Re: The problem with nested rulesets (aka groups)

Gary,
There is no "problem" with groups of rules. And metadata is not just a
"comment".

> is that people will reasonably expect that they serve some more useful 
> purpose than merely a device to put a comment on a collection of rules.  
> For example, at least one real product (Haley) supports nested rulesets 
> (groups) and allows one to attach a condition to the group, with the 
> semantics that the condition is ANDed with the condition of each of the 
> group members.  This is useful because often a group of rules will all 
> be about the same frames or relations and you don't have to repeat that 
> in each group member. In PRD, one might also reasonably expect to attach 
> a priority or mutual exclusion constraint to the group.
> 
> If the sole purpose of groups in RIF is to avoid repeating a comment, 
> surely we can do that by putting the comment at the document (non-nested 
> ruleset) level, giving it an IRI or id, and then referring to it from 
> several rules.
>

Received on Sunday, 27 April 2008 06:34:46 UTC