difference by assigning variables using '==' and 'var: value' in recursion

[wdduncan]

I am experimenting with using recursion to return RDF triples. In my example use case, I am wanting query for all types of teeth.

Here are are my basic setup:

  logica_lib.RunPredicateFromString("""
  @Engine("sqlite");

  Tooth(subject: "tooth", predicate: "type", object: "tooth");
  Tooth(subject: "canine", predicate: "type", object: "tooth");
  Tooth(subject: "upper-left canine", predicate: "type", object: "canine");
  Tooth(subject: "upper-right canine", predicate: "type", object: "canine");
  Tooth(subject: "bottom-left canine", predicate: "type", object: "canine");
  Tooth(subject: "bottom-right canine", predicate: "type", object: "canine");

  ChildOf(x, y) :- *** see #1 and #2 below ***;

  DescendantOf(x, y) :- ChildOf(x, y);
  DescendantOf(x, z) :- ChildOf(x, y), DescendantOf(y, z);
  IsaTooth(x) = y :- y == "tooth", DescendantOf(x, y);

  Teeth(tooth: subject) distinct :- 
    Tooth(subject:, predicate:, object:),
    IsaTooth(subject);
""", 'Teeth')

I can define ChildOf in two ways.

1. assign variables using ==:

  ChildOf(x, y) :- Tooth(subject:, predicate:, object:), 
    subject == x, 
    predicate == "type", 
    object == y;

2. assign value in the predicate:

  ChildOf(x, y) :- Tooth(subject: x, predicate: "type", object: y);

Each returns the same table.

What is the difference in the two ChildOf predicates?

[EvgSkv]

@wdduncan As far as I see there is precisely no difference, it compiles to exactly the same internal representation and evenaully same SQL (maybe with different order of constraints in the WHERE clause). It is also logically equivalent, of course.

Which way to write is purely a readability question, so chose whatever you find more intuitive.