- From: Christian de Sainte Marie <csma@ilog.fr>
- Date: Wed, 08 Mar 2006 17:19:59 +0100
- To: "Ginsberg, Allen" <AGINSBERG@imc.mitre.org>
- CC: RIF WG <public-rif-wg@w3.org>, john.hall@modelsys.com, Said Tabet <stabet@comcast.net>, Donald Chapin <donald.chapin@btinternet.com>
Allen, Ginsberg, Allen wrote: > > I beg to differ on this. [2] does make the case for interchange: it > specifically says that meta-data indicating that a rule is of the > human-machine interactive type is required to capture the meaning in > the RIF so rules can be interchanged accurately. Right. Actually, this is the point I tried to make: your example above illustrates a requirement on the RIF, rather than a usage of, or a need for the RIF. My point was simply that I prefered the scenario part of John's version as a use case, because it tells a story of several players between whom rules need be interchanged, like all other use cases do. And the story, illustrated with examples of rules, tells us, albeit implicitely, what kind of requirements are set by this sort of usage of the RIF. > Also, I have difficulty seeing a narrative "flow" or logical sequence > of premises and inferences in the original scenario. [2] attempts to > lay out the logic of the scenario in a step-by-step fashion. Agreed. Actually, I prefer the original scenario (the one in the MSWord version John sent yesterday). I modified [3] to copy that version instead, for your convenience. > It is > true that not all the specific requirements are explicitly covered, but > that is true of all the use-cases in the UCR. Right again. I did not mean to say that the UC should cover all the underlying requirements: only that a story of interchange illustrated what was required from the RIF as well, and made a better case for the use of the RIF (than the direct description and illustration of some of the requirements). > The key design goal of this use-case is to allow for interchange of > rules that are of a human-machine interaction type. What that amounts > to in terms of requirements (e.g., does it require deontic tags?) needs > to be figured out when we do the requirements. Agreed. This is clear from the scenario as well. Christian [3] http://www.w3.org/2005/rules/wg/wiki/UCR/Managing_Inter-Organizational_Business_Policies_and_Practices_-_Scenario
Received on Wednesday, 8 March 2006 16:19:39 UTC