parid9005 | Tue, 25 Feb 2003 17:04:28
Every assertion has one and only one a-node. cf. parid0191
(add)
New sub-section under Rules for Assertions stating that assertion
has only one a-node.
—pdurusau@emory.edu
parid9005 | Tue, 25 Feb 2003 17:04:28
Every assertion may have one t-node.
(add)
New sub-section under Rules for Assertions stating that an
assertion may have one t-node. parid0178
—pdurusau@emory.edu
parid9005 | Tue, 25 Feb 2003 17:04:28
If an assertion has a t-node, then it has a AT-arc and a TA-arc with the a-node and t-node as endpoints.
(add)
New sub-section under Rules for Assertions stating the arcs that
are present if a t-node exists in the assertion. cf. parid0177
—pdurusau@emory.edu
parid9005 | Tue, 25 Feb 2003 17:04:28
Every assertion must have at least two c-nodes.
(add)
New sub-section under Rules for Assertions stating that an assertion must have at least two c-nodes. Note that I have not seen a answer to Bernard Valant's question about why two or more AC arcs? Posted 16 November 2002, Bernard asks: "Why "two or more"? There are many cases of assertions with a single role type (take "sibling" for example)" Is this a symptom of conflating instances and classes? cf. my comments on role type in parid2260. (on this rule, cf. parid0180)
—pdurusau@emory.edu
parid9005 | Tue, 25 Feb 2003 17:04:28
Every assertion must have an AC-arc and CA-arc for each c-node in the assertion.
(add)
New sub-section under Rules for Assertions stating the arcs that
must be present between the a-node and c-nodes in an assertion. cf.
parid0179
—pdurusau@emory.edu
parid9005 | Tue, 25 Feb 2003 17:04:28
Note: In a minimum assertion, there exist two (2) AC-arcs and two (2) CA-arcs between the a-node and the minimum required c-nodes. cf. parid0180
(add)
Note (for immediately preceding REF) to clarify any lingering doubt
about the arcs that are present in a minimal assertion. Not really
necessary but notes rarely are. ;-)
—pdurusau@emory.edu
parid9005 | Tue, 25 Feb 2003 17:04:28
Every assertion must have a unique r-node for each c-node.
(add)
New sub-section under Rules for Assertions stating that an r-node
must exist for every c-node in an assertion. cf. parid0182
—pdurusau@emory.edu
parid9005 | Tue, 25 Feb 2003 17:04:28
Note: The requirement of a unique r-node is to make it explicit that c-nodes can never share an r-node via CR-arcs.
(add)
add as note to preceeding new entry.
—pdurusau@emory.edu
parid9005 | Tue, 25 Feb 2003 17:04:28
Every assertion must have an CR-arc and RC-arc for each c-node in the assertion.
(add)
This is the equivalent of requiring the number of RC-arcs to match the number of AC-arcs, but via requiring an r-node for every c-node and then requiring the arcs to be present. cf. parid0181
—pdurusau@emory.edu
parid9005 | Tue, 25 Feb 2003 17:04:28
Every assertion must have at least one x-node.
(add)
Sets up requirement for at least one Cx arc and to have at least
one role player. cf. parid0184
—pdurusau@emory.edu
parid9005 | Tue, 25 Feb 2003 17:04:28
Every assertion must have at least one Cx-arc and xC-arc for at least one c-node.
(add)
Builds on x-node requirement to require arc to role player.
—pdurusau@emory.edu