Paul : -- data from many sources available for rule-inputs. (eg COBOL data) issue for SemWeb community, aka the future rules in cobol world --- gives them a path to the future -- [[ very impressed at "no defining rules"!! ]] Lots of *structure* rules, vs *behavior* (process) world in which vendors like me live, PLUS enforcement... which of these will rule interop format address? -- what is the use here? Ed -- It's okay to have multiple semantics, labeled, in XML I think - need to standardize what rule engines can do. horn datalog + NAF, plus... How to combine with FOL ? -- WALL INCLUSION. -- WG should standardize a surface (presentation) syntax. IT should be n-ary, etc. WE like frames. **************************************************************** BenG ommission: two communities learning about each other [Simple Business Rules] ---- [KR/Prolog/SemWeb] what's the nucleus in common between them a. Negation? (NON-MON) b. Actions? In or out? JP: yes, both. FA: yes, both. isn't actions just a superset of regular logic. Ed: we need some negation scheme. But carefully. Closed relations, specific sets, maybe? Pat: I'm scared because I've heard "NAF" to mean several things. "Need a way to say something is not in a set" -- that's perfectly classical. Undoing a shopping cart can be done in other ways. Ian: DLP is very small. It's NOT datalog, it's not LP. side-by-side picture is Disingenuous. The WG will take wind from RDF/OWL sails MikeU: Keep emphasis on use cases JosDB: lots of cases use rules engines, working with DLP fragment AnthonyF: stuck between sc. and charib. -- large research, large vendors. Interop rule systems. We do our small task well. cf CFP not asking for language, ... Interop of rules engines? 1. exchange form for rules, with possibly DIFFERENT semantics -- presentation of rules, etc. -- NOT modal, etc Waleed: hard to understand RuleML ** make output SIMPLER Vassilis: aside from NAF, lots of important stuff eg Uncertainly Mark: "Is there a basic subset that would satisfy the BR world, and somehow grow into the whole thing" ... CSMA: striking mismatch between use cases and candidtate technologies -- standard must be based on use cases -- must be able to last! -- RDF is more an inconvenience than a benefit; and YET if you want to Tim: do you mean rules in RDF? CSMA: should the common rules language be in RDF? use cases aren't ready for that Tim: for the semweb community, making a rule language WORK ON rdf is good, but exchanging rules in RDF is only an interesting academic excesice. Pat: n-ary predicates are required; anything else is dumbass coming from IKRA meeting -- they want the UNA !!! they want to use names to index facts. Bijan: Use cases don't solve theoretical issues, they dont solve reconsilliation of community issues; use cases are easy and hard. Hard to make them nail something. It's a nice mantra, but it doesn't necessarily mean progress. Danny: We will have use cases, so we don't have to dicsuss it any more Gary Ng: three camps dont communicate much feature table would be nice -- semantics vs. what it does not allow you to say then researches can be guided by features. HarryHalpin: Does business comm. know what Rule Light is? What about uncertainty, etc.., is that needed in Rule Light? [[ I think the Use Cases seen here were the most sophisticated ones ]] Suzette: something simpler Murray: common ground: beng: policy, CA rules need for sharable knowledge, beyond business rules. policy knowledge SWS, rules, process representation --- important problem, but I think we need to then look at the differrent problem sets; more general than use cases MikeUs: Glaring Ommission: ontology-to-ontoly mapping, for data integration RaviRaman: we heard about each C.T.'s, my CT can do anything yours can do trying to stay away from these issues is like putting your finger in the dyke ed's sugg: abs semantic, tagging rule sets, SWS, ... allows EXCHANGE of rules, given tags, and maybe that's enough to get started... no need to convert immediately and completely AdrianWalker: 2 glaring ommissions - when you run rules, declarative, and you start to run them distributed, you NEED explanations at a business analyst level, NO ONE will trust it - surface syntax that google will happily index; so people can find the rules Danny: "roogle" Mark: Back to AnthonyF --- subset? might not be If-Then, might be at arithmentic in conditional expressions (MathML), .... Tim pushes URIs -- maybe we can *just* do that as a standard for now. Maybe it would be fruitful to break the problem down into even smaller, complementary scopes. == panel summary, < 1 m == Don Chapin: Subset for BR and KR com -- SBVR use cases, rules, good input PV: This is a Broad Church for rules. This might be intractable. But I;m hopefull. I don's see any showstoppers. JRP: I would agree. I'm encouraged. I agree with Mark, on first common step. TimBL: Yes we started with some gaps of connectivity, but we've established a lot more connections. Certainly I've got a much better understand of BR. As the hours pass, our vocabulary has been missing less and less. Eg Fuzzy -- wont be in core, but very useful to see as an anchor point. Gap seems to be shrinking. Lots of work to do; go of digging in and comparing Something very simple and very useful Ed: if we have a WG -- we want to know in advance what the WG will do. OMG and RuleML will soldier on - need to make sure we're not competing Kurt: 1. if there is a W3C standard here (as Ed says) it would be absolutely WONDERFUL. Eff. of orgs will 2. simple rules to make me a good chess player Harold: comparison tables (a la Gary) would be very good to do builtings for cwm, forum[?] and swrl BR - SW communicating much beter. I liked SBVR. English rules, controlled english, ... see if we can abstract formal semantics Use energy from this WS to expand Website, toward WG JFA: 2nd Mark -- keep dynamic of WS -- define small subset to keep us together, working. find a way to keep momentum going Danny: what I heard -- ambitious goal: bring together lots of communities Users please stick with this, we need you in thie last hour I've heard "there's interest in doing something, it has to be well defined, it has to be simple, it has to be compatible and learn from work out there, and finally everyone has an interest in being able to evolve from what they are doing now, to take careful steps to interoperation" ultimate goal to find that path Here's how we'll do that. -> write a report, presentations are available, ... -> co-chairs and PC will be working on a report buy early june at the very latest. PC will review report. COrrections from everyone, but skip public review. Welcome feedback within about a month. -> Volunteers to work with us on developing Activity Proposal. Contact Sandro on that. We'll work out a process for that. -> We welcome fellows to come work at one of out host sites. Pat: Is this more like Precision Engineering, Carpentry, or Birck Laying Danny: Bricklaying