Knowledge Distribution

Inflow@eworld.com
Sat, 30 Dec 1995 10:54:20 -0800

Replying to LO4477 --

In responding to a description of QuestMaps, Hal Popplewell wrote...

> This being done, do we not, then have knowledge encoded in "near" rule
> form and couldn't this knowledge translated to executable rules for
> permanent use in future decisions (intelligent decision support)?
>
> Capturing legacy knowledge in executable form is a new (and I suspect
> increasingly essential) way of being a learning organization.

Will this work in real knowledge-worker environments (non-linear, crative,
adaptive, etc.) like software/product development?

IMHO, I think QuestMaps are useful tools -- we all need maps to see where
we are at and where we are going -- they are great for group learning.
But trying to capture this information into executable rules to be
processed by an 'inference engine' sounds like we are trying to mechanize
an organic system/process [and repeating the failures of rule-based AI(not
ALL rule-based AI was a failure)].

Many knowledge processes appear self-organizing, with some 'legacy'
knowledge[captured], and much created knowledge[learning based on the
particular situation/environment/needs]. Therefore, maps are useful, but
pre-compiled solutions are not. Throughout most of the trip you are not
sure HOW you are going to get to where you are going -- you are constantly
looking at maps.

If it was so easy, Microsoft [who has a ton of legacy knowledge in
software development] would have figured this out and released Windows 95
in 1994 as they originally promise (or was it 1993?) ;-)

Valdis Krebs
Krebs & Associates
Los Angeles, CA

--
inflow@eworld.com