[ACT-R-users] Strict harvesting theoretical motivation

[db30 at andrew.cmu.edu]

--On Wednesday, October 21, 2009 11:03 AM -0400 "Stu @ AGS TechNet" 
<stu at agstechnet.com> wrote:

> Can anyone point me to a reference that discusses the theoretical
> underpinnings for strict harvesting (with respect to buffer clearing)?
>

Strict harvesting isn't a theoretical component of the system.  It's a
modeling convenience that was put into ACT-R 6.0 to address issues that
people had using ACT-R 5.0, particularly with respect to the perceptual
buffers.  Keeping the chunk around was the less common action in most
models and there were lots of explicit buffer clearings required in the
productions.  That was often confusing and required more work to get it
"right".  So, the clearing was made the default action instead.

It can be turned off on a buffer by buffer basis with the :do-not-harvest
parameter, and it can also be prevented in a production by using the buffer
on the RHS (even without making any modification).

Hope that helps,
Dan