RE: Approaching an XML syntax for RIF --> other

> E.g., for the condition in Harold's example to be actually 
> interchangeable between applications, the interchanging parties would 
> need to agree on the definitions of 'purchase' as a predicate and of 
> 'book' as a function, in most cases including a specification of the 
> types of the arguments etc.

That we mapped all of Rel, Fun, Ind, and Data to Con, as in

      <Atom>
        <Rel>purchase</Rel>
        <Var>Buyer</Var>
        <Var>Seller</Var>
        <Expr>
          <Fun>book</Fun>
          <Var>Author</Var>
          <Ind>LeRif</Ind>
        </Expr>
        <Data>$49</Data>
      </Atom>

-->

      <Atom>
        <Con>purchase</Con>
        <Var>Buyer</Var>
        <Var>Seller</Var>
        <Expr>
          <Con>book</Con>
          <Var>Author</Var>
          <Con>LeRif</Con>
        </Expr>
        <Con>$49</Con>
      </Atom>
 
was to accommodate multisorted logic, a previous design choice.
 
-- Harold


-----Original Message-----
From: public-rif-wg-request@w3.org [mailto:public-rif-wg-request@w3.org]
On Behalf Of Christian de Sainte Marie
Sent: Tuesday, January 30, 2007 9:22 AM
To: Paul Vincent
Cc: Dave Reynolds; Sandro Hawke; public-rif-wg@w3.org
Subject: Re: Approaching an XML syntax for RIF --> other


Paul Vincent wrote:
> 
> It seems to me that one way for RIF "widespread adoption" implies the
need for XML domain specific languages to adopt RIF for validation and
behaviour rules (for example vis a vis ACORD SPX, or MISMO BREW) amongst
others. In these cases:
> - some industry body has already specified the XML schema (/class
model)
> - RIF statements will need to be defined, in some tool, to conform
with said schema
> - an XML syntax for RIF rules will need to accommodate compatibility
with an existing schema.
> 
> This implies at least a step 4 below - for example -
> 4.  Verify the RIF output against any source definition of terms and
facts used in the rule expressions. 

Paul, you are thinking of the application's data model, right?

That is, the model for the application data (or references to 
application data) that makes the payload of the interchanged RIF rules, 
and that the two sides in an interchange must share (and agree on its 
interpretation) for the rules to have the same meaning for both parties 
(in addition to a common understanding of the semantics of the rules) 
(see picture 2 in the processing models section of UCR [1]).

E.g., for the condition in Harold's example to be actually 
interchangeable between applications, the interchanging parties would 
need to agree on the definitions of 'purchase' as a predicate and of 
'book' as a function, in most cases including a specification of the 
types of the arguments etc.

We never really discussed how RIF relates to that shared data model 
(although this is related to the discussion wrt data models as 
constraints), but we are probably coming close to a point where we will 
need to do so.

Paul's question/comment, as I understand it, is wrt the fact that some 
industry already have developed and agreed on a common data model for 
sharing data (or are in the process of doing so), e.g. in the form of an

XML Schema, and that they will want to be able to use that shared data 
model for the payload of the rules they need to interchange, and that 
this is likely to be a strong argument in favor or against adoption of
RIF.

Christian

[1] 
http://www.w3.org/2005/rules/wg/wiki/UCR/Usage_Scenarios_and_Processing_
Models?action=AttachFile&do=get&target=RIFprocessingmodel.png

Received on Tuesday, 30 January 2007 14:18:02 UTC