- From: Dave Reynolds <der@hplb.hpl.hp.com>
- Date: Thu, 09 Feb 2006 13:42:53 +0000
- To: Dieter Fensel <dieter.fensel@deri.org>
- CC: public-rif-wg@w3.org
Dieter Fensel wrote: > This completely neglects the idea of defining (as a side effect) a > useful rule language for the web. That is an interesting way of putting it and I think it goes to the heart of this "is RIF also a language" discussion. [What follows is not as clear as I'd like to be. To claw back some brevity I'll state as facts some things which are, at best, over simplifications.] If we look at the business rules market then we have a number of mature and successful products. One goal for RIF in that market is to enable users to move rules between systems, in which case RIF is for interchange between well-established systems. No vendor will change their language to move towards some invented RIF language. That's as it should be. In the semantic web area we have quite a large number of proposals and implemented systems with a very wide diversity but not the same maturity. This diversity inhibits rule developers because they may get locked into a particular rule implementation and can't easily share their rules with other web developers who want to reuse them for similar tasks. RIF could bring some coherence to this space. If there were to be a RIF profile which defined a reasonably simple, understandable, RDF/OWL compatible rule format and semantics that covered "80%" of "interesting" cases then: (a) that might give a profile that rule authors could write to in order to gain some implementation independence and ability to share rules; (b) at least some semantic web rule language implementers might migrate their systems to be more compatible with that profile, possibly even as far as doing a direct implementation. A nice feature of the way the RIF work is structured is that having such a profile would in no way preclude other extensions needed to handle exchange between systems with more advanced capabilities. Thus whilst the goal of RIF is interchange of rules, it *may* also have the side effect of providing the seed for a semantic web rule language. I don't think that side effect is in conflict with the main goal. Dave
Received on Thursday, 9 February 2006 13:43:12 UTC