- From: Ginsberg, Allen <AGINSBERG@imc.mitre.org>
- Date: Tue, 21 Feb 2006 10:36:30 -0500
- To: "Dave Reynolds" <der@hplb.hpl.hp.com>, "Vincent, Paul D" <PaulVincent@fairisaac.com>
- Cc: "RIF" <public-rif-wg@w3.org>
Some comments on Dave's response to Paul: > No! I'm not proposing transport at all. In fact the mention of "RIF RPC" in > the design goals page worries me and I wanted to pick up on that sometime. I used the term "RIF RPC" in the design goals discussion to quickly illustrate or sketch some idea of a mechanism whereby a client process could make a RIF-based request to a server process. Nothing more was intended by the use of "RPC" in that context. > It seems like the combination of <rules being exchanged between different > parties> and <vendor-neutral format for rules> is not sufficient to place > such a case within RIF from your POV. If so that sounds like a useful > boundary case. Perhaps could could explain what it is that makes a rule > exchange sufficient of a rule "interchange" for RIF to become relevant? Under the definition of "rule interchange" as given in UCR document, there is a trivial sense in which an exchange of identical rules would qualify as interchange. However, in general "rule interchange" conveys the idea of some non-trivial mapping between distinct rule-representation languages or rule-systems. Nonetheless, it could be that simply exchanging identical rule-sets might fall under the RIF in a non-trivial way. For example, metadata tags associated with the rule sets might provide information concerning authorship,etc. Allen
Received on Tuesday, 21 February 2006 15:36:33 UTC