The ORM Foundation

In the attached file I have shown a simple ORM model, using two different types of constraints on the ternary predicate.  The verbalizations appear to be equivalent except for the part about the chosen constraint type.  My question: are these equivalent?  It certainly is true that the verbalization for the not equal value comparison constraint is much clearer and more direct that the verbalization for the irreflexive ring constraint.  The value comparison constraint was in the solution submission from one of my students for our latest assignment.  I note that it was introduced into NORMA much later than the ring constraint.

 Hi Gordon,

In your PDF you say " Note the formatting  problem when doing Copy and Paste on the verbalization"

I could not see any "formatting problem"

Please will you clarify this point.

Thanks

Ken  .
 

Gordon,

Regarding the formatting problem: "all the indents are removed so all lines start at the left margin. "

I made some tests on extracting the contents of the Verbalization window and have found a workaround:

The Tests

Procedure 1: Lasso the contents of the verbaliser window then:.  

1.1: Copy & paste  to Word  - Indents: lost, Text colours: preserved

Procedure 2: Lasso the contents of the verbaliser window then right-click and choose from the menu options.   

2.1: Export to OneNote - Indents: lost, Text colours: lost

2.2: Print (via Print Preview)  - Indents: preserved, Text colours: preserved  (Print to printer or PDF)

Ken 

 Hi Gordon

Yes, an irreflexive ring constraint is equivalent to a value-comparison constraint with the inequality operator. E.g. a simple irreflexive constraint on a binary predicate R formalizes as  forall x not xRx, and a corresponding value-inequality constraint formalizes as  forall x, y (xRy -> ~x=y). It's easy to prove in logic that these expressions are equivalent.

As minor issues, your ring constraint case would verbalize better if the object type was called Guest (as in the value constraint case), and the diagram for the latter case would be cleaner if the role connectors connected to the join of adjacent roles (as in the former case).

Cheers

Terry