Visual Planner TODO List:

Fix bug that results when an aggregate operator
exists at the top of the plan tree.  All operators
used in a path tree *have* unique IDs established
by the pointer reference to that node.  Nodes 
created for use in the executor do not have IDs
properly assigned and instead use a common ID.
The Visual Planner client
stores the root of each plan tree in a hash map
for reference.  If the IDs of each plan tree are
not unique, this scheme does not work.

Devise a compression scheme to make for more 
efficient transmission of XML data.  Currently,
all path nodes are listed in the stream in their
entirety, whether or not they had already been
referenced previously.  This results in a large
amount of duplicate data.  By utilizing the unique
IDs of each path, it should be possible to place
each path in the XML stream once and then refer
to it by ID in later use.

Using some of the server components used by VP,
is it possible (or desirable) to create a new 
command 'planner' that works in way similar to 
'explain' but instead provides a list of all
paths considered by the planner in pretty print
format?  If that results in too much information,
in what way could this information be simplified?

The potential strength of the Visual Planner tool
is that it presents a complex operation in a 
simplified manner.  Is there a way to provide 
further annotation to a path node that could
be then conveyed to the client application? If so,
what sort of information would benefit the user?