Issue 15140: ComputerExecutableRuleset should have Reference and Modifies associations with Class (prr-rtf) Source: TIBCO (Mr. Paul Vincent, pvincent(at)tibco.com) Nature: Clarification Severity: Significant Summary: In implementations of PRR in a UML tool, it has been found that rulesets (as well as rules) should be able to reference and modify a Class. This should be at the ComputerExecutableRuleset level per Figure 7.1. This should not impact any current implementations by adding this. Resolution: Revised Text: Actions taken: March 23, 2010: received issue Discussion: End of Annotations:===== m: webmaster@omg.org Date: 23 Mar 2010 14:04:13 -0500 To: Subject: Issue/Bug Report ******************************************************************************* Name: Paul Vincent Company: TIBCO Software mailFrom: pvincent@tibco.com Notification: Yes Specification: Production Rule Representation Section: 7 FormalNumber: formal/2009-12-01 Version: 1.0 RevisionDate: December 2009 Page: 5+ Title: ComputerExecutableRuleset should have Reference and Modifies associations with Class Nature: Clarification Severity: Significant test: 3qw8 B1: Report Issue Description: In implementations of PRR in a UML tool, it has been found that rulesets (as well as rules) should be able to reference and modify a Class. This should be at the ComputerExecutableRuleset level per Figure 7.1. From: "Owen, James C" To: Juergen Boldt , "issues@omg.org" , "prr-rtf@omg.org" Date: Tue, 6 Apr 2010 15:45:39 -0500 Subject: RE: issue 15140 -- PRR RTF issue Thread-Topic: issue 15140 -- PRR RTF issue Thread-Index: AcrVyARWVdi7viQOQkGEbkoRP2kVRgAAfe0g Accept-Language: en-US X-MS-Has-Attach: yes X-MS-TNEF-Correlator: acceptlanguage: en-US A Ruleset is a collection of Rules and would not need to modify a class/object. Only the Rule itself would need to do this. SDG jco From: Juergen Boldt [mailto:juergen@omg.org] Sent: Tuesday, April 06, 2010 3:30 PM To: issues@omg.org; prr-rtf@omg.org Subject: issue 15140 -- PRR RTF issue From: webmaster@omg.org Date: 23 Mar 2010 14:04:13 -0500 To: Subject: Issue/Bug Report ******************************************************************************* Name: Paul Vincent Company: TIBCO Software mailFrom: pvincent@tibco.com Notification: Yes Specification: Production Rule Representation Section: 7 FormalNumber: formal/2009-12-01 Version: 1.0 RevisionDate: December 2009 Page: 5+ Title: ComputerExecutableRuleset should have Reference and Modifies associations with Class Nature: Clarification Severity: Significant test: 3qw8 B1: Report Issue Description: In implementations of PRR in a UML tool, it has been found that rulesets (as well as rules) should be able to reference and modify a Class. This should be at the ComputerExecutableRuleset level per Figure 7.1. This should not impact any current implementations by adding this. Juergen Boldt Director, Member Services Object Management Group 140 Kendrick St Building A Suite 300 Needham, MA 02494 USA tel: +1 781 444 0404 x 132 fax: +1 781 444 0320 email: juergen@omg.org www.omg.org This email and any files transmitted with it are confidential, proprietary and intended solely for the individual or entity to whom they are addressed. If you have received this email in error please delete it immediately. ' X-IronPort-AV: E=Sophos;i="4.52,178,1270450800"; d="jpg'145?scan'145,208,217,145";a="9859536" Subject: RE: issue 15140 -- PRR RTF issue Date: Fri, 9 Apr 2010 09:16:58 -0700 X-MS-Has-Attach: yes X-MS-TNEF-Correlator: Thread-Topic: issue 15140 -- PRR RTF issue Thread-Index: AcrVyARWVdi7viQOQkGEbkoRP2kVRgAAfe0gABJn7xYAemUFIAAAUtyw From: "Paul Vincent" To: "Owen, James C" Cc: X-OriginalArrivalTime: 09 Apr 2010 16:17:01.0120 (UTC) FILETIME=[15D28C00:01CAD800] Hi James . thx for the reply . cc.ing th the RTF for other opinions. Like all issues the resolutions need to go to vote by the RTF members so its important that they are discussed in full... and I apologise for not ccing the original reply to you as I was using a Dumberry... I agree with your para 1. Para 2 is true at runtime and when designing the rules (which is why we have PRR associations between rules and the classes they reference). In *modelling*, using top-down, I may well specify that a ruleset is executed in some activity diagram (/ruleflow/BPM diagram), without filling in the rules. At the modelling level, I might well want to specify the sorts of information that the enclosed rules will use (and maybe update), when they are written. Therefore, although production rule runtime rulesets have no explicit association with the classes they modify (except maybe at the interface level for rulesets-as-services as in Blaze or JRules), there is a case for this association to exist and be defined explicitly at model-time. So this issue primarily supports the idea of top-down modeling. Does it also support bottom-up? Well, a tool could well generate the associations for a ruleset from the rules that are placed in it . so I think the bottom-uup approach is also covered (and this issue if implemented simply supports the idea of *documenting* the associations at the ruleset level, which as you say, really only exist at the rule level). [Disclosure and example: in TIBCO BusinessEvents, the rulesets are simply convenience folders with no runtime meaning whatsoever. However, the IDE can display links between rulesets and classes as a kind of cross-reference report, to show what affects what. So in other words, TIBCO *only* supports the bottom-up approach to development, but still is compatible with the idea of linking rulesets to classes...] Cheers Paul Vincent +1 650 206 2493 / mobile +44 781 493 7229 From: Owen, James C [mailto:JamesOwen@fico.com] Sent: 09 April 2010 17:01 To: Paul Vincent Subject: RE: issue 15140 -- PRR RTF issue Paul: When designing a rulebase from the bottom up, we start with rules and THEN break them up into rulesets. When designing from top down, we design (model) the ruleset as part of the overall architecture and then write the rules for each ruleset. When designing using a ruleflow, it.s much like a model of where a ruleset should go . but even then you don.t really know which objects will be needed until you actually begin to write the rules. Either way, it is ONLY the rules that determine which objects+attributes will be needed to enable a condition element to make a decision and which object+attribute will be modified in the action side. The ruleset itself does not touch nor relate to the objects, only the rules. I can model something from a top-down approach and write the rules later OR I can write the rules (and rulesets) and then design the model to show what we have done. Sad to say, being a fan of MDA, I have seen a lot more of the latter than the former. So, how do we, as a standardization organization, support both approaches? If we promulgate the top-down approach only we will be ignored by the industry architects and rule builders as some kind of academic hindrance. If we go with bottom-up approach then we will have betrayed our basic intents and purposes. Any suggestions? SDG James Owen FICO Product Management Senior Manager +1.817.656.4553 Office +1.214.684.5272 Cell From: Paul Vincent [mailto:pvincent@tibco.com] Sent: Wednesday, April 07, 2010 12:32 AM To: Owen, James C Subject: Re: issue 15140 -- PRR RTF issue ... At runtime But you don't model rules at runtime. Knowing the context of a rulleset - what facts it will use / touch - is what this is about. Regards, Paul Vincent, TIBCO, from my handheld -------------------------------------------------------------------------------- From: Owen, James C To: Juergen Boldt ; issues@omg.org ; prr-rtf@omg.org Sent: Tue Apr 06 13:45:39 2010 Subject: RE: issue 15140 -- PRR RTF issue A Ruleset is a collection of Rules and would not need to modify a class/object. Only the Rule itself would need to do this. SDG jco From: Juergen Boldt [mailto:juergen@omg.org] Sent: Tuesday, April 06, 2010 3:30 PM To: issues@omg.org; prr-rtf@omg.org Subject: issue 15140 -- PRR RTF issue From: webmaster@omg.org Date: 23 Mar 2010 14:04:13 -0500 To: Subject: Issue/Bug Report ******************************************************************************* Name: Paul Vincent Company: TIBCO Software mailFrom: pvincent@tibco.com Notification: Yes Specification: Production Rule Representation Section: 7 FormalNumber: formal/2009-12-01 Version: 1.0 RevisionDate: December 2009 Page: 5+ Title: ComputerExecutableRuleset should have Reference and Modifies associations with Class Nature: Clarification Severity: Significant test: 3qw8 B1: Report Issue Description: In implementations of PRR in a UML tool, it has been found that rulesets (as well as rules) should be able to reference and modify a Class. This should be at the ComputerExecutableRuleset level per Figure 7.1. This should not impact any current implementations by adding this. Juergen Boldt Director, Member Services Object Management Group 140 Kendrick St Building A Suite 300 Needham, MA 02494 USA tel: +1 781 444 0404 x 132 fax: +1 781 444 0320 email: juergen@omg.org www.omg.org This email and any files transmitted with it are confidential, proprietary and intended solely for the individual or entity to whom they are addressed. If you have received this email in error please delete it immediately. This email and any files transmitted with it are confidential, proprietary and intended solely for the individual or entity to whom they are addressed. If you have received this email in error please delete it immediately. This should not impact any current implementations by adding this.