W3C home > Mailing lists > Public > public-rif-comments@w3.org > June 2010

Clarification regarding the semantics of RIF-PRD

From: Carlos Damásio <cd@di.fct.unl.pt>
Date: Fri, 11 Jun 2010 12:36:46 +0100
Message-Id: <CA92B754-3B9A-464B-ADC9-0CEA1AB9D0BA@di.fct.unl.pt>
To: public-rif-comments@w3.org
Dear RIF Working Group,

I have some doubts regarding two issues in the current RIF-PRD  
proposed recommendation.

1) Handling of frame atomic formulas.

According to Definition of Matching Substitution (sect. 2.2.1) the  
empty substitution does not
match the ground formula s[p1->v1] to { s[p1->v1 p2->v2] } since s[p1- 
 >v1] does not belong to { s[p1->v1 p2->v2] }.
Is this the intended behaviour?

A related problem is when the action is retract( s[p1->v1] ) applied  
to state { s[p1->v1 p2->v2] }.
The result according to the specification is apparently { s[p1->v1 p2- 
 >v2] }.
Is this the intended behaviour?

In summary, is { s[p1->v1 ... pn -> vn] } equivalent to { s[p1- 
 >v1], ..., s[pn->vn]} or not?

2) Rule variables (section 4.2.3)

The concept of rule variable is never formally defined (the notion of  
rule variable appears afterwards in Section 7.3 and 8.5.1.3 Forall).
My interpretation is the following:
- for unconditional action blocks, rule variables are empty (i.e. free  
variables).
- for conditional action blocks, rule variables are the free variables.
- for rules with variable declarations, the rule variables are the  
universally quantified variables.

Is this correct? Wouldn't it make more sense to equate rule variables  
to free variables in all cases?
In this way, universally quantified variables would be ignored for  
comparing instances of two rules, while free variables
would be used for that purpose in all cases. This behaviour is  
probably more coherent and easier to explain to users.

A related problem occurs in Section 7.3, where in case 2 of the  
algorithm clashes of names among rule variables are not taken care.

Best regards,

Carlos Viegas Damásio

----------------------------------------------------------------------------
Carlos Viegas Damásio                   | Tel: +351 21 294 8536 (ext.  
10758)
Professor Associado (Associate Prof.)   | Fax: +351 21 294 8541
Centro Inteligência Artificial (CENTRIA)|
Departamento de Informática             |
Fac. de Ciências e Tecnologia da        |
Universidade Nova de Lisboa             |
Qta. da Torre - 2829-516 Caparica       | e-mail: cd@di.fct.unl.pt
PORTUGAL                                | http://centria.di.fct.unl.pt/~cd
Received on Friday, 11 June 2010 11:37:17 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 11 June 2010 11:37:17 GMT