Tobin, when I used the term "platforms", I was referring to a base
from which further development builds. It is easy to measure whether
a change in welding procedure increases or decreases performance. I
is not so easy to measure whether teaching chemistry or computer
technology or metalurgy or time-and-motion studies to welders will
increase performance.
Why?
Partly because of the time difference. We (usually) presume that the
direct measure of improvement will take longer to occur if at all.
Partly because if performance does improve, will we know it can be
attributed to the platform development of education?
This is the challenge of attribution of value to earlier stages of an
evolutionary process. Here is the real challenge. Was the "idiot"
in Einstein's genetic chain a major contributing source to his
brilliance? (I made up the relative to make my point.) We don't
know.
To create the conditions of effective adaptation, we need to assign
values to the contribution of earlier steps that may, in themselves
have been total failures but which provided the possibility of future
success.
Processes for creating/evaluating effective processes, in the world
where "platforms" are a concern, are those which can track the
effects of earlier generation. I would suggest that a beginning has
been made in this field by John Holland and those working on genetic
algorithms, artificial life, etc.
-- Michael McMaster Michael@kbddean.demon.co.uk