[UCR] Use case review comments

Section 1.1 may  have missed what I regard to be an important aspect of a 
RIF and that is that the ability of the the recieving application to deal 
with the encoded message has much more to do with the message being 
properly identified and less to do with the specific content of the 
message.  It may even exclude the following two cases.
 
Imagine an application that recieves rule sets from multiple sources and 
charged with merging the rule sets into rules re-expressed in RIF.   I 
would expect the ability to recieve such messages crucial to the overall 
framework.  On recognizing a message it cannot decode, the application can 
take appropriate action, including that of being a broker passing off 
homgeneous rule sets to appropriate engines.
 
A second case that seems to be missing (and possibly excluded) by the 
present definition is that of an RIF editor.  On being handed an RIF 
message, the user of the application restructures the rules with some goal 
in mind and saves them.
The application need never executes any of the rules - though it may 
benefit from rule processing to help simplify or restructure a 
collection.  
 
This is somewhat related to the human rules discussion we had earlier.  
The content of the human rules must be clearly identified according to 
vocabulary in use, even though there is no intent that the recieving 
application be able to do anything more than pass on the rule to a human, 
appropriately identified as to origin.
 
These two cases both point towards a need for some basics of a primitive 
transport layer, possibly including support for identification of status 
or response.  A closely related example might be the semantics tag and its 
uses in MathML, where multiple representations are explicitly identified 
as alternatives.
 
I would be concerned if we are explicitly excluding such uses of RIF.
 
Apart from this the examples, and the notion of translation to and from 
RIF seem about right.
 
Stan Devitt
Agfa Healthcare.