1. Task 1 mentions early and late binding time for monitor points, then
   clarifies that by saying that 'late' means enabling at compilation
   time, then using runtime mechanisms mentioned in Task 2. Task 2 lists
   only reading, resetting, and constraint triggering. Does this mean
   that interactive toggling isn't required?

2. Are non-CORBA use cases envisioned? By 'non-CORBA' we mean applications
   without CORBA or anything equivalent to it.

3. Some older feedback on the SOW indicates that a complex rule/constraint
   language is too heavyweight. In that feedback, the Extended Trader
   Constraint Language (ETCL) that's used in the Notification Service was
   explicitly rejected, but it has since crept into our discussion on
   this end. We need to get clear about what is desired, especially in
   light of the answer to 2.

4. The SOW seems to imply that distributed ACE logging is needed to meet
   the requirements (remote access, wide range of info destinations).
   ACE distributed logging consists of a central logging server and client
   logging 'proxies', where users of the service connect to the latter.
   While this approach will certainly meet the logging requirements, it
   might make it harder to meet the overhead requirements. By getting clearer
   on the logging requirements, we can better avoid overkill on the
   implementation.