Transparency of Model Structures in Learning Environments
Posted: Wed Mar 24, 1999 8:10 am
A few thoughts regarding Sauds inquiry about making model struture
transparent:
- Lead with behavior, then with structure. Weve found that people get a lot
more interested in structure, once they experienced some behavior that they
wish to have explained.
- Use simplified versions of structure. Generally models that provide
credible outputs for a learning environment demand structural complexity
which may not add to the dynamics, but certainly adds to the spagetti.
Simplify struture to the minium required to explain the behavior of
interest.
- Use a slow reveal technique, like ithinks logic tracing capability. The
idea is that even simplified structure can be overwhelming to the
uninitiated. So start with a little structure and build up gradually adding
more pieces. I would think you could program something like this in your VB
interface.
Phil Odence
HPS
From: "Phil Odence" <podence@hps-inc.com>
transparent:
- Lead with behavior, then with structure. Weve found that people get a lot
more interested in structure, once they experienced some behavior that they
wish to have explained.
- Use simplified versions of structure. Generally models that provide
credible outputs for a learning environment demand structural complexity
which may not add to the dynamics, but certainly adds to the spagetti.
Simplify struture to the minium required to explain the behavior of
interest.
- Use a slow reveal technique, like ithinks logic tracing capability. The
idea is that even simplified structure can be overwhelming to the
uninitiated. So start with a little structure and build up gradually adding
more pieces. I would think you could program something like this in your VB
interface.
Phil Odence
HPS
From: "Phil Odence" <podence@hps-inc.com>