Issue 6184: UML 2 Super/Metamodel::Kernel::Packages/missing redefinition (uml2-superstructure-ftf) Source: Simula Research Laboratory (Mr. Bran Selic, selic(at)acm.org) Nature: Uncategorized Issue Severity: Summary: Kernel/Packages has association package:Package <--> ownedClassifier:Type without a redefinition while its ownedType in Basic and Constructs Resolution: This issue was resolved by the resolution to issue 6918. Revised Text: Actions taken: September 7, 2003: received issue December 2, 2004: closed issue Discussion: End of Annotations:===== To: issues@omg.org Subject: UML 2 Super/Metamodel::Kernel::Packages/missing redefinition X-Mailer: Lotus Notes Release 5.0.11 July 24, 2002 From: Branislav Selic Date: Sun, 7 Sep 2003 08:47:26 -0400 X-MIMETrack: Serialize by Router on D25ML05/25/M/IBM(Release 6.0.2CF1|June 9, 2003) at 09/07/2003 08:47:28, Serialize complete at 09/07/2003 08:47:28 Kernel/Packages has association package:Package <--> ownedClassifier:Type without a redefinition while its ownedType in Basic and Constructs. Bran Selic IBM Software Group -- Rational Software 770 Palladium Drive Kanata, Ontario, Canada K2V 1C8 ph. (613) 591-7915 fax (613) 599-3912 OMG Issue No: 6184 Title: UML 2 Super/Metamodel::Kernel::Packages/missing redefinition Source: Bran Selic, IBM (bselic@ca.ibm.com) Summary: Kernel/Packages has association package:Package <--> ownedClassifier:Type without a redefinition while its ownedType in Basic and Constructs Discussion: A fuller statement of the problem is that the Superstructure model for Package seems to be ignoring and not reusing the Infrastructure Constructs::Package with regard to a composition relationship whereby a Package owns [*] Type, a relationship central to both Infra and Super.. A summary of the analysis of this is: the Superstructure metamodel ignores the Infrastructure with regard to Package <.> Type composition, and seems to invent an altogether new composition, derived from another source, and perhaps by mistake gives it a name that matches the name of a similar composition in the infrastructure, causing confusion. Three possible resolutions, 1) keep the name 'ownedType' and revise the superstructure derivation and redefines relationships, so this composition redefines the one of the same name in the Infrastructure, or 2) keep the derivation and redefines relationships given in the superstructure, and revise the name in the superstructure to avoid the suggestion that this is the same relationship as the one in the superstructure. 3) keep both the derivation and the name, and state that in addition to everything else, this is a redefinition of the relationship in the Infrastructure Constructs. 3 is the best solution at this stage in the game as it does not involve any restructuring of the metamodel and acknowledges the importance of the infrastructure as the basis of the superstructure. Background documents: See Core::Basic, the Packages diagram in the Infrastructure FAS, section 10.4.1 Also see Core::Constructs, the Packages diagram, section 11.8.2 In disagreement with that, we have as the issue reports, in Superstructure Classes, section 7.13, the following confused situtation 1. The syntax diagram Figure 43 in section 7.13 shows Package with a composition association to Type of cardinality [*] but with a different name than that used in the infrastructure. Here it is 'ownedClassifier' 2. The syntax diagram shows this composition as subsetting a different composition, one of PackageableElement, so no relationship to the Infrastructure metamodel is recognized. 3. The text spec for section 7.13.1, Package associations on page 100, says: Compare with Figure 46 on page 139 of the Infrastructure, Namespaces, Section 11.6, where there is indeed the Namespace::ownedMember, which the ownedMember:PackageableElement of the Superstructure is said to redefine. To correct these problems replace the following sentences from the FAS: Old Text: Associations subsection for 7.13.1 on page 100 ownedType: Type[*] References the owned members that are Types. Subsets Package::ownedMember. New Text: ownedType: Type[*] References the owned members that are Types. Subsets Package::ownedMember. Redefines Constructs::Package::ownedType Disposition: Resolved Subject: Two Classes Issues for Ballot 20 Date: Wed, 21 Jul 2004 11:45:05 -0700 X-MS-Has-Attach: yes X-MS-TNEF-Correlator: Thread-Topic: Two Classes Issues for Ballot 20 Thread-Index: AcRumSVvPbiPVFL/Q8+3v1ZIOlTV4gAr/EoA From: "Karl Frank" To: "Branislav Selic" Cc: , X-OriginalArrivalTime: 21 Jul 2004 18:45:19.0202 (UTC) FILETIME=[DEF81C20:01C46F52] For consistency with Bran's proposal for resolving issue I have revised the terminology in the attached, using as the counterpart to the in Bran's proposal. We seem to have discovered the same solution to the same problem but used different terms for the non-terminal constituents of a property string. New Text: The BNF for property strings on association ends is: ::= '{' [ ',' ]* '}' ::= ( 'subsets' | 'redefines' | 'union' | 'ordered ' | 'unique' ) where and are names of user-provided properties and association ends found in the model context. If an association end is navigable, attribute-properties defined for attributes are legal as end-properties in the property string for that association end. -------------------------------------------------------------------- Please see attached. These both overlap infra and super, but the proposed resolution to 6184 does not require any change to infra. 6157 fixes a hole in BNF regarding property-string. - Karl To: "Karl Frank" Cc: mu2i-ftf@omg.org, uml2-superstructure-ftf@omg.org Subject: Re: Two Classes Issues for Ballot 20 X-Mailer: Lotus Notes Release 6.0.2CF1 June 9, 2003 From: Branislav Selic Date: Wed, 21 Jul 2004 23:39:39 -0400 X-MIMETrack: Serialize by Router on D25ML01/25/M/IBM(Release 6.0.2CF2|July 23, 2003) at 07/21/2004 23:39:42, Serialize complete at 07/21/2004 23:39:42 Karl, Your issue 6184 has already been resolved by the resolution to issue 6918. I suggest that you close this as a simple Duplicate. Folks, one of the reasons I work over the weekend to publish an updated version of the spec is precisely so that such duplication of effort does not occur. So, please check the latest version (up to and including ballot 18) before you commence on any resolutions. You can find it at: ftp://ftp.omg.org/pub/UML_2.0_superstructure-ftf/_UML2-Super.book.PDF.040719.zip Thanks, Bran "Karl Frank" 07/21/2004 01:39 PM To Branislav Selic/Ottawa/IBM@IBMCA cc , Subject Two Classes Issues for Ballot 20 Please see attached. These both overlap infra and super, but the proposed resolution to 6184 does not require any change to infra. 6157 fixes a hole in BNF regarding property-string. - Karl[attachment "ForBallot20_6157_6184.doc" deleted by Branislav Selic/Ottawa/IBM] e-mail: bselic@ca.ibm.com