= '''Discussion: Part/whole relations and ISO 15926-2''' = This page is intended for a particular example given in ISO 15926-2, page 102, where part/whole relations are discussed. Of particular difficulty is the fact that the diagram illustrating the discussion seems to confuse the notion of entity type with that of a defined class or relationship. In particular, the name ''assembly of individual'' is given as the name of a relation, even though this also names one of the most important entity types for relations.[[BR]][[BR]]The following graph is intended as an attempted reconstruction of the mentioned graphic. More discussion will follow, in the hope that confusion can be laid to rest. {{{ #!LatexEquation \TemplateDiagram{ % column 1 !c{WandI}{whole \& individual}{Role and Domain} ( [d(1.5)] !c{6ofM8}{6 of M8 bolts}{Role and Domain} [d] !c{my6ofM8}{\#my 6 of M8 bolts}{Physical Object} ) % column 2 ( [r] !t{corws_a}{CO Relationship with Signature}{assembly of individual} %% ( % detour for an additional entity type %% [d(.5)r(.4)] %% !e{ai}{Assembly of Individual} %% !l{corws_a}{ai}{}{} %% ) [d(1.5)] !t{corws_6M8_a}{CO Relationship with Signature}{6 M8 bolt assembly} [d] !q{or}{Other Relationship} ) % column 3 ( [rr] !c{PandI}{part \& individual}{Role and Domain} [d(1.5)] !c{PandM8b}{Part \& M8 Bolt}{Role and Domain} [d] !c{1}{\#1}{Physical Object} ) % subclass arrows ( !b{WandI}{6ofM8} !b{corws_a}{corws_6M8_a} !b{PandI}{PandM8b} ) % membership arrows ( !m{6ofM8}{my6ofM8} !m{corws_6M8_a}{or} !m{PandM8b}{1} ) % relation symbols tied to boxes, top -- bottom ( !s{corws_a}{WandI}{CO End 2}{PandI}{CO End 1} !s{corws_6M8_a}{6ofM8}{CO End 2}{PandM8b}{CO End 1} !s{or}{my6ofM8}{End 2}{1}{End 1} ) % cardinality symbols tied to central corws ( [d(1)r(0.3)] !c{card01}{0:1}{Cardinality} ) ( [d(1)r(1.7)] !c{card66}{6:6}{Cardinality} ) %% (!l{corws_6M8_a}{card01}{End 2 Cardinality}{.5}) %% (!l{corws_6M8_a}{card66}{End 1 Cardinality}{.3}) (!s{corws_6M8_a}{card01}{End 2 Cardinality}{card66}{End 1 Cardinality} } }}} Suggestions: * The ''other relationship'' instance should have an (additional) entity type ''assembly of individual'' * The relation which is called "assembly of individual" (overusing the name, clearly) should have an additional entity type ''class of assembly of individual'' * Perhaps the graph, as given in ISO 15926-2, is confusing only because the text is at the point in question ''only'' intending to make a point about assignment of ''cardinality constraints''. Indeed, the section header is called precisely "Cardinality constraints".