Portlet lifecycles and CCP's

[ricky_frost@peoplesoft.com]

I've been trying to understand the relationship between POP's and CCP's and
the Portlet lifecycle. It appears that before a blockingInteraction can be
successful a portlet must be consumer-configured. This happens by one of
two ways. Either by first calling the optional Management clonePortlet, or
by requesting cloneBeforeWrite on the blockingInteraction itself. So it
seems that a Producer not offering a Management option might still allow
the cloneBeforeWrite request. My question: how then will these
"cloneBeforeWrite" portlet instances get cleaned up?

While searching for an answer I came across section 8 of the 1.0
specification where there is a reference to "releasePortlet" as follows:

Any Producer that supports cloning portlets on performBlockingInteraction()
MUST support the releasePortlet() operation.

Can someone explain this? There are no other references in the document...
Ricky Frost