[great-users] More Formal Definition of "Packets"

[Chris vanBuskirk]

The GReAT Users' Manual introduces the term "packets" with the
following:

==========
[each rule has a]
MATCH CONDITION:  can be either "all matches" or "any match". If it is
"allmatches" for each input <<packet>> the rule will find all the
matches and for all the valid matches it will execute the effecter.  
==========

Later, the word "packet" is mentioned 41 more times, but there doesn't
seem to be a real definition of "packet" anywhere near the beginning of
the document.  

IMHO this is a deficit.  I offer the following as a starting point for
agreeing on a more formal definition:

"The set of input ports for any rule/test-case form a tuple that
reference a sub-graph(s), within the overall graph(s), from which the
pattern matching algorithm is to initiate its search."

{Note that a set of input ports of size one (though pointing to only one
graph "node") is still a proper 'sub-graph' of the original graph(s).}  

Any reactions?

Sincerely,
  Chris