3.3. Collateral and parallel clauses

A collateral- or parallel-clause consists of a sequence of units separated by and-also-symbols (viz., ","), and is enclosed by parentheses or by a BEGIN-END pair; a parallel-clause begins moreover with PAR.

Collateral-clauses, but not parallel-clauses, may yield stowed values composed from the yields of the constituent units.

Examples of collateral-clauses yielding stowed values:

[ ] INT q = (1, 4, 9, 16, 25);
STRUCT(INT price, STRING category) bike := (150, "sport")

.

PROC VOID eat, speak; SEMA mouth = LEVEL 1;
PAR BEGIN
      DO
          DOWN mouth;
          eat;
          UP mouth
      OD,
      DO
          DOWN mouth;
          speak;
          UP mouth
      OD
  END 

.}

3.3.1. Syntax

b) SOID NEST joined portrait{a,b,c,d,34g } : where SOID balances SOID1 and SOID2{32e } , SOID1 NEST unit{32d } , and also{94f} token, SOID2 NEST unit{32d } or alternatively SOID2 NEST joined portrait{b}.

c) strong void NEST parallel clause{5D ,551a} : parallel{94f } token, strong void NEST joined portrait{b} PACK.

d) strong ROWS of MODE NEST collateral clause{5D ,551a} : where (ROWS) is (row), strong MODE NEST joined portrait{b} PACK ; where (ROWS) is (row ROWS1), strong ROWS1 of MODE NEST joined portrait{b} PACK ; EMPTY PACK.

e) strong structured with FIELDS FIELD mode NEST collateral clause{5D ,551a} : NEST FIELDS FIELD portrait{f} PACK.

f) NEST FIELDS FIELD portrait{e,f} : NEST FIELDS portrait{f,g}, and also{94f } token, NEST FIELD portrait{g}. {FIELD :: MODE field TAG.}

g) NEST MODE field TAG portrait{f} : strong MODE NEST unit{32d } .

h) *structure display : strong structured with FIELDS FIELD mode NEST collateral clause{e}.

i) *row display : strong ROWS of MODE NEST collateral clause{d}.

j) *display : strong STOWED NEST collateral clause{d,e}.

k) *vacuum : EMPTY PACK. {Examples:

}

a)

(x := 1, y := 2)

b)

x := 1, y := 2

c)

PAR(task1, task2)

d)

(1, 2) (in [ ] REAL(1, 2))

e)

(1, 2) (in COMPL(1, 2))

f)

1, 2

g)

1

MODE M = STRUCT(REF M m); M nobuo, yoneda;, the assignation nobuo := (yoneda) would be syntactically ambiguous and could produce different effects; however, m OF nobuo := yoneda is unambiguous.

Row-displays contain zero, two or more constituent units. It is also possible to present a single value as a multiple value, e.g., [1: 1] INT v := 123, but this uses a coercion known as rowing {6.6 }.}

3.3.2. Semantics

b) The yield W of a STOWED-collateral-clause C is determined as follows:

If the direct descendent of C is a vacuum,

then{'STOWED' is some 'ROWS of MODE' and} each bound pair in the descriptor of W is (1,0) {and it has one ghost element whose value is irrelevant};

otherwise,

• let V₁, ... , V_m be the {collateral} yields of the constituent units of C;

Case A: 'STOWED' is some 'structured with FIELDS mode':

· the fields of W, taken in order, are V₁, ... , V_m;

Case B: 'STOWED' is some 'row of MODE1':

· W is composed of

1. a descriptor ((1,m)),
2. V₁, ... , V_m;

For i = 1... m,

· the element selected by the index (i) in W is V_i;

Case C: 'STOWED' is some 'row ROWS of MODE2':

• it is required that the descriptors of V₁, ... , V_m be identical;
• let the descriptor of {say} V₁ be ((l₁,u₁), ..., (l_n,u_n));
• W is composed of
1. a descriptor ((1, m), (l₁,u₁), ..., (l_n,u_n));
2. the elements of V₁, ... , V_m;

For i = 1... m,

• the element selected by an index (i, i₁, ... , i_n ) in W is that selected by (i₁, ... , i_n) in V_i.