- From: Ginsberg, Allen <AGINSBERG@imc.mitre.org>
- Date: Thu, 18 May 2006 12:03:01 -0400
- To: <public-rif-wg@w3.org>
Dear RIFers, This is a somewhat revised version of the CSFs for RIF Goal 1 (Exchange of Rules) that I sent to Frank, Paula, and the team chairs. Allen GOAL-1: Promote Exchange of Rules Critical Success Factors: CSF-1. An Extensible Standard for Representing Rule Language Families. One way of promoting the exchange of rules is to develop a standard rule language and 1) try to get everyone to use it, and 2) provide reliable tools for porting existing rules into the standard (and vice versa). For a large number of reasons, that is not feasible, and probably not desirable. A single (extensible) standard for representing Rule Languages Families is, however, feasible and will also facilitate the exchange of rules. By a "Rule Language Family" we mean a set of existing rule-languages that share an "overall common structure". A precise specification of what the latter means may be one of the outcomes of the work of this group, but, for now, we do not need to develop such a definition. For now, it is enough to acknowledge that such families exist. Various dialects of Prolog belong to such a family. It is also clear that many production-rule systems belong to the same family. Requirement: The RIF standard presupposes that given sets of rule-language families are provided. Initially, these sets may be specified purely extensionally (i.e., by enumerating the specific rule-langugage members). Also, it is possible that intially (in phase 1 of the RIF activity) that only one such set is provided. Requirement: For each rule-language family (that the RIF supports) the RIF must provide a vendor/platform-independent canonical format for representing rules in that family. To say that the RIF is an "extensible standard" means that as new members of existing rule language families, or entirely new rule languages families come into being (or are considered for inclusion in the RIF) it must be possible to revise or grow the core RIF specification to allow rules in the new language to have a canonical representation in the RIF. So this leads to the following requirement: Requirement: The RIF core must be structured in such a way that rules belonging to a new rule-language or an entirely new family of rule language can, if desired, be represented with a vendor/platform-independent canonical format without thereby making the canonical format for established RIF Rule Language Families invalid. CSF-2. Intra-Rule-Language-Family Rule-Exchange Support Exchanging rule-sets within the same rule-language family may or may not be trivial, depending on the family. Whatever the case, it is critical to the success of the RIF that exchange of rule-sets belonging to rule-languages in the same family should not only be possible, but that the use of RIF-based technology should make this process reliable and straightforward. 3. (**Not a CSF **) Inter-Rule-Language-Family Rule-Exchange Support Exchanging rule-sets belonging to different rule-language families may be assumed to be non-trivial in the general case. To what extent RIF-based technology, or any technology, can make such exchange a reliable and straightforward process is difficult to say. Obviously, therefore, we do not contemplate making such support a critical success factor for the RIF. Having said that, however, we do believe that the possibilities for developing technologies for this kind of rule exchange may be enhanced by the development of the RIF.
Received on Thursday, 18 May 2006 16:03:14 UTC