Relationship Grammar

The relationships section follows the class section. Here each relationship is defined. Here are a couple of examples so we can look at the basic structure.

relationships
    R1
    has service features define by, 1 Bank
    defines service features of, M Shaft
    Shaft.Bank -> Bank.Name
--
    R2
    travels through, 1 Shaft
    is conduit for, 1 Cabin
    Cabin.Shaft -> Shaft.ID
--

We start off with the unindented relationships keyword to mark the beginning of the section. Then we have any number of relationship descriptions.

The relevant grammar is:

// Relationships
// ---
rel_section = relationship_header rel*  // All relationships in the subsystem
relationship_header = "relationships" EOL*
rel = rname (ordinal_rel / binary_rel / gen_rel) block_end  // A rel is an ordinal, association or a generalization
rname = INDENT rnum EOL*  // Name of a rel is its rnum
...
rnum = r'O?R[1-9][0-9]*' // Relationship number

Here we see the rel token represents a relationship description.

The indented rname token defines a relationship's name as a short string consisting of an optional O character (if this is an ordinal relationship) followed by an R followed by a number.

We also see that there are three possible types. Each is indented to line up with the relationship name. And each is terminated by an unindented -- separator.

See the sidebar for a breakdown of each of these relationship types.