The properties of the words as the basic units of the language

29)Basic features of semantic relationships in conversion.
Various classes of semantic relationships are examined: inclusion, possession,
attachment, attribution, antonyms, synonyms, and case. Have~has relationships,
which can easily be confused with these, are also discussed. The guidelines that
result from the analysis include ways to recognize when certain classes of semantic
relationships exist and how relationships from these classes should be represented in a
relational data model. The importance of obtaining the correct rain/max cardinalities,
which can only be done by having a clear understanding of the application (and
the relationships used to model that application), is highlighted.
It is shown that the verb phrase, has, can be ambiguous and should be changed
to one that is more appropriate for the application. In some situations this might
not have a direct impact on the structure of the final design, but changing the name
of a relationship to reflect more of the application's semantics should make the
database easier for the end-user to work with.
A prototype system, the Semantic Relationship Analyzer, is presented. It
implements the analysis carried out in this research. The system engages a user in a
dialogue that elicits the relationships of the user's application. These relationships
are then analyzed for their most appropriate semantic interpretations. The system's
capabilities include: correcting erroneous rain/max cardinalities; detecting missing
relationships; and inferring attributes and integrity constraints. A relational database
design is obtained as the system's output, which is comprised of a set of relations,
along with some attributes and semantic integrity constraints.
It appears that the only way certain semantic relationships (e.g., case relationships) could be really useful.
30.Initialization and acronyms.