Issue 5697: TaggedValue (cwm-rtf) Source: International Business Machines (Dr. Daniel T. Chang, dtchang(at)us.ibm.com) Nature: Uncategorized Issue Severity: Summary: CWM_1.0.dtd is missing the following element: <!ELEMENT CWM:ModelElement.taggedValue (CWM:TaggedValue)*> This element should be a child element of CWM:ModelElement and all elements that correspond to subclasses of ModelElement. Without this element, CWM_1.0.dtd does not conform to the CWM 1.0 Metamodel. According to the metamodel, any ModelElement (or its subclass) can own TaggedValue through the TaggedElement aggregation. Resolution: Revised Text: Actions taken: October 24, 2002: received issue Discussion: End of Annotations:===== From: "Tolbert, Doug M" To: "Juergen Boldt (E-mail)" Cc: Dan Chang Subject: RE: TaggedValue Date: Wed, 23 Oct 2002 16:06:11 -0500 X-Mailer: Internet Mail Service (5.5.2656.59) Juergen, Please log this in on the CWM issues list. Dan, soonest we can address this is the 1.2 release, which will be discussed in Washington DC next month. I recall that the intent was that there should be an XOR constraint on these associations, but there seems to be no such OCL in the spec. In the meantime, can you just insure in your implementation that TaggedValues don't participate in both associations simultaneously? Thanks, Doug -----Original Message----- From: Dan Chang [mailto:dtchang@us.ibm.com] Sent: Wednesday, October 23, 2002 11:10 AM To: Tolbert, Doug M Subject: TaggedValue Importance: High Doug, We have encountered a problem in our CWM enablement: CWM_1.0.dtd would not allow one to attach TaggedValue to ModelElement. I would appreciate your immediate attention. I believe the error is caused by associations defined on TaggedValue. In Figure 7-3-3 (CWM 1.0), TaggedValue is owned by both ModelElement and Stereotype. I don't think this should be allowed. The consequence is that, arbitrarily, in CWM_1.0.dtd, ModelElement does not have a child element ModelElement.taggedValue but Stereotype does have a child element Stereotype.requiredTag. The consequence is that one cannot simply "attach" TaggedValue to ModelElement as what TaggedValue is intended for. I believe the association between Stereotype and TaggedValue should be a normal association. Regards, Dan IBM Almaden Research Center/Silicon Valley Lab Notes: Dan Chang/Santa Teresa/IBM@IBMUS Internet: dtchang@us.ibm.com Phone: (408)-927-1714/(408)-463-2319 Office: ARC(G1-018)/SVL(D168) Subject: RE: TaggedValue To: "Tolbert, Doug M" Cc: "Juergen Boldt (E-mail)" X-Mailer: Lotus Notes Release 5.0.7 March 21, 2001 From: "Dan Chang" Date: Wed, 23 Oct 2002 15:41:33 -0700 X-MIMETrack: Serialize by Router on D03NM800/03/M/IBM(Release 5.0.10 |March 22, 2002) at 10/23/2002 04:41:35 PM Doug, Thank you for confirming the error. The problem right now is: CWM_1.0.dtd (presumably the same for CWM_1.1.dtd) does not allow anyone to associate TaggedValue to ModelElement. Our implementation does not use Stereotype. It only uses TaggedValue. However, the XML file that we export will not validate against CWM_1.0.dtd because of this error. Anyone else who uses TaggedValue alone should encounter the same problem. BTW, at present I do not plan to attend the Washington DC meeting because of my main work load. If so, I hope you will bring up this issue for discussion and resolution because of its significance. (I assume any resolution will be voted on thereafter by e-mail.) Thanks much. Regards, Dan IBM Almaden Research Center/Silicon Valley Lab Notes: Dan Chang/Santa Teresa/IBM@IBMUS Internet: dtchang@us.ibm.com Phone: (408)-927-1714/(408)-463-2319 Office: ARC(G1-018)/SVL(D168) "Tolbert, Doug M" sys.com> cc: Dan Chang Subject: RE: TaggedValue 10/23/2002 02:06 PM Juergen, Please log this in on the CWM issues list. Dan, soonest we can address this is the 1.2 release, which will be discussed in Washington DC next month. I recall that the intent was that there should be an XOR constraint on these associations, but there seems to be no such OCL in the spec. In the meantime, can you just insure in your implementation that TaggedValues don't participate in both associations simultaneously? Thanks, Doug -----Original Message----- From: Dan Chang [mailto:dtchang@us.ibm.com] Sent: Wednesday, October 23, 2002 11:10 AM To: Tolbert, Doug M Subject: TaggedValue Importance: High Doug, We have encountered a problem in our CWM enablement: CWM_1.0.dtd would not allow one to attach TaggedValue to ModelElement. I would appreciate your immediate attention. I believe the error is caused by associations defined on TaggedValue. In Figure 7-3-3 (CWM 1.0), TaggedValue is owned by both ModelElement and Stereotype. I don't think this should be allowed. The consequence is that, arbitrarily, in CWM_1.0.dtd, ModelElement does not have a child element ModelElement.taggedValue but Stereotype does have a child element Stereotype.requiredTag. The consequence is that one cannot simply "attach" TaggedValue to ModelElement as what TaggedValue is intended for. I believe the association between Stereotype and TaggedValue should be a normal association. Regards, Dan IBM Almaden Research Center/Silicon Valley Lab Notes: Dan Chang/Santa Teresa/IBM@IBMUS Internet: dtchang@us.ibm.com Phone: (408)-927-1714/(408)-463-2319 Office: ARC(G1-018)/SVL(D168) Subject: RE: issue 5697 -- CWM RTF issue Date: Thu, 24 Oct 2002 15:05:43 +0100 X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: issue 5697 -- CWM RTF issue Thread-Index: AcJ7YHZvUwgTTkk0R4y8pFFM7uiKNgAAeeDw From: "Pete Rivett" To: X-MIME-Autoconverted: from quoted-printable to 8bit by emerald.omg.org id g9OE5Xm13137 It's fine in MOF for a class to potentially be owned via more than one composition: the restriction is that a single instance can only have one owner. So that means that the compositions must be 0..1 rather than 1..1 which they are in this case in CWM. As it stands it means that a ModelElement can own tags and a Stereotype can own tags but the same tag cannot be owned by both. The particular issue of the missing reference from ModelElement to TaggedValue I thought was fixed by issue 4408 (which was dealt with as an urgent Issue). Pete > -----Original Message----- > From: Juergen Boldt [mailto:juergen@omg.org] > Sent: 24 October 2002 14:17 > To: issues@omg.org; cwm-rtf@omg.org > Subject: issue 5697 -- CWM RTF issue > > > This is issue # 5697 Dan Chang [mailto:dtchang@us.ibm.com] > > TaggedValue > > We have encountered a problem in our CWM enablement: > CWM_1.0.dtd would not allow one to attach TaggedValue to > ModelElement. I would appreciate your immediate attention. > > > I believe the error is caused by associations defined on > TaggedValue. In Figure 7-3-3 (CWM 1.0), TaggedValue is owned > by both ModelElement and Stereotype. I don't think this > should be allowed. The consequence is that, arbitrarily, in > CWM_1.0.dtd, ModelElement does not have a child element > ModelElement.taggedValue but Stereotype does have a child > element Stereotype.requiredTag. The consequence is that one > cannot simply "attach" TaggedValue to ModelElement as what > TaggedValue is intended for. I believe the association > between Stereotype and TaggedValue should be a normal association > > ================================= > Juergen Boldt > Director, Member Services > > Object Management Group > 250 First Avenue, Suite 100 > Needham, MA 02494 > > Tel. +1 781 444 0404 ext. 132 > Fax: +1 781 444 0320 > email: juergen@omg.org > www www.omg.org > > ================================ > > The information contained in this email and any attached files are confidential and intended solely for the addressee(s). The e-mail may be legally privileged or prohibited from disclosure and unauthorised use. If you are not the named addressee you may not use, copy or disclose this information to any other person. If you received this message in error please notify the sender immediately. Any views or opinions presented here may be solely those of the originator and do not necessarily reflect those of the Company. Subject: RE: issue 5697 -- CWM RTF issue To: cwm-rtf@omg.org X-Mailer: Lotus Notes Release 5.0.9a January 7, 2002 From: John_Poole@hyperion.com Date: Thu, 24 Oct 2002 13:47:12 -0400 X-MIMETrack: Serialize by Router on Stmfd-Gateway1/na/Hyperion(Release 5.0.9a |January 7, 2002) at 10/24/2002 01:47:13 PM Hi all, I think I know (approximately) where the problem lies... By way of comparison, there is an identical situation in the CWM OLAP metamodel, in which an instance of DimensionDeployment may be owned by at most one instance of HierarchyLevelAssociation or by at most one instance of ValueBasedHierarchy, but not by both at the same time, and this situation is formally expressed by an OCL constraint of the form: context DimensionDeployment inv: self.hierarchyLevelAssociation->isEmpty xor self.valueBasedHierarchy->isEmpty (Of course, the constraint is just documentation, is not embedded in the Rose model, and could not have possibly influenced the DTD generation in any way). The main thing that is different between this part of the OLAP model and the Core classes ModelElement and TaggedValue is the absense of an explicit reference on ModelElement corresponding to the TaggedElement::taggedValue association end, where as both HierarchyLevelAssociation and ValueBasedHierarchy both have explicit references to DimensionDeployment. There is another difference in that both the association ends touching DimenisonDeployment are also ordered, whereas this is not the case with the taggedValue / requiredTag ends. Other than these two points, I see nothing else in the class properties and/or features that might otherwise distinguish the two situations. So, I suspect that the absense of a TaggedValue reference on ModelElement is the root cause of the problem. Is it possible that the DTD generation process was driven off the references, and not the association ends? Or could this point to a possible error in XMI itself? In any event, the absence of a TaggedValue reference in ModelElement had previously been raised by issue 4408 and certainly is indeed problematic for any CWM implementation not using MOF. Also, we will need to update the CWM specification to include an XOR constraint in the Core model for ownership of TaggedValue. SO IN SUMMARY: I believe this is either a DTD generator problem or possibly an XMI issue. It is not a problem in the CWM model itself. I think Dan's current open issue should be amended to require that an appropriate XOR constraint be added to the spec., and that the DTD be fixed. In the meantime, someone needs to investigate the DTD generator that was used (I'd offer to do it myself, but I do not own a copy of the XMI tool used to produce the CWM 1.0 DTD). Best regards, John "Pete Rivett" tive.com> cc: Subject: RE: issue 5697 -- CWM RTF issue 10/24/2002 10:05 AM It's fine in MOF for a class to potentially be owned via more than one composition: the restriction is that a single instance can only have one owner. So that means that the compositions must be 0..1 rather than 1..1 which they are in this case in CWM. As it stands it means that a ModelElement can own tags and a Stereotype can own tags but the same tag cannot be owned by both. The particular issue of the missing reference from ModelElement to TaggedValue I thought was fixed by issue 4408 (which was dealt with as an urgent Issue). Pete > -----Original Message----- > From: Juergen Boldt [mailto:juergen@omg.org] > Sent: 24 October 2002 14:17 > To: issues@omg.org; cwm-rtf@omg.org > Subject: issue 5697 -- CWM RTF issue > > > This is issue # 5697 Dan Chang [mailto:dtchang@us.ibm.com] > > TaggedValue > > We have encountered a problem in our CWM enablement: > CWM_1.0.dtd would not allow one to attach TaggedValue to > ModelElement. I would appreciate your immediate attention. > > > I believe the error is caused by associations defined on > TaggedValue. In Figure 7-3-3 (CWM 1.0), TaggedValue is owned > by both ModelElement and Stereotype. I don't think this > should be allowed. The consequence is that, arbitrarily, in > CWM_1.0.dtd, ModelElement does not have a child element > ModelElement.taggedValue but Stereotype does have a child > element Stereotype.requiredTag. The consequence is that one > cannot simply "attach" TaggedValue to ModelElement as what > TaggedValue is intended for. I believe the association > between Stereotype and TaggedValue should be a normal association > > ================================= > Juergen Boldt > Director, Member Services > > Object Management Group > 250 First Avenue, Suite 100 > Needham, MA 02494 > > Tel. +1 781 444 0404 ext. 132 > Fax: +1 781 444 0320 > email: juergen@omg.org > www www.omg.org > > ================================ > > The information contained in this email and any attached files are confidential and intended solely for the addressee(s). The e-mail may be legally privileged or prohibited from disclosure and unauthorised use. If you are not the named addressee you may not use, copy or disclose this information to any other person. If you received this message in error please notify the sender immediately. Any views or opinions presented here may be solely those of the originator and do not necessarily reflect those of the Company. Subject: Re: issue 5697 -- CWM RTF issue To: Juergen Boldt Cc: cwm-rtf@omg.org, issues@omg.org X-Mailer: Lotus Notes Release 5.0.7 March 21, 2001 From: "Dan Chang" Date: Thu, 24 Oct 2002 12:28:49 -0700 X-MIMETrack: Serialize by Router on D03NM800/03/M/IBM(Release 5.0.10 |March 22, 2002) at 10/24/2002 01:28:49 PM Jurgen, Please replace the text with the following, which is explicit on what the problem is. Sorry for the trouble. CWM_1.0.dtd is missing the following element: This element should be a child element of CWM:ModelElement and all elements that correspond to subclasses of ModelElement. Without this element, CWM_1.0.dtd does not conform to the CWM 1.0 Metamodel. According to the metamodel, any ModelElement (or its subclass) can own TaggedValue through the TaggedElement aggregation. Regards, Dan IBM Almaden Research Center/Silicon Valley Lab Notes: Dan Chang/Santa Teresa/IBM@IBMUS Internet: dtchang@us.ibm.com Phone: (408)-927-1714/(408)-463-2319 Office: ARC(G1-018)/SVL(D168) Juergen Boldt To: issues@omg.org, cwm-rtf@omg.org cc: 10/24/2002 06:16 Subject: issue 5697 -- CWM RTF issue AM This is issue # 5697 Dan Chang [mailto:dtchang@us.ibm.com] TaggedValue We have encountered a problem in our CWM enablement: CWM_1.0.dtd would not allow one to attach TaggedValue to ModelElement. I would appreciate your immediate attention. I believe the error is caused by associations defined on TaggedValue. In Figure 7-3-3 (CWM 1.0), TaggedValue is owned by both ModelElement and Stereotype. I don't think this should be allowed. The consequence is that, arbitrarily, in CWM_1.0.dtd, ModelElement does not have a child element ModelElement.taggedValue but Stereotype does have a child element Stereotype.requiredTag. The consequence is that one cannot simply "attach" TaggedValue to ModelElement as what TaggedValue is intended for. I believe the association between Stereotype and TaggedValue should be a normal association ================================= Juergen Boldt Director, Member Services Object Management Group 250 First Avenue, Suite 100 Needham, MA 02494 Tel. +1 781 444 0404 ext. 132 Fax: +1 781 444 0320 email: juergen@omg.org www www.omg.org ================================ Subject: RE: issue 5697 -- CWM RTF issue Date: Fri, 25 Oct 2002 12:26:50 +0100 X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: issue 5697 -- CWM RTF issue Thread-Index: AcJ7hewThWbGSq1pRS6BSqEjA6wijQAkgCvQ From: "Pete Rivett" To: , X-MIME-Autoconverted: from quoted-printable to 8bit by emerald.omg.org id g9PBQnm08147 With respect to the DTD, I don't think there's a generator problem: I think it's just that we have not generated/issued a CWM 1.0 DTD with the fix to issue 4408 applied. In fact I don't think we've issued ANY updated CWM 1.0 artifacts (mdl, XMI, IDL) with just that emergency fix applied. Is that something we want to do? Then there's the issue of how we catalog it and distinguish it from the previous CWM 1.0 DTD. Pete > -----Original Message----- > From: John_Poole@hyperion.com [mailto:John_Poole@hyperion.com] > Sent: 24 October 2002 18:47 > To: cwm-rtf@omg.org > Subject: RE: issue 5697 -- CWM RTF issue > > > > Hi all, > > I think I know (approximately) where the problem lies... > > By way of comparison, there is an identical situation in the > CWM OLAP metamodel, in which an instance of > DimensionDeployment may be owned by at most one instance of > HierarchyLevelAssociation or by at most one instance of > ValueBasedHierarchy, but not by both at the same time, and > this situation is formally expressed by an OCL constraint of the form: > > context DimensionDeployment inv: > self.hierarchyLevelAssociation->isEmpty xor > self.valueBasedHierarchy->isEmpty > > (Of course, the constraint is just documentation, is not > embedded in the Rose model, and could not have possibly > influenced the DTD generation in any way). > > The main thing that is different between this part of the > OLAP model and the Core classes ModelElement and TaggedValue > is the absense of an explicit reference on ModelElement > corresponding to the TaggedElement::taggedValue association > end, where as both HierarchyLevelAssociation and > ValueBasedHierarchy both have explicit references to > DimensionDeployment. There is another difference in that both > the association ends touching DimenisonDeployment are also > ordered, whereas this is not the case with the taggedValue / > requiredTag ends. Other than these two points, I see nothing > else in the class properties and/or features that might > otherwise distinguish the two situations. > > So, I suspect that the absense of a TaggedValue reference on > ModelElement is the root cause of the problem. Is it > possible that the DTD generation process was driven off the > references, and not the association ends? Or could this > point to a possible error in XMI itself? > > In any event, the absence of a TaggedValue reference in > ModelElement had previously been raised by issue 4408 and > certainly is indeed problematic for any CWM implementation > not using MOF. > > Also, we will need to update the CWM specification to include > an XOR constraint in the Core model for ownership of TaggedValue. > > SO IN SUMMARY: I believe this is either a DTD generator > problem or possibly an XMI issue. It is not a problem in the > CWM model itself. I think Dan's current open issue should be > amended to require that an appropriate XOR constraint be > added to the spec., and that the DTD be fixed. In the > meantime, someone needs to investigate the DTD generator that > was used (I'd offer to do it myself, but I do not own a copy > of the XMI tool used to produce the CWM 1.0 DTD). > > Best regards, > John > > > > > > > "Pete Rivett" > > > tive.com> cc: > > Subject: RE: > issue 5697 -- CWM RTF issue > 10/24/2002 10:05 > > AM > > > > > > > > > > It's fine in MOF for a class to potentially be owned via more than one > composition: the restriction is that a single instance can > only have one owner. So that means that the compositions must > be 0..1 rather than 1..1 which they are in this case in CWM. > As it stands it means that a ModelElement can own tags and a > Stereotype can own tags but the same tag cannot be owned by both. > > The particular issue of the missing reference from > ModelElement to TaggedValue I thought was fixed by issue 4408 > (which was dealt with as an urgent Issue). > > Pete > > > -----Original Message----- > > From: Juergen Boldt [mailto:juergen@omg.org] > > Sent: 24 October 2002 14:17 > > To: issues@omg.org; cwm-rtf@omg.org > > Subject: issue 5697 -- CWM RTF issue > > > > > > This is issue # 5697 Dan Chang [mailto:dtchang@us.ibm.com] > > > > TaggedValue > > > > We have encountered a problem in our CWM enablement: > CWM_1.0.dtd would > > not allow one to attach TaggedValue to ModelElement. I would > > appreciate your immediate attention. > > > > > > I believe the error is caused by associations defined on > TaggedValue. > > In Figure 7-3-3 (CWM 1.0), TaggedValue is owned by both > ModelElement > > and Stereotype. I don't think this should be allowed. The > consequence > > is that, arbitrarily, in CWM_1.0.dtd, ModelElement does not have a > > child element ModelElement.taggedValue but Stereotype does have a > > child element Stereotype.requiredTag. The consequence is that one > > cannot simply "attach" TaggedValue to ModelElement as what > > TaggedValue is intended for. I believe the association > > between Stereotype and TaggedValue should be a normal association > > > > ================================= > > Juergen Boldt > > Director, Member Services > > > > Object Management Group > > 250 First Avenue, Suite 100 > > Needham, MA 02494 > > > > Tel. +1 781 444 0404 ext. 132 > > Fax: +1 781 444 0320 > > email: juergen@omg.org > > www www.omg.org > > > > ================================ > > > > > > The information contained in this email and any attached > files are confidential and intended solely for the > addressee(s). The e-mail may be legally privileged or > prohibited from disclosure and unauthorised use. > > If you are not the named addressee you may not use, copy or > disclose this information to any other person. If you > received this message in error please notify the sender immediately. > > Any views or opinions presented here may be solely those of > the originator and do not necessarily reflect those of the Company. > > > > > The information contained in this email and any attached files are confidential and intended solely for the addressee(s). The e-mail may be legally privileged or prohibited from disclosure and unauthorised use. If you are not the named addressee you may not use, copy or disclose this information to any other person. If you received this message in error please notify the sender immediately. Any views or opinions presented here may be solely those of the originator and do not necessarily reflect those of the Company. From: "Tolbert, Doug M" To: Pete Rivett , John_Poole@hyperion.com, cwm-rtf@omg.org Subject: RE: issue 5697 -- CWM RTF issue Date: Fri, 25 Oct 2002 10:58:49 -0500 X-Mailer: Internet Mail Service (5.5.2656.59) I'm not sure OMG processes would allow us to formally issue a corrected 1.0 DTD. I think, rather, that one would be expected to move to 1.1 instead to get the fixes. Pete, with your AB hat on, do you know what the official OMG position on this would be? If not, maybe we should start asking Andrew. Doug -----Original Message----- From: Pete Rivett [mailto:Pete.Rivett@adaptive.com] Sent: Friday, October 25, 2002 4:27 AM To: John_Poole@hyperion.com; cwm-rtf@omg.org Subject: RE: issue 5697 -- CWM RTF issue With respect to the DTD, I don't think there's a generator problem: I think it's just that we have not generated/issued a CWM 1.0 DTD with the fix to issue 4408 applied. In fact I don't think we've issued ANY updated CWM 1.0 artifacts (mdl, XMI, IDL) with just that emergency fix applied. Is that something we want to do? Then there's the issue of how we catalog it and distinguish it from the previous CWM 1.0 DTD. Pete > -----Original Message----- > From: John_Poole@hyperion.com [mailto:John_Poole@hyperion.com] > Sent: 24 October 2002 18:47 > To: cwm-rtf@omg.org > Subject: RE: issue 5697 -- CWM RTF issue > > > > Hi all, > > I think I know (approximately) where the problem lies... > > By way of comparison, there is an identical situation in the > CWM OLAP metamodel, in which an instance of > DimensionDeployment may be owned by at most one instance of > HierarchyLevelAssociation or by at most one instance of > ValueBasedHierarchy, but not by both at the same time, and > this situation is formally expressed by an OCL constraint of the form: > > context DimensionDeployment inv: > self.hierarchyLevelAssociation->isEmpty xor > self.valueBasedHierarchy->isEmpty > > (Of course, the constraint is just documentation, is not > embedded in the Rose model, and could not have possibly > influenced the DTD generation in any way). > > The main thing that is different between this part of the > OLAP model and the Core classes ModelElement and TaggedValue > is the absense of an explicit reference on ModelElement > corresponding to the TaggedElement::taggedValue association > end, where as both HierarchyLevelAssociation and > ValueBasedHierarchy both have explicit references to > DimensionDeployment. There is another difference in that both > the association ends touching DimenisonDeployment are also > ordered, whereas this is not the case with the taggedValue / > requiredTag ends. Other than these two points, I see nothing > else in the class properties and/or features that might > otherwise distinguish the two situations. > > So, I suspect that the absense of a TaggedValue reference on > ModelElement is the root cause of the problem. Is it > possible that the DTD generation process was driven off the > references, and not the association ends? Or could this > point to a possible error in XMI itself? > > In any event, the absence of a TaggedValue reference in > ModelElement had previously been raised by issue 4408 and > certainly is indeed problematic for any CWM implementation > not using MOF. > > Also, we will need to update the CWM specification to include > an XOR constraint in the Core model for ownership of TaggedValue. > > SO IN SUMMARY: I believe this is either a DTD generator > problem or possibly an XMI issue. It is not a problem in the > CWM model itself. I think Dan's current open issue should be > amended to require that an appropriate XOR constraint be > added to the spec., and that the DTD be fixed. In the > meantime, someone needs to investigate the DTD generator that > was used (I'd offer to do it myself, but I do not own a copy > of the XMI tool used to produce the CWM 1.0 DTD). > > Best regards, > John > > > > > > > "Pete Rivett" > > > tive.com> cc: > > Subject: RE: > issue 5697 -- CWM RTF issue > 10/24/2002 10:05 > > AM > > > > > > > > > > It's fine in MOF for a class to potentially be owned via more than one > composition: the restriction is that a single instance can > only have one owner. So that means that the compositions must > be 0..1 rather than 1..1 which they are in this case in CWM. > As it stands it means that a ModelElement can own tags and a > Stereotype can own tags but the same tag cannot be owned by both. > > The particular issue of the missing reference from > ModelElement to TaggedValue I thought was fixed by issue 4408 > (which was dealt with as an urgent Issue). > > Pete > > > -----Original Message----- > > From: Juergen Boldt [mailto:juergen@omg.org] > > Sent: 24 October 2002 14:17 > > To: issues@omg.org; cwm-rtf@omg.org > > Subject: issue 5697 -- CWM RTF issue > > > > > > This is issue # 5697 Dan Chang [mailto:dtchang@us.ibm.com] > > > > TaggedValue > > > > We have encountered a problem in our CWM enablement: > CWM_1.0.dtd would > > not allow one to attach TaggedValue to ModelElement. I would > > appreciate your immediate attention. > > > > > > I believe the error is caused by associations defined on > TaggedValue. > > In Figure 7-3-3 (CWM 1.0), TaggedValue is owned by both > ModelElement > > and Stereotype. I don't think this should be allowed. The > consequence > > is that, arbitrarily, in CWM_1.0.dtd, ModelElement does not have a > > child element ModelElement.taggedValue but Stereotype does have a > > child element Stereotype.requiredTag. The consequence is that one > > cannot simply "attach" TaggedValue to ModelElement as what > > TaggedValue is intended for. I believe the association > > between Stereotype and TaggedValue should be a normal association > > > > ================================= > > Juergen Boldt > > Director, Member Services > > > > Object Management Group > > 250 First Avenue, Suite 100 > > Needham, MA 02494 > > > > Tel. +1 781 444 0404 ext. 132 > > Fax: +1 781 444 0320 > > email: juergen@omg.org > > www www.omg.org > > > > ================================ > > > > > > The information contained in this email and any attached > files are confidential and intended solely for the > addressee(s). The e-mail may be legally privileged or > prohibited from disclosure and unauthorised use. > > If you are not the named addressee you may not use, copy or > disclose this information to any other person. If you > received this message in error please notify the sender immediately. > > Any views or opinions presented here may be solely those of > the originator and do not necessarily reflect those of the Company. > > > > > The information contained in this email and any attached files are confidential and intended solely for the addressee(s). The e-mail may be legally privileged or prohibited from disclosure and unauthorised use. If you are not the named addressee you may not use, copy or disclose this information to any other person. If you received this message in error please notify the sender immediately. Any views or opinions presented here may be solely those of the originator and do not necessarily reflect those of the Company. Subject: RE: issue 5697 -- CWM RTF issue To: "Tolbert, Doug M" Cc: cwm-rtf@omg.org, John_Poole@hyperion.com, Pete Rivett X-Mailer: Lotus Notes Release 5.0.8 June 18, 2001 From: "Sridhar Iyengar" Date: Fri, 25 Oct 2002 12:31:08 -0400 X-MIMETrack: Serialize by Router on D03NM120/03/M/IBM(Release 5.0.10 |March 22, 2002) at 10/25/2002 10:32:16 AM I have already sent a question to Andrew. There is an 'emergency fix' process to fix critical bugs. In general a proposal to fix a vote can be taken. Sridhar --------------------------------------------------------------------------------------------------------- Sridhar Iyengar Distinguished Engineer Application and Integration Middleware IBM : Research Triangle Park, NC E-Mail: siyengar@us.ibm.com Phone: (919) 486-1768 --------------------------------------------------------------------------------------------------------- |---------+----------------------------> | | "Tolbert, Doug M"| | | | | | | | | 10/25/2002 11:58 | | | AM | | | | |---------+----------------------------> >---------------------------------------------------------------------------------------------------------------------------------------------| | | | To: Pete Rivett , John_Poole@hyperion.com, cwm-rtf@omg.org | | cc: | | Subject: RE: issue 5697 -- CWM RTF issue | | | | | >---------------------------------------------------------------------------------------------------------------------------------------------| I'm not sure OMG processes would allow us to formally issue a corrected 1.0 DTD. I think, rather, that one would be expected to move to 1.1 instead to get the fixes. Pete, with your AB hat on, do you know what the official OMG position on this would be? If not, maybe we should start asking Andrew. Doug -----Original Message----- From: Pete Rivett [mailto:Pete.Rivett@adaptive.com] Sent: Friday, October 25, 2002 4:27 AM To: John_Poole@hyperion.com; cwm-rtf@omg.org Subject: RE: issue 5697 -- CWM RTF issue With respect to the DTD, I don't think there's a generator problem: I think it's just that we have not generated/issued a CWM 1.0 DTD with the fix to issue 4408 applied. In fact I don't think we've issued ANY updated CWM 1.0 artifacts (mdl, XMI, IDL) with just that emergency fix applied. Is that something we want to do? Then there's the issue of how we catalog it and distinguish it from the previous CWM 1.0 DTD. Pete > -----Original Message----- > From: John_Poole@hyperion.com [mailto:John_Poole@hyperion.com] > Sent: 24 October 2002 18:47 > To: cwm-rtf@omg.org > Subject: RE: issue 5697 -- CWM RTF issue > > > > Hi all, > > I think I know (approximately) where the problem lies... > > By way of comparison, there is an identical situation in the > CWM OLAP metamodel, in which an instance of > DimensionDeployment may be owned by at most one instance of > HierarchyLevelAssociation or by at most one instance of > ValueBasedHierarchy, but not by both at the same time, and > this situation is formally expressed by an OCL constraint of the form: > > context DimensionDeployment inv: > self.hierarchyLevelAssociation->isEmpty xor > self.valueBasedHierarchy->isEmpty > > (Of course, the constraint is just documentation, is not > embedded in the Rose model, and could not have possibly > influenced the DTD generation in any way). > > The main thing that is different between this part of the > OLAP model and the Core classes ModelElement and TaggedValue > is the absense of an explicit reference on ModelElement > corresponding to the TaggedElement::taggedValue association > end, where as both HierarchyLevelAssociation and > ValueBasedHierarchy both have explicit references to > DimensionDeployment. There is another difference in that both > the association ends touching DimenisonDeployment are also > ordered, whereas this is not the case with the taggedValue / > requiredTag ends. Other than these two points, I see nothing > else in the class properties and/or features that might > otherwise distinguish the two situations. > > So, I suspect that the absense of a TaggedValue reference on > ModelElement is the root cause of the problem. Is it > possible that the DTD generation process was driven off the > references, and not the association ends? Or could this > point to a possible error in XMI itself? > > In any event, the absence of a TaggedValue reference in > ModelElement had previously been raised by issue 4408 and > certainly is indeed problematic for any CWM implementation > not using MOF. > > Also, we will need to update the CWM specification to include > an XOR constraint in the Core model for ownership of TaggedValue. > > SO IN SUMMARY: I believe this is either a DTD generator > problem or possibly an XMI issue. It is not a problem in the > CWM model itself. I think Dan's current open issue should be > amended to require that an appropriate XOR constraint be > added to the spec., and that the DTD be fixed. In the > meantime, someone needs to investigate the DTD generator that > was used (I'd offer to do it myself, but I do not own a copy > of the XMI tool used to produce the CWM 1.0 DTD). > > Best regards, > John > > > > > > > "Pete Rivett" > > > tive.com> cc: > > Subject: RE: > issue 5697 -- CWM RTF issue > 10/24/2002 10:05 > > AM > > > > > > > > > > It's fine in MOF for a class to potentially be owned via more than one > composition: the restriction is that a single instance can > only have one owner. So that means that the compositions must > be 0..1 rather than 1..1 which they are in this case in CWM. > As it stands it means that a ModelElement can own tags and a > Stereotype can own tags but the same tag cannot be owned by both. > > The particular issue of the missing reference from > ModelElement to TaggedValue I thought was fixed by issue 4408 > (which was dealt with as an urgent Issue). > > Pete > > > -----Original Message----- > > From: Juergen Boldt [mailto:juergen@omg.org] > > Sent: 24 October 2002 14:17 > > To: issues@omg.org; cwm-rtf@omg.org > > Subject: issue 5697 -- CWM RTF issue > > > > > > This is issue # 5697 Dan Chang [mailto:dtchang@us.ibm.com] > > > > TaggedValue > > > > We have encountered a problem in our CWM enablement: > CWM_1.0.dtd would > > not allow one to attach TaggedValue to ModelElement. I would > > appreciate your immediate attention. > > > > > > I believe the error is caused by associations defined on > TaggedValue. > > In Figure 7-3-3 (CWM 1.0), TaggedValue is owned by both > ModelElement > > and Stereotype. I don't think this should be allowed. The > consequence > > is that, arbitrarily, in CWM_1.0.dtd, ModelElement does not have a > > child element ModelElement.taggedValue but Stereotype does have a > > child element Stereotype.requiredTag. The consequence is that one > > cannot simply "attach" TaggedValue to ModelElement as what > > TaggedValue is intended for. I believe the association > > between Stereotype and TaggedValue should be a normal association > > > > ================================= > > Juergen Boldt > > Director, Member Services > > > > Object Management Group > > 250 First Avenue, Suite 100 > > Needham, MA 02494 > > > > Tel. +1 781 444 0404 ext. 132 > > Fax: +1 781 444 0320 > > email: juergen@omg.org > > www www.omg.org > > > > ================================ > > > > > > The information contained in this email and any attached > files are confidential and intended solely for the > addressee(s). The e-mail may be legally privileged or > prohibited from disclosure and unauthorised use. > > If you are not the named addressee you may not use, copy or > disclose this information to any other person. If you > received this message in error please notify the sender immediately. > > Any views or opinions presented here may be solely those of > the originator and do not necessarily reflect those of the Company. > > > > > The information contained in this email and any attached files are confidential and intended solely for the addressee(s). The e-mail may be legally privileged or prohibited from disclosure and unauthorised use. If you are not the named addressee you may not use, copy or disclose this information to any other person. If you received this message in error please notify the sender immediately. Any views or opinions presented here may be solely those of the originator and do not necessarily reflect those of the Company. Subject: RE: issue 5697 -- CWM RTF issue Date: Fri, 25 Oct 2002 17:52:49 +0100 X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: issue 5697 -- CWM RTF issue Thread-Index: AcJ8RBqwY7eSlG3dRd2Sspt0HarCCgAArgHw From: "Pete Rivett" To: "Sridhar Iyengar" , "Tolbert, Doug M" Cc: , X-MIME-Autoconverted: from quoted-printable to 8bit by emerald.omg.org id g9PGqmm17629 We already did an emergency resolution for exactly this issue (numbered 4408 at the time) - we just have not followed through with republishing the 1.0 artifacts (since CWM 1.1 came along very soon after). Pete > -----Original Message----- > From: Sridhar Iyengar [mailto:siyengar@us.ibm.com] > Sent: 25 October 2002 17:31 > To: Tolbert, Doug M > Cc: cwm-rtf@omg.org; John_Poole@hyperion.com; Pete Rivett > Subject: RE: issue 5697 -- CWM RTF issue > > > > I have already sent a question to Andrew. There is an 'emergency fix' > process to fix critical bugs. In general a proposal to fix > a vote can be > taken. > > Sridhar > -------------------------------------------------------------- > ------------------------------------------- > > Sridhar Iyengar > Distinguished Engineer > Application and Integration Middleware > IBM : Research Triangle Park, NC > E-Mail: siyengar@us.ibm.com > Phone: (919) 486-1768 > -------------------------------------------------------------- > ------------------------------------------- > > > |---------+----------------------------> > | | "Tolbert, Doug M"| > | | | | sys.com> | > | | | > | | 10/25/2002 11:58 | > | | AM | > | | | > |---------+----------------------------> > > >------------------------------------------------------------- > -------------------------------------------------------------- > ------------------| > | > > | > | To: Pete Rivett , > John_Poole@hyperion.com, cwm-rtf@omg.org > | > | cc: > > | > | Subject: RE: issue 5697 -- CWM RTF issue > > | > | > > | > | > > | > > >------------------------------------------------------------- > -------------------------------------------------------------- > ------------------| > > > > I'm not sure OMG processes would allow us to formally issue a > corrected 1.0 DTD. I think, rather, that one would be > expected to move to 1.1 instead to get the fixes. > > Pete, with your AB hat on, do you know what the official OMG > position on this would be? If not, maybe we should start > asking Andrew. > > Doug > > -----Original Message----- > From: Pete Rivett [mailto:Pete.Rivett@adaptive.com] > Sent: Friday, October 25, 2002 4:27 AM > To: John_Poole@hyperion.com; cwm-rtf@omg.org > Subject: RE: issue 5697 -- CWM RTF issue > > > With respect to the DTD, I don't think there's a generator > problem: I think it's just that we have not generated/issued > a CWM 1.0 DTD with the fix to issue 4408 applied. In fact I > don't think we've issued ANY updated CWM 1.0 artifacts (mdl, > XMI, IDL) with just that emergency fix applied. > > Is that something we want to do? > Then there's the issue of how we catalog it and distinguish > it from the previous CWM 1.0 DTD. > > Pete > > > -----Original Message----- > > From: John_Poole@hyperion.com [mailto:John_Poole@hyperion.com] > > Sent: 24 October 2002 18:47 > > To: cwm-rtf@omg.org > > Subject: RE: issue 5697 -- CWM RTF issue > > > > > > > > Hi all, > > > > I think I know (approximately) where the problem lies... > > > > By way of comparison, there is an identical situation in > the CWM OLAP > > metamodel, in which an instance of DimensionDeployment may > be owned by > > at most one instance of HierarchyLevelAssociation or by at most one > > instance of ValueBasedHierarchy, but not by both at the > same time, and > > this situation is formally expressed by an OCL constraint > of the form: > > > > context DimensionDeployment inv: > > self.hierarchyLevelAssociation->isEmpty xor > > self.valueBasedHierarchy->isEmpty > > > > (Of course, the constraint is just documentation, is not > embedded in > > the Rose model, and could not have possibly influenced the DTD > > generation in any way). > > > > The main thing that is different between this part of the > OLAP model > > and the Core classes ModelElement and TaggedValue is the > absense of an > > explicit reference on ModelElement corresponding to the > > TaggedElement::taggedValue association end, where as both > > HierarchyLevelAssociation and ValueBasedHierarchy both have > explicit > > references to DimensionDeployment. There is another > difference in that > > both the association ends touching DimenisonDeployment are also > > ordered, whereas this is not the case with the taggedValue / > > requiredTag ends. Other than these two points, I see nothing > > else in the class properties and/or features that might > > otherwise distinguish the two situations. > > > > So, I suspect that the absense of a TaggedValue reference on > > ModelElement is the root cause of the problem. Is it possible that > > the DTD generation process was driven off the references, > and not the > > association ends? Or could this point to a possible error in XMI > > itself? > > > > In any event, the absence of a TaggedValue reference in > ModelElement > > had previously been raised by issue 4408 and certainly is indeed > > problematic for any CWM implementation not using MOF. > > > > Also, we will need to update the CWM specification to > include an XOR > > constraint in the Core model for ownership of TaggedValue. > > > > SO IN SUMMARY: I believe this is either a DTD generator problem or > > possibly an XMI issue. It is not a problem in the CWM > model itself. > > I think Dan's current open issue should be amended to > require that an > > appropriate XOR constraint be added to the spec., and that > the DTD be > > fixed. In the meantime, someone needs to investigate the DTD > > generator that was used (I'd offer to do it myself, but I > do not own a > > copy of the XMI tool used to produce the CWM 1.0 DTD). > > > > Best regards, > > John > > > > > > > > > > > > > > "Pete Rivett" > > > > > > > tive.com> cc: > > > > Subject: RE: issue > > 5697 -- CWM RTF issue > > 10/24/2002 10:05 > > > > AM > > > > > > > > > > > > > > > > > > > > It's fine in MOF for a class to potentially be owned via > more than one > > composition: the restriction is that a single instance can > only have > > one owner. So that means that the compositions must be 0..1 rather > > than 1..1 which they are in this case in CWM. As it stands it means > > that a ModelElement can own tags and a Stereotype can own > tags but the > > same tag cannot be owned by both. > > > > The particular issue of the missing reference from ModelElement to > > TaggedValue I thought was fixed by issue 4408 (which was > dealt with as > > an urgent Issue). > > > > Pete > > > > > -----Original Message----- > > > From: Juergen Boldt [mailto:juergen@omg.org] > > > Sent: 24 October 2002 14:17 > > > To: issues@omg.org; cwm-rtf@omg.org > > > Subject: issue 5697 -- CWM RTF issue > > > > > > > > > This is issue # 5697 Dan Chang [mailto:dtchang@us.ibm.com] > > > > > > TaggedValue > > > > > > We have encountered a problem in our CWM enablement: > > CWM_1.0.dtd would > > > not allow one to attach TaggedValue to ModelElement. I would > > > appreciate your immediate attention. > > > > > > > > > I believe the error is caused by associations defined on > > TaggedValue. > > > In Figure 7-3-3 (CWM 1.0), TaggedValue is owned by both > > ModelElement > > > and Stereotype. I don't think this should be allowed. The > > consequence > > > is that, arbitrarily, in CWM_1.0.dtd, ModelElement does > not have a > > > child element ModelElement.taggedValue but Stereotype does have a > > > child element Stereotype.requiredTag. The consequence is that one > > > cannot simply "attach" TaggedValue to ModelElement as what > > > TaggedValue is intended for. I believe the association between > > > Stereotype and TaggedValue should be a normal association > > > > > > ================================= > > > Juergen Boldt > > > Director, Member Services > > > > > > Object Management Group > > > 250 First Avenue, Suite 100 > > > Needham, MA 02494 > > > > > > Tel. +1 781 444 0404 ext. 132 > > > Fax: +1 781 444 0320 > > > email: juergen@omg.org > > > www www.omg.org > > > > > > ================================ > > > > > > > > > > The information contained in this email and any attached files are > > confidential and intended solely for the addressee(s). The > e-mail may > > be legally privileged or prohibited from disclosure and > unauthorised > > use. > > > > If you are not the named addressee you may not use, copy or > disclose > > this information to any other person. If you received this > message in > > error please notify the sender immediately. > > > > Any views or opinions presented here may be solely those of the > > originator and do not necessarily reflect those of the Company. > > > > > > > > > > > > The information contained in this email and any attached > files are confidential and intended solely for the > addressee(s). The e-mail may be legally privileged or > prohibited from disclosure and unauthorised use. > > If you are not the named addressee you may not use, copy or > disclose this information to any other person. If you > received this message in error please notify the sender immediately. > > Any views or opinions presented here may be solely those of > the originator and do not necessarily reflect those of the Company. > > > > > > The information contained in this email and any attached files are confidential and intended solely for the addressee(s). The e-mail may be legally privileged or prohibited from disclosure and unauthorised use. If you are not the named addressee you may not use, copy or disclose this information to any other person. If you received this message in error please notify the sender immediately. Any views or opinions presented here may be solely those of the originator and do not necessarily reflect those of the Company. From: "Tolbert, Doug M" To: Pete Rivett , Sridhar Iyengar Cc: cwm-rtf@omg.org, John_Poole@hyperion.com Subject: RE: issue 5697 -- CWM RTF issue Date: Fri, 25 Oct 2002 11:59:06 -0500 X-Mailer: Internet Mail Service (5.5.2656.59) I don't think there was ever an understanding on the CWM team the 1.0 COULD be fixed. We thought 1.1 was the only choice. Is this incorrect? Doug -----Original Message----- From: Pete Rivett [mailto:Pete.Rivett@adaptive.com] Sent: Friday, October 25, 2002 9:53 AM To: Sridhar Iyengar; Tolbert, Doug M Cc: cwm-rtf@omg.org; John_Poole@hyperion.com Subject: RE: issue 5697 -- CWM RTF issue We already did an emergency resolution for exactly this issue (numbered 4408 at the time) - we just have not followed through with republishing the 1.0 artifacts (since CWM 1.1 came along very soon after). Pete > -----Original Message----- > From: Sridhar Iyengar [mailto:siyengar@us.ibm.com] > Sent: 25 October 2002 17:31 > To: Tolbert, Doug M > Cc: cwm-rtf@omg.org; John_Poole@hyperion.com; Pete Rivett > Subject: RE: issue 5697 -- CWM RTF issue > > > > I have already sent a question to Andrew. There is an 'emergency fix' > process to fix critical bugs. In general a proposal to fix > a vote can be > taken. > > Sridhar > -------------------------------------------------------------- > ------------------------------------------- > > Sridhar Iyengar > Distinguished Engineer > Application and Integration Middleware > IBM : Research Triangle Park, NC > E-Mail: siyengar@us.ibm.com > Phone: (919) 486-1768 > -------------------------------------------------------------- > ------------------------------------------- > > > |---------+----------------------------> > | | "Tolbert, Doug M"| > | | | | sys.com> | > | | | > | | 10/25/2002 11:58 | > | | AM | > | | | > |---------+----------------------------> > > >------------------------------------------------------------- > -------------------------------------------------------------- > ------------------| > | > > | > | To: Pete Rivett , > John_Poole@hyperion.com, cwm-rtf@omg.org > | > | cc: > > | > | Subject: RE: issue 5697 -- CWM RTF issue > > | > | > > | > | > > | > > >------------------------------------------------------------- > -------------------------------------------------------------- > ------------------| > > > > I'm not sure OMG processes would allow us to formally issue a > corrected 1.0 DTD. I think, rather, that one would be > expected to move to 1.1 instead to get the fixes. > > Pete, with your AB hat on, do you know what the official OMG > position on this would be? If not, maybe we should start > asking Andrew. > > Doug > > -----Original Message----- > From: Pete Rivett [mailto:Pete.Rivett@adaptive.com] > Sent: Friday, October 25, 2002 4:27 AM > To: John_Poole@hyperion.com; cwm-rtf@omg.org > Subject: RE: issue 5697 -- CWM RTF issue > > > With respect to the DTD, I don't think there's a generator > problem: I think it's just that we have not generated/issued > a CWM 1.0 DTD with the fix to issue 4408 applied. In fact I > don't think we've issued ANY updated CWM 1.0 artifacts (mdl, > XMI, IDL) with just that emergency fix applied. > > Is that something we want to do? > Then there's the issue of how we catalog it and distinguish > it from the previous CWM 1.0 DTD. > > Pete > > > -----Original Message----- > > From: John_Poole@hyperion.com [mailto:John_Poole@hyperion.com] > > Sent: 24 October 2002 18:47 > > To: cwm-rtf@omg.org > > Subject: RE: issue 5697 -- CWM RTF issue > > > > > > > > Hi all, > > > > I think I know (approximately) where the problem lies... > > > > By way of comparison, there is an identical situation in > the CWM OLAP > > metamodel, in which an instance of DimensionDeployment may > be owned by > > at most one instance of HierarchyLevelAssociation or by at most one > > instance of ValueBasedHierarchy, but not by both at the > same time, and > > this situation is formally expressed by an OCL constraint > of the form: > > > > context DimensionDeployment inv: > > self.hierarchyLevelAssociation->isEmpty xor > > self.valueBasedHierarchy->isEmpty > > > > (Of course, the constraint is just documentation, is not > embedded in > > the Rose model, and could not have possibly influenced the DTD > > generation in any way). > > > > The main thing that is different between this part of the > OLAP model > > and the Core classes ModelElement and TaggedValue is the > absense of an > > explicit reference on ModelElement corresponding to the > > TaggedElement::taggedValue association end, where as both > > HierarchyLevelAssociation and ValueBasedHierarchy both have > explicit > > references to DimensionDeployment. There is another > difference in that > > both the association ends touching DimenisonDeployment are also > > ordered, whereas this is not the case with the taggedValue / > > requiredTag ends. Other than these two points, I see nothing > > else in the class properties and/or features that might > > otherwise distinguish the two situations. > > > > So, I suspect that the absense of a TaggedValue reference on > > ModelElement is the root cause of the problem. Is it possible that > > the DTD generation process was driven off the references, > and not the > > association ends? Or could this point to a possible error in XMI > > itself? > > > > In any event, the absence of a TaggedValue reference in > ModelElement > > had previously been raised by issue 4408 and certainly is indeed > > problematic for any CWM implementation not using MOF. > > > > Also, we will need to update the CWM specification to > include an XOR > > constraint in the Core model for ownership of TaggedValue. > > > > SO IN SUMMARY: I believe this is either a DTD generator problem or > > possibly an XMI issue. It is not a problem in the CWM > model itself. > > I think Dan's current open issue should be amended to > require that an > > appropriate XOR constraint be added to the spec., and that > the DTD be > > fixed. In the meantime, someone needs to investigate the DTD > > generator that was used (I'd offer to do it myself, but I > do not own a > > copy of the XMI tool used to produce the CWM 1.0 DTD). > > > > Best regards, > > John > > > > > > > > > > > > > > "Pete Rivett" > > > > > > > tive.com> cc: > > > > Subject: RE: issue > > 5697 -- CWM RTF issue > > 10/24/2002 10:05 > > > > AM > > > > > > > > > > > > > > > > > > > > It's fine in MOF for a class to potentially be owned via > more than one > > composition: the restriction is that a single instance can > only have > > one owner. So that means that the compositions must be 0..1 rather > > than 1..1 which they are in this case in CWM. As it stands it means > > that a ModelElement can own tags and a Stereotype can own > tags but the > > same tag cannot be owned by both. > > > > The particular issue of the missing reference from ModelElement to > > TaggedValue I thought was fixed by issue 4408 (which was > dealt with as > > an urgent Issue). > > > > Pete > > > > > -----Original Message----- > > > From: Juergen Boldt [mailto:juergen@omg.org] > > > Sent: 24 October 2002 14:17 > > > To: issues@omg.org; cwm-rtf@omg.org > > > Subject: issue 5697 -- CWM RTF issue > > > > > > > > > This is issue # 5697 Dan Chang [mailto:dtchang@us.ibm.com] > > > > > > TaggedValue > > > > > > We have encountered a problem in our CWM enablement: > > CWM_1.0.dtd would > > > not allow one to attach TaggedValue to ModelElement. I would > > > appreciate your immediate attention. > > > > > > > > > I believe the error is caused by associations defined on > > TaggedValue. > > > In Figure 7-3-3 (CWM 1.0), TaggedValue is owned by both > > ModelElement > > > and Stereotype. I don't think this should be allowed. The > > consequence > > > is that, arbitrarily, in CWM_1.0.dtd, ModelElement does > not have a > > > child element ModelElement.taggedValue but Stereotype does have a > > > child element Stereotype.requiredTag. The consequence is that one > > > cannot simply "attach" TaggedValue to ModelElement as what > > > TaggedValue is intended for. I believe the association between > > > Stereotype and TaggedValue should be a normal association > > > > > > ================================= > > > Juergen Boldt > > > Director, Member Services > > > > > > Object Management Group > > > 250 First Avenue, Suite 100 > > > Needham, MA 02494 > > > > > > Tel. +1 781 444 0404 ext. 132 > > > Fax: +1 781 444 0320 > > > email: juergen@omg.org > > > www www.omg.org > > > > > > ================================ > > > > > > > > > > The information contained in this email and any attached files are > > confidential and intended solely for the addressee(s). The > e-mail may > > be legally privileged or prohibited from disclosure and > unauthorised > > use. > > > > If you are not the named addressee you may not use, copy or > disclose > > this information to any other person. If you received this > message in > > error please notify the sender immediately. > > > > Any views or opinions presented here may be solely those of the > > originator and do not necessarily reflect those of the Company. > > > > > > > > > > > > The information contained in this email and any attached > files are confidential and intended solely for the > addressee(s). The e-mail may be legally privileged or > prohibited from disclosure and unauthorised use. > > If you are not the named addressee you may not use, copy or > disclose this information to any other person. If you > received this message in error please notify the sender immediately. > > Any views or opinions presented here may be solely those of > the originator and do not necessarily reflect those of the Company. > > > > > > The information contained in this email and any attached files are confidential and intended solely for the addressee(s). The e-mail may be legally privileged or prohibited from disclosure and unauthorised use. If you are not the named addressee you may not use, copy or disclose this information to any other person. If you received this message in error please notify the sender immediately. Any views or opinions presented here may be solely those of the originator and do not necessarily reflect those of the Company. Subject: RE: issue 5697 -- CWM RTF issue To: "Tolbert, Doug M" Cc: cwm-rtf@omg.org, John_Poole@hyperion.com, Pete Rivett X-Mailer: Lotus Notes Release 5.0.8 June 18, 2001 From: "Sridhar Iyengar" Date: Fri, 25 Oct 2002 13:20:34 -0400 X-MIMETrack: Serialize by Router on D03NM120/03/M/IBM(Release 5.0.10 |March 22, 2002) at 10/25/2002 11:20:40 AM Doug>We thought 1.1 was the only choice. Is this incorrect? This is incorrrect - If there is a genuine critical bug - The OMG 'urgent issue' process can be used to fix critical bugs - you dont have to wait for the next RTF revision. This has been done for other specs before. Pete>- we just have not followed through with republishing Pete> the 1.0 artifacts (since CWM 1.1 came along very soon after). So will this be followed thru as part of current revision? If not does the RTF want to raise another urgent issue to resolve this? Sridhar --------------------------------------------------------------------------------------------------------- Sridhar Iyengar Distinguished Engineer Application and Integration Middleware IBM : Research Triangle Park, NC E-Mail: siyengar@us.ibm.com Phone: (919) 486-1768 --------------------------------------------------------------------------------------------------------- |---------+----------------------------> | | "Tolbert, Doug M"| | | | | | | | | 10/25/2002 12:59 | | | PM | | | | |---------+----------------------------> >---------------------------------------------------------------------------------------------------------------------------------------------| | | | To: Pete Rivett , Sridhar Iyengar/Raleigh/IBM@IBMUS | | cc: cwm-rtf@omg.org, John_Poole@hyperion.com | | Subject: RE: issue 5697 -- CWM RTF issue | | | | | >---------------------------------------------------------------------------------------------------------------------------------------------| I don't think there was ever an understanding on the CWM team the 1.0 COULD be fixed. We thought 1.1 was the only choice. Is this incorrect? Doug -----Original Message----- From: Pete Rivett [mailto:Pete.Rivett@adaptive.com] Sent: Friday, October 25, 2002 9:53 AM To: Sridhar Iyengar; Tolbert, Doug M Cc: cwm-rtf@omg.org; John_Poole@hyperion.com Subject: RE: issue 5697 -- CWM RTF issue We already did an emergency resolution for exactly this issue (numbered 4408 at the time) - we just have not followed through with republishing the 1.0 artifacts (since CWM 1.1 came along very soon after). Pete > -----Original Message----- > From: Sridhar Iyengar [mailto:siyengar@us.ibm.com] > Sent: 25 October 2002 17:31 > To: Tolbert, Doug M > Cc: cwm-rtf@omg.org; John_Poole@hyperion.com; Pete Rivett > Subject: RE: issue 5697 -- CWM RTF issue > > > > I have already sent a question to Andrew. There is an 'emergency fix' > process to fix critical bugs. In general a proposal to fix > a vote can be > taken. > > Sridhar > -------------------------------------------------------------- > ------------------------------------------- > > Sridhar Iyengar > Distinguished Engineer > Application and Integration Middleware > IBM : Research Triangle Park, NC > E-Mail: siyengar@us.ibm.com > Phone: (919) 486-1768 > -------------------------------------------------------------- > ------------------------------------------- > > > |---------+----------------------------> > | | "Tolbert, Doug M"| > | | | | sys.com> | > | | | > | | 10/25/2002 11:58 | > | | AM | > | | | > |---------+----------------------------> > > >------------------------------------------------------------- > -------------------------------------------------------------- > ------------------| > | > > | > | To: Pete Rivett , > John_Poole@hyperion.com, cwm-rtf@omg.org > | > | cc: > > | > | Subject: RE: issue 5697 -- CWM RTF issue > > | > | > > | > | > > | > > >------------------------------------------------------------- > -------------------------------------------------------------- > ------------------| > > > > I'm not sure OMG processes would allow us to formally issue a > corrected 1.0 DTD. I think, rather, that one would be > expected to move to 1.1 instead to get the fixes. > > Pete, with your AB hat on, do you know what the official OMG > position on this would be? If not, maybe we should start > asking Andrew. > > Doug > > -----Original Message----- > From: Pete Rivett [mailto:Pete.Rivett@adaptive.com] > Sent: Friday, October 25, 2002 4:27 AM > To: John_Poole@hyperion.com; cwm-rtf@omg.org > Subject: RE: issue 5697 -- CWM RTF issue > > > With respect to the DTD, I don't think there's a generator > problem: I think it's just that we have not generated/issued > a CWM 1.0 DTD with the fix to issue 4408 applied. In fact I > don't think we've issued ANY updated CWM 1.0 artifacts (mdl, > XMI, IDL) with just that emergency fix applied. > > Is that something we want to do? > Then there's the issue of how we catalog it and distinguish > it from the previous CWM 1.0 DTD. > > Pete > > > -----Original Message----- > > From: John_Poole@hyperion.com [mailto:John_Poole@hyperion.com] > > Sent: 24 October 2002 18:47 > > To: cwm-rtf@omg.org > > Subject: RE: issue 5697 -- CWM RTF issue > > > > > > > > Hi all, > > > > I think I know (approximately) where the problem lies... > > > > By way of comparison, there is an identical situation in > the CWM OLAP > > metamodel, in which an instance of DimensionDeployment may > be owned by > > at most one instance of HierarchyLevelAssociation or by at most one > > instance of ValueBasedHierarchy, but not by both at the > same time, and > > this situation is formally expressed by an OCL constraint > of the form: > > > > context DimensionDeployment inv: > > self.hierarchyLevelAssociation->isEmpty xor > > self.valueBasedHierarchy->isEmpty > > > > (Of course, the constraint is just documentation, is not > embedded in > > the Rose model, and could not have possibly influenced the DTD > > generation in any way). > > > > The main thing that is different between this part of the > OLAP model > > and the Core classes ModelElement and TaggedValue is the > absense of an > > explicit reference on ModelElement corresponding to the > > TaggedElement::taggedValue association end, where as both > > HierarchyLevelAssociation and ValueBasedHierarchy both have > explicit > > references to DimensionDeployment. There is another > difference in that > > both the association ends touching DimenisonDeployment are also > > ordered, whereas this is not the case with the taggedValue / > > requiredTag ends. Other than these two points, I see nothing > > else in the class properties and/or features that might > > otherwise distinguish the two situations. > > > > So, I suspect that the absense of a TaggedValue reference on > > ModelElement is the root cause of the problem. Is it possible that > > the DTD generation process was driven off the references, > and not the > > association ends? Or could this point to a possible error in XMI > > itself? > > > > In any event, the absence of a TaggedValue reference in > ModelElement > > had previously been raised by issue 4408 and certainly is indeed > > problematic for any CWM implementation not using MOF. > > > > Also, we will need to update the CWM specification to > include an XOR > > constraint in the Core model for ownership of TaggedValue. > > > > SO IN SUMMARY: I believe this is either a DTD generator problem or > > possibly an XMI issue. It is not a problem in the CWM > model itself. > > I think Dan's current open issue should be amended to > require that an > > appropriate XOR constraint be added to the spec., and that > the DTD be > > fixed. In the meantime, someone needs to investigate the DTD > > generator that was used (I'd offer to do it myself, but I > do not own a > > copy of the XMI tool used to produce the CWM 1.0 DTD). > > > > Best regards, > > John > > > > > > > > > > > > > > "Pete Rivett" > > > > > > > tive.com> cc: > > > > Subject: RE: issue > > 5697 -- CWM RTF issue > > 10/24/2002 10:05 > > > > AM > > > > > > > > > > > > > > > > > > > > It's fine in MOF for a class to potentially be owned via > more than one > > composition: the restriction is that a single instance can > only have > > one owner. So that means that the compositions must be 0..1 rather > > than 1..1 which they are in this case in CWM. As it stands it means > > that a ModelElement can own tags and a Stereotype can own > tags but the > > same tag cannot be owned by both. > > > > The particular issue of the missing reference from ModelElement to > > TaggedValue I thought was fixed by issue 4408 (which was > dealt with as > > an urgent Issue). > > > > Pete > > > > > -----Original Message----- > > > From: Juergen Boldt [mailto:juergen@omg.org] > > > Sent: 24 October 2002 14:17 > > > To: issues@omg.org; cwm-rtf@omg.org > > > Subject: issue 5697 -- CWM RTF issue > > > > > > > > > This is issue # 5697 Dan Chang [mailto:dtchang@us.ibm.com] > > > > > > TaggedValue > > > > > > We have encountered a problem in our CWM enablement: > > CWM_1.0.dtd would > > > not allow one to attach TaggedValue to ModelElement. I would > > > appreciate your immediate attention. > > > > > > > > > I believe the error is caused by associations defined on > > TaggedValue. > > > In Figure 7-3-3 (CWM 1.0), TaggedValue is owned by both > > ModelElement > > > and Stereotype. I don't think this should be allowed. The > > consequence > > > is that, arbitrarily, in CWM_1.0.dtd, ModelElement does > not have a > > > child element ModelElement.taggedValue but Stereotype does have a > > > child element Stereotype.requiredTag. The consequence is that one > > > cannot simply "attach" TaggedValue to ModelElement as what > > > TaggedValue is intended for. I believe the association between > > > Stereotype and TaggedValue should be a normal association > > > > > > ================================= > > > Juergen Boldt > > > Director, Member Services > > > > > > Object Management Group > > > 250 First Avenue, Suite 100 > > > Needham, MA 02494 > > > > > > Tel. +1 781 444 0404 ext. 132 > > > Fax: +1 781 444 0320 > > > email: juergen@omg.org > > > www www.omg.org > > > > > > ================================ > > > > > > > > > > The information contained in this email and any attached files are > > confidential and intended solely for the addressee(s). The > e-mail may > > be legally privileged or prohibited from disclosure and > unauthorised > > use. > > > > If you are not the named addressee you may not use, copy or > disclose > > this information to any other person. If you received this > message in > > error please notify the sender immediately. > > > > Any views or opinions presented here may be solely those of the > > originator and do not necessarily reflect those of the Company. > > > > > > > > > > > > The information contained in this email and any attached > files are confidential and intended solely for the > addressee(s). The e-mail may be legally privileged or > prohibited from disclosure and unauthorised use. > > If you are not the named addressee you may not use, copy or > disclose this information to any other person. If you > received this message in error please notify the sender immediately. > > Any views or opinions presented here may be solely those of > the originator and do not necessarily reflect those of the Company. > > > > > > The information contained in this email and any attached files are confidential and intended solely for the addressee(s). The e-mail may be legally privileged or prohibited from disclosure and unauthorised use. If you are not the named addressee you may not use, copy or disclose this information to any other person. If you received this message in error please notify the sender immediately. Any views or opinions presented here may be solely those of the originator and do not necessarily reflect those of the Company. Subject: RE: issue 5697 -- CWM RTF issue To: "Pete Rivett" Cc: cwm-rtf@omg.org X-Mailer: Lotus Notes Release 5.0.9a January 7, 2002 From: John_Poole@hyperion.com Date: Fri, 25 Oct 2002 13:36:20 -0400 X-MIMETrack: Serialize by Router on Stmfd-Gateway1/na/Hyperion(Release 5.0.9a |January 7, 2002) at 10/25/2002 01:36:22 PM But my point is that the DTD should not have been wrong in the first place. The CWM ModelElement represented in the DTD should still have contained an XML Element representing owned TaggedValues, by virtue of the composition relationship from ModelElement to TaggedValue in the CWM model. It seems like the DTD only got corrected after a "taggedValue" reference was added to ModelElement; hence, my conclusion that perhaps the DTD generation process is being driven by the presence (or absence) of references, rather than the associations upon which the references are based. Do you agree? Regards, John "Pete Rivett" tive.com> cc: Subject: RE: issue 5697 -- CWM RTF issue 10/25/02 07:26 AM With respect to the DTD, I don't think there's a generator problem: I think it's just that we have not generated/issued a CWM 1.0 DTD with the fix to issue 4408 applied. In fact I don't think we've issued ANY updated CWM 1.0 artifacts (mdl, XMI, IDL) with just that emergency fix applied. Is that something we want to do? Then there's the issue of how we catalog it and distinguish it from the previous CWM 1.0 DTD. Pete > -----Original Message----- > From: John_Poole@hyperion.com [mailto:John_Poole@hyperion.com] > Sent: 24 October 2002 18:47 > To: cwm-rtf@omg.org > Subject: RE: issue 5697 -- CWM RTF issue > Subject: RE: issue 5697 -- CWM RTF issue To: John_Poole@hyperion.com Cc: cwm-rtf@omg.org, "Pete Rivett" X-Mailer: Lotus Notes Release 5.0.7 March 21, 2001 From: "Dan Chang" Date: Fri, 25 Oct 2002 11:15:16 -0700 X-MIMETrack: Serialize by Router on D03NM800/03/M/IBM(Release 5.0.10 |March 22, 2002) at 10/25/2002 12:15:17 PM John, I agree with you. The DTD should not have been wrong in the first place. Though fixing #4408 will incidentally fix #5697. If one takes a look at UML 1.3, ModelElement doesn't have a reference to TaggedVaule, but its DTD is correct. Regards, Dan IBM Almaden Research Center/Silicon Valley Lab Notes: Dan Chang/Santa Teresa/IBM@IBMUS Internet: dtchang@us.ibm.com Phone: (408)-927-1714/(408)-463-2319 Office: ARC(G1-018)/SVL(D168) John_Poole@hyperi on.com To: "Pete Rivett" cc: cwm-rtf@omg.org 10/25/2002 10:36 Subject: RE: issue 5697 -- CWM RTF issue AM But my point is that the DTD should not have been wrong in the first place. The CWM ModelElement represented in the DTD should still have contained an XML Element representing owned TaggedValues, by virtue of the composition relationship from ModelElement to TaggedValue in the CWM model. It seems like the DTD only got corrected after a "taggedValue" reference was added to ModelElement; hence, my conclusion that perhaps the DTD generation process is being driven by the presence (or absence) of references, rather than the associations upon which the references are based. Subject: RE: issue 5697 -- CWM RTF issue Date: Fri, 25 Oct 2002 12:26:50 +0100 X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: issue 5697 -- CWM RTF issue Thread-Index: AcJ7hewThWbGSq1pRS6BSqEjA6wijQAkgCvQ From: "Pete Rivett" To: , X-MIME-Autoconverted: from quoted-printable to 8bit by emerald.omg.org id g9PBQnm08147 With respect to the DTD, I don't think there's a generator problem: I think it's just that we have not generated/issued a CWM 1.0 DTD with the fix to issue 4408 applied. In fact I don't think we've issued ANY updated CWM 1.0 artifacts (mdl, XMI, IDL) with just that emergency fix applied. Is that something we want to do? Then there's the issue of how we catalog it and distinguish it from the previous CWM 1.0 DTD. Pete > -----Original Message----- > From: John_Poole@hyperion.com [mailto:John_Poole@hyperion.com] > Sent: 24 October 2002 18:47 > To: cwm-rtf@omg.org > Subject: RE: issue 5697 -- CWM RTF issue > > > > Hi all, > > I think I know (approximately) where the problem lies... > > By way of comparison, there is an identical situation in the > CWM OLAP metamodel, in which an instance of > DimensionDeployment may be owned by at most one instance of > HierarchyLevelAssociation or by at most one instance of > ValueBasedHierarchy, but not by both at the same time, and > this situation is formally expressed by an OCL constraint of the form: > > context DimensionDeployment inv: > self.hierarchyLevelAssociation->isEmpty xor > self.valueBasedHierarchy->isEmpty > > (Of course, the constraint is just documentation, is not > embedded in the Rose model, and could not have possibly > influenced the DTD generation in any way). > > The main thing that is different between this part of the > OLAP model and the Core classes ModelElement and TaggedValue > is the absense of an explicit reference on ModelElement > corresponding to the TaggedElement::taggedValue association > end, where as both HierarchyLevelAssociation and > ValueBasedHierarchy both have explicit references to > DimensionDeployment. There is another difference in that both > the association ends touching DimenisonDeployment are also > ordered, whereas this is not the case with the taggedValue / > requiredTag ends. Other than these two points, I see nothing > else in the class properties and/or features that might > otherwise distinguish the two situations. > > So, I suspect that the absense of a TaggedValue reference on > ModelElement is the root cause of the problem. Is it > possible that the DTD generation process was driven off the > references, and not the association ends? Or could this > point to a possible error in XMI itself? > > In any event, the absence of a TaggedValue reference in > ModelElement had previously been raised by issue 4408 and > certainly is indeed problematic for any CWM implementation > not using MOF. > > Also, we will need to update the CWM specification to include > an XOR constraint in the Core model for ownership of TaggedValue. > > SO IN SUMMARY: I believe this is either a DTD generator > problem or possibly an XMI issue. It is not a problem in the > CWM model itself. I think Dan's current open issue should be > amended to require that an appropriate XOR constraint be > added to the spec., and that the DTD be fixed. In the > meantime, someone needs to investigate the DTD generator that > was used (I'd offer to do it myself, but I do not own a copy > of the XMI tool used to produce the CWM 1.0 DTD). > > Best regards, > John > > > > > > > "Pete Rivett" > > > tive.com> cc: > > Subject: RE: > issue 5697 -- CWM RTF issue > 10/24/2002 10:05 > > AM > > > > > > > > > > It's fine in MOF for a class to potentially be owned via more than one > composition: the restriction is that a single instance can > only have one owner. So that means that the compositions must > be 0..1 rather than 1..1 which they are in this case in CWM. > As it stands it means that a ModelElement can own tags and a > Stereotype can own tags but the same tag cannot be owned by both. > > The particular issue of the missing reference from > ModelElement to TaggedValue I thought was fixed by issue 4408 > (which was dealt with as an urgent Issue). > > Pete > > > -----Original Message----- > > From: Juergen Boldt [mailto:juergen@omg.org] > > Sent: 24 October 2002 14:17 > > To: issues@omg.org; cwm-rtf@omg.org > > Subject: issue 5697 -- CWM RTF issue > > > > > > This is issue # 5697 Dan Chang [mailto:dtchang@us.ibm.com] > > > > TaggedValue > > > > We have encountered a problem in our CWM enablement: > CWM_1.0.dtd would > > not allow one to attach TaggedValue to ModelElement. I would > > appreciate your immediate attention. > > > > > > I believe the error is caused by associations defined on > TaggedValue. > > In Figure 7-3-3 (CWM 1.0), TaggedValue is owned by both > ModelElement > > and Stereotype. I don't think this should be allowed. The > consequence > > is that, arbitrarily, in CWM_1.0.dtd, ModelElement does not have a > > child element ModelElement.taggedValue but Stereotype does have a > > child element Stereotype.requiredTag. The consequence is that one > > cannot simply "attach" TaggedValue to ModelElement as what > > TaggedValue is intended for. I believe the association > > between Stereotype and TaggedValue should be a normal association > > > > ================================= > > Juergen Boldt > > Director, Member Services > > > > Object Management Group > > 250 First Avenue, Suite 100 > > Needham, MA 02494 > > > > Tel. +1 781 444 0404 ext. 132 > > Fax: +1 781 444 0320 > > email: juergen@omg.org > > www www.omg.org > > > > ================================ > > > > > > The information contained in this email and any attached > files are confidential and intended solely for the > addressee(s). The e-mail may be legally privileged or > prohibited from disclosure and unauthorised use. > > If you are not the named addressee you may not use, copy or > disclose this information to any other person. If you > received this message in error please notify the sender immediately. > > Any views or opinions presented here may be solely those of > the originator and do not necessarily reflect those of the Company. > > > > > The information contained in this email and any attached files are confidential and intended solely for the addressee(s). The e-mail may be legally privileged or prohibited from disclosure and unauthorised use. If you are not the named addressee you may not use, copy or disclose this information to any other person. If you received this message in error please notify the sender immediately. Any views or opinions presented here may be solely those of the originator and do not necessarily reflect those of the Company. Subject: conference call To: "CWM-RTF" X-Mailer: Lotus Notes Release 5.0.7 March 21, 2001 From: "Dan Chang" Date: Fri, 25 Oct 2002 11:42:02 -0700 X-MIMETrack: Serialize by Router on D03NM800/03/M/IBM(Release 5.0.10 |March 22, 2002) at 10/25/2002 12:42:02 PM I have one priority issue: #5697, which has to do with correcting an error in CWM_1.0.dtd. In addition, I believe that the resolution to #4408, which was raised as an urgent issue, should have been applied to the CWM 1.0 specification and associated files. Regards, Dan IBM Almaden Research Center/Silicon Valley Lab Notes: Dan Chang/Santa Teresa/IBM@IBMUS Internet: dtchang@us.ibm.com Phone: (408)-927-1714/(408)-463-2319 Office: ARC(G1-018)/SVL(D168) Subject: RE: issue 5697 -- CWM RTF issue - the UML DTD is wrong Date: Sat, 26 Oct 2002 09:19:10 +0100 X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: issue 5697 -- CWM RTF issue - the UML DTD is wrong Thread-Index: AcJ8UqvGTPahhJd4T82bQEv19atWqQAcuGPQ From: "Pete Rivett" To: "Dan Chang" Cc: X-MIME-Autoconverted: from quoted-printable to 8bit by emerald.omg.org id g9Q8J4m02667 Hi Dan, Actually the part of the UML DTD you point out is *incorrect* (due to its inconsistency with the metamodel) and this is causing a fair amount of hassle for interoperability, since some people have implementations based on the DTD and others based on the metamodel and they can't interoperate. See UML Issue 5525. The CWM situation is different since the metamodel *does* have the reference (with the resolution to 4408). Pete > -----Original Message----- > From: Dan Chang [mailto:dtchang@us.ibm.com] > Sent: 25 October 2002 19:15 > To: John_Poole@hyperion.com > Cc: cwm-rtf@omg.org; Pete Rivett > Subject: RE: issue 5697 -- CWM RTF issue > > > > John, > > I agree with you. The DTD should not have been wrong in the > first place. Though fixing #4408 will incidentally fix #5697. > > If one takes a look at UML 1.3, ModelElement doesn't have a > reference to TaggedVaule, but its DTD is correct. > > Regards, Dan > > IBM Almaden Research Center/Silicon Valley Lab > Notes: Dan Chang/Santa Teresa/IBM@IBMUS > Internet: dtchang@us.ibm.com > Phone: (408)-927-1714/(408)-463-2319 > Office: ARC(G1-018)/SVL(D168) > > > > > > John_Poole@hyperi > > > on.com To: > "Pete Rivett" > > cc: > cwm-rtf@omg.org > > 10/25/2002 10:36 Subject: RE: > issue 5697 -- CWM RTF issue > > AM > > > > > > > > > > > > > But my point is that the DTD should not have been wrong in > the first place. The CWM ModelElement represented in the DTD > should still have contained an XML Element representing owned > TaggedValues, by virtue of the composition relationship from > ModelElement to TaggedValue in the CWM model. > > It seems like the DTD only got corrected after a > "taggedValue" reference was added to ModelElement; hence, my > conclusion that perhaps the DTD generation process is being > driven by the presence (or absence) of references, rather > than the associations upon which the references are based. > > Do you agree? > > Regards, > John > > > > > > "Pete Rivett" > > Poole/na/Hyperion@Hyperion, > tive.com> cc: > > Subject: RE: > issue 5697 -- CWM RTF issue > 10/25/02 07:26 AM > > > > > > > > With respect to the DTD, I don't think there's a generator > problem: I think it's just that we have not generated/issued > a CWM 1.0 DTD with the fix to issue 4408 applied. In fact I > don't think we've issued ANY updated CWM 1.0 artifacts (mdl, > XMI, IDL) with just that emergency fix applied. > > Is that something we want to do? > Then there's the issue of how we catalog it and distinguish > it from the previous CWM 1.0 DTD. > > Pete > > > -----Original Message----- > > From: John_Poole@hyperion.com [mailto:John_Poole@hyperion.com] > > Sent: 24 October 2002 18:47 > > To: cwm-rtf@omg.org > > Subject: RE: issue 5697 -- CWM RTF issue > > > > > > > > Hi all, > > > > I think I know (approximately) where the problem lies... > > > > By way of comparison, there is an identical situation in > the CWM OLAP > > metamodel, in which an instance of DimensionDeployment may > be owned by > > at most one instance of HierarchyLevelAssociation or by at most one > > instance of ValueBasedHierarchy, but not by both at the > same time, and > > this situation is formally expressed by an OCL constraint > of the form: > > > > context DimensionDeployment inv: > > self.hierarchyLevelAssociation->isEmpty xor > > self.valueBasedHierarchy->isEmpty > > > > (Of course, the constraint is just documentation, is not > embedded in > > the Rose model, and could not have possibly influenced the DTD > > generation in any way). > > > > The main thing that is different between this part of the > OLAP model > > and the Core classes ModelElement and TaggedValue is the > absense of an > > explicit reference on ModelElement corresponding to the > > TaggedElement::taggedValue association end, where as both > > HierarchyLevelAssociation and ValueBasedHierarchy both have > explicit > > references to DimensionDeployment. There is another > difference in that > > both the association ends touching DimenisonDeployment are also > > ordered, whereas this is not the case with the taggedValue / > > requiredTag ends. Other than these two points, I see nothing > > else in the class properties and/or features that might > > otherwise distinguish the two situations. > > > > So, I suspect that the absense of a TaggedValue reference on > > ModelElement is the root cause of the problem. Is it possible that > > the DTD generation process was driven off the references, > and not the > > association ends? Or could this point to a possible error in XMI > > itself? > > > > In any event, the absence of a TaggedValue reference in > ModelElement > > had previously been raised by issue 4408 and certainly is indeed > > problematic for any CWM implementation not using MOF. > > > > Also, we will need to update the CWM specification to > include an XOR > > constraint in the Core model for ownership of TaggedValue. > > > > SO IN SUMMARY: I believe this is either a DTD generator problem or > > possibly an XMI issue. It is not a problem in the CWM > model itself. > > I think Dan's current open issue should be amended to > require that an > > appropriate XOR constraint be added to the spec., and that > the DTD be > > fixed. In the meantime, someone needs to investigate the DTD > > generator that was used (I'd offer to do it myself, but I > do not own a > > copy of the XMI tool used to produce the CWM 1.0 DTD). > > > > Best regards, > > John > > > > > > > > > > > > > > "Pete Rivett" > > > > > > > tive.com> cc: > > > > Subject: RE: issue > > 5697 -- CWM RTF issue > > 10/24/2002 10:05 > > > > AM > > > > > > > > > > > > > > > > > > > > It's fine in MOF for a class to potentially be owned via > more than one > > composition: the restriction is that a single instance can > only have > > one owner. So that means that the compositions must be 0..1 rather > > than 1..1 which they are in this case in CWM. As it stands it means > > that a ModelElement can own tags and a Stereotype can own > tags but the > > same tag cannot be owned by both. > > > > The particular issue of the missing reference from ModelElement to > > TaggedValue I thought was fixed by issue 4408 (which was > dealt with as > > an urgent Issue). > > > > Pete > > > > > -----Original Message----- > > > From: Juergen Boldt [mailto:juergen@omg.org] > > > Sent: 24 October 2002 14:17 > > > To: issues@omg.org; cwm-rtf@omg.org > > > Subject: issue 5697 -- CWM RTF issue > > > > > > > > > This is issue # 5697 Dan Chang [mailto:dtchang@us.ibm.com] > > > > > > TaggedValue > > > > > > We have encountered a problem in our CWM enablement: > > CWM_1.0.dtd would > > > not allow one to attach TaggedValue to ModelElement. I would > > > appreciate your immediate attention. > > > > > > > > > I believe the error is caused by associations defined on > > TaggedValue. > > > In Figure 7-3-3 (CWM 1.0), TaggedValue is owned by both > > ModelElement > > > and Stereotype. I don't think this should be allowed. The > > consequence > > > is that, arbitrarily, in CWM_1.0.dtd, ModelElement does > not have a > > > child element ModelElement.taggedValue but Stereotype does have a > > > child element Stereotype.requiredTag. The consequence is that one > > > cannot simply "attach" TaggedValue to ModelElement as what > > > TaggedValue is intended for. I believe the association between > > > Stereotype and TaggedValue should be a normal association > > > > > > ================================= > > > Juergen Boldt > > > Director, Member Services > > > > > > Object Management Group > > > 250 First Avenue, Suite 100 > > > Needham, MA 02494 > > > > > > Tel. +1 781 444 0404 ext. 132 > > > Fax: +1 781 444 0320 > > > email: juergen@omg.org > > > www www.omg.org > > > > > > ================================ > > > > > > > > > > The information contained in this email and any attached files are > > confidential and intended solely for the addressee(s). The > e-mail may > > be legally privileged or prohibited from disclosure and > unauthorised > > use. > > > > If you are not the named addressee you may not use, copy or > disclose > > this information to any other person. If you received this > message in > > error please notify the sender immediately. > > > > Any views or opinions presented here may be solely those of the > > originator and do not necessarily reflect those of the Company. > > > > > > > > > > > > The information contained in this email and any attached > files are confidential and intended solely for the > addressee(s). The e-mail may be legally privileged or > prohibited from disclosure and unauthorised use. > > If you are not the named addressee you may not use, copy or > disclose this information to any other person. If you > received this message in error please notify the sender immediately. > > Any views or opinions presented here may be solely those of > the originator and do not necessarily reflect those of the Company. > > > > > > > > > > The information contained in this email and any attached files are confidential and intended solely for the addressee(s). The e-mail may be legally privileged or prohibited from disclosure and unauthorised use. If you are not the named addressee you may not use, copy or disclose this information to any other person. If you received this message in error please notify the sender immediately. Any views or opinions presented here may be solely those of the originator and do not necessarily reflect those of the Company. Subject: RE: issue 5697 -- CWM RTF issue - the UML DTD is wrong To: "Pete Rivett" Cc: cwm-rtf@omg.org X-Mailer: Lotus Notes Release 5.0.7 March 21, 2001 From: "Dan Chang" Date: Mon, 28 Oct 2002 09:15:58 -0800 X-MIMETrack: Serialize by Router on D03NM800/03/M/IBM(Release 5.0.10 |March 22, 2002) at 10/28/2002 10:16:06 AM Does the UML team agree with your branding of UML 1.3 DTD as "incorrect"? UML 1.3, unlike MOF, does not have explicit references. Also, are you saying every composition must have an explicit reference defined on the owner ModelElement for XMI to work? Is this specified in MOF or XMI? Regards, Dan IBM Almaden Research Center/Silicon Valley Lab Notes: Dan Chang/Santa Teresa/IBM@IBMUS Internet: dtchang@us.ibm.com Phone: (408)-927-1714/(408)-463-2319 Office: ARC(G1-018)/SVL(D168) "Pete Rivett" cc: Subject: RE: issue 5697 -- CWM RTF issue - the UML DTD is wrong 10/26/2002 01:19 AM Hi Dan, Actually the part of the UML DTD you point out is *incorrect* (due to its inconsistency with the metamodel) and this is causing a fair amount of hassle for interoperability, since some people have implementations based on the DTD and others based on the metamodel and they can't interoperate. See UML Issue 5525. The CWM situation is different since the metamodel *does* have the reference (with the resolution to 4408). Subject: RE: issue 5697 -- CWM RTF issue - the UML DTD is wrong Date: Mon, 28 Oct 2002 17:52:07 -0000 X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: issue 5697 -- CWM RTF issue - the UML DTD is wrong Thread-Index: AcJ+pbmNnEfwNiUNRQOowj97ioNYcQAA6HPA From: "Pete Rivett" To: "Dan Chang" Cc: X-MIME-Autoconverted: from quoted-printable to 8bit by emerald.omg.org id g9SHq4m19487 Hi Dan, UML has a separate 'concrete' or 'interchange' metamodel which does have explicit references. The UML issue I raised discussed the pros and cons of 2 possible resolutions: - make the metamodel match the DTD - make the DTD match the metamodel I'm not a member of the UML RTF so don't know if it's been discussed on their telcons: I am on the mailing list and have seen little if no discussion there. The issue was however discussed at the UML XMI Interoperability Workshop at OMG TC in Helsinki, and it was agreed by the vendors present that this was a genuine problem (in fact the 'public enemy number one' standing in the way of UML interoperability). No every composition does not need an explicit reference from the owner: however if not then the owned elements must be held as 'top level' elements in the XMI file which have a reference back to the owner. This is what several metamodel-driven tools implement (e.g. Unisys Rose integration, Interactive Objects, Sun MDR). BTW I have knocked up a XSLT stylesheet to convert XMI files with nested Tags into the correct 'top level' representation. Pete > -----Original Message----- > From: Dan Chang [mailto:dtchang@us.ibm.com] > Sent: 28 October 2002 17:16 > To: Pete Rivett > Cc: cwm-rtf@omg.org > Subject: RE: issue 5697 -- CWM RTF issue - the UML DTD is wrong > > > > Does the UML team agree with your branding of UML 1.3 DTD as > "incorrect"? UML 1.3, unlike MOF, does not have explicit references. > > Also, are you saying every composition must have an explicit > reference defined on the owner ModelElement for XMI to work? > Is this specified in MOF or XMI? > > Regards, Dan > > IBM Almaden Research Center/Silicon Valley Lab > Notes: Dan Chang/Santa Teresa/IBM@IBMUS > Internet: dtchang@us.ibm.com > Phone: (408)-927-1714/(408)-463-2319 > Office: ARC(G1-018)/SVL(D168) > > > > > > "Pete Rivett" > > > Chang/Santa Teresa/IBM@IBMUS > > tive.com> cc: > > > Subject: RE: > issue 5697 -- CWM RTF issue - the UML DTD is wrong > > 10/26/2002 01:19 > > > AM > > > > > > > > > > > > Hi Dan, > Actually the part of the UML DTD you point out is *incorrect* > (due to its inconsistency with the metamodel) and this is > causing a fair amount of hassle for interoperability, since > some people have implementations based on the DTD and others > based on the metamodel and they can't interoperate. See UML > Issue 5525. > > The CWM situation is different since the metamodel *does* > have the reference (with the resolution to 4408). > > Pete > > > -----Original Message----- > > From: Dan Chang [mailto:dtchang@us.ibm.com] > > Sent: 25 October 2002 19:15 > > To: John_Poole@hyperion.com > > Cc: cwm-rtf@omg.org; Pete Rivett > > Subject: RE: issue 5697 -- CWM RTF issue > > > > > > > > John, > > > > I agree with you. The DTD should not have been wrong in the first > > place. Though fixing #4408 will incidentally fix #5697. > > > > If one takes a look at UML 1.3, ModelElement doesn't have a > reference > > to TaggedVaule, but its DTD is correct. > > > > Regards, Dan > > > > IBM Almaden Research Center/Silicon Valley Lab > > Notes: Dan Chang/Santa Teresa/IBM@IBMUS > > Internet: dtchang@us.ibm.com > > Phone: (408)-927-1714/(408)-463-2319 > > Office: ARC(G1-018)/SVL(D168) > > > > > > > > > > > > John_Poole@hyperi > > > > > > on.com To: > > "Pete Rivett" > > > > cc: cwm-rtf@omg.org > > > > 10/25/2002 10:36 Subject: RE: > > issue 5697 -- CWM RTF issue > > > > AM > > > > > > > > > > > > > > > > > > > > > > > > > > But my point is that the DTD should not have been wrong in > the first > > place. The CWM ModelElement represented in the DTD should > still have > > contained an XML Element representing owned TaggedValues, > by virtue of > > the composition relationship from ModelElement to > TaggedValue in the > > CWM model. > > > > It seems like the DTD only got corrected after a "taggedValue" > > reference was added to ModelElement; hence, my conclusion > that perhaps > > the DTD generation process is being driven by the presence (or > > absence) of references, rather than the associations upon which the > > references are based. > > > > Do you agree? > > > > Regards, > > John > > > > > > > > > > > > "Pete Rivett" > > > > > Poole/na/Hyperion@Hyperion, > > tive.com> cc: > > > > Subject: RE: issue > > 5697 -- CWM RTF issue > > 10/25/02 07:26 AM > > > > > > > > > > > > > > > > With respect to the DTD, I don't think there's a generator > > problem: I think it's just that we have not > generated/issued a CWM 1.0 > > DTD with the fix to issue 4408 applied. In fact I don't think we've > > issued ANY updated CWM 1.0 artifacts (mdl, XMI, IDL) with just that > > emergency fix applied. > > > > Is that something we want to do? > > Then there's the issue of how we catalog it and distinguish it from > > the previous CWM 1.0 DTD. > > > > Pete > > > > > -----Original Message----- > > > From: John_Poole@hyperion.com [mailto:John_Poole@hyperion.com] > > > Sent: 24 October 2002 18:47 > > > To: cwm-rtf@omg.org > > > Subject: RE: issue 5697 -- CWM RTF issue > > > > > > > > > > > > Hi all, > > > > > > I think I know (approximately) where the problem lies... > > > > > > By way of comparison, there is an identical situation in > > the CWM OLAP > > > metamodel, in which an instance of DimensionDeployment may > > be owned by > > > at most one instance of HierarchyLevelAssociation or by > at most one > > > instance of ValueBasedHierarchy, but not by both at the > > same time, and > > > this situation is formally expressed by an OCL constraint > > of the form: > > > > > > context DimensionDeployment inv: > > > self.hierarchyLevelAssociation->isEmpty xor > > > self.valueBasedHierarchy->isEmpty > > > > > > (Of course, the constraint is just documentation, is not > > embedded in > > > the Rose model, and could not have possibly influenced the DTD > > > generation in any way). > > > > > > The main thing that is different between this part of the > > OLAP model > > > and the Core classes ModelElement and TaggedValue is the > > absense of an > > > explicit reference on ModelElement corresponding to the > > > TaggedElement::taggedValue association end, where as both > > > HierarchyLevelAssociation and ValueBasedHierarchy both have > > explicit > > > references to DimensionDeployment. There is another > > difference in that > > > both the association ends touching DimenisonDeployment are also > > > ordered, whereas this is not the case with the taggedValue / > > > requiredTag ends. Other than these two points, I see > nothing else in > > > the class properties and/or features that might otherwise > > > distinguish the two situations. > > > > > > So, I suspect that the absense of a TaggedValue reference on > > > ModelElement is the root cause of the problem. Is it > possible that > > > the DTD generation process was driven off the references, > > and not the > > > association ends? Or could this point to a possible error in XMI > > > itself? > > > > > > In any event, the absence of a TaggedValue reference in > > ModelElement > > > had previously been raised by issue 4408 and certainly is indeed > > > problematic for any CWM implementation not using MOF. > > > > > > Also, we will need to update the CWM specification to > > include an XOR> > > constraint in the Core model for ownership of TaggedValue. > > > > > > SO IN SUMMARY: I believe this is either a DTD generator > problem or > > > possibly an XMI issue. It is not a problem in the CWM > > model itself. > > > I think Dan's current open issue should be amended to > > require that an > > > appropriate XOR constraint be added to the spec., and that > > the DTD be > > > fixed. In the meantime, someone needs to investigate the DTD > > > generator that was used (I'd offer to do it myself, but I > > do not own a > > > copy of the XMI tool used to produce the CWM 1.0 DTD). > > > > > > Best regards, > > > John > > > > > > > > > > > > > > > > > > > > > "Pete Rivett" > > > > > > > > > > > tive.com> cc: > > > > > > Subject: > RE: issue > > > 5697 -- CWM RTF issue > > > 10/24/2002 10:05 > > > > > > AM > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > It's fine in MOF for a class to potentially be owned via > > more than one > > > composition: the restriction is that a single instance can > > only have > > > one owner. So that means that the compositions must be > 0..1 rather > > > than 1..1 which they are in this case in CWM. As it > stands it means > > > that a ModelElement can own tags and a Stereotype can own > > tags but the > > > same tag cannot be owned by both. > > > > > > The particular issue of the missing reference from > ModelElement to > > > TaggedValue I thought was fixed by issue 4408 (which was > > dealt with as > > > an urgent Issue). > > > > > > Pete > > > > > > > -----Original Message----- > > > > From: Juergen Boldt [mailto:juergen@omg.org] > > > > Sent: 24 October 2002 14:17 > > > > To: issues@omg.org; cwm-rtf@omg.org > > > > Subject: issue 5697 -- CWM RTF issue > > > > > > > > > > > > This is issue # 5697 Dan Chang [mailto:dtchang@us.ibm.com] > > > > > > > > TaggedValue > > > > > > > > We have encountered a problem in our CWM enablement: > > > CWM_1.0.dtd would > > > > not allow one to attach TaggedValue to ModelElement. I would > > > > appreciate your immediate attention. > > > > > > > > > > > > I believe the error is caused by associations defined on > > > TaggedValue. > > > > In Figure 7-3-3 (CWM 1.0), TaggedValue is owned by both > > > ModelElement > > > > and Stereotype. I don't think this should be allowed. The > > > consequence > > > > is that, arbitrarily, in CWM_1.0.dtd, ModelElement does > > not have a > > > > child element ModelElement.taggedValue but Stereotype > does have a > > > > child element Stereotype.requiredTag. The consequence > is that one > > > > cannot simply "attach" TaggedValue to ModelElement as what > > > > TaggedValue is intended for. I believe the association between > > > > Stereotype and TaggedValue should be a normal association > > > > > > > > ================================= > > > > Juergen Boldt > > > > Director, Member Services > > > > > > > > Object Management Group > > > > 250 First Avenue, Suite 100 > > > > Needham, MA 02494 > > > > > > > > Tel. +1 781 444 0404 ext. 132 > > > > Fax: +1 781 444 0320 > > > > email: juergen@omg.org > > > > www www.omg.org > > > > > > > > ================================ > > > > > > > > > > > > > > The information contained in this email and any attached > files are > > > confidential and intended solely for the addressee(s). The > > e-mail may > > > be legally privileged or prohibited from disclosure and > > unauthorised > > > use. > > > > > > If you are not the named addressee you may not use, copy or > > disclose > > > this information to any other person. If you received this > > message in > > > error please notify the sender immediately. > > > > > > Any views or opinions presented here may be solely those of the > > > originator and do not necessarily reflect those of the Company. > > > > > > > > > > > > > > > > > > > The information contained in this email and any attached files are > > confidential and intended solely for the addressee(s). The > e-mail may > > be legally privileged or prohibited from disclosure and > unauthorised > > use. > > > > If you are not the named addressee you may not use, copy or > disclose > > this information to any other person. If you received this > message in > > error please notify the sender immediately. > > > > Any views or opinions presented here may be solely those of the > > originator and do not necessarily reflect those of the Company. > > > > > > > > > > > > > > > > > > > > > > The information contained in this email and any attached > files are confidential and intended solely for the > addressee(s). The e-mail may be legally privileged or > prohibited from disclosure and unauthorised use. > > If you are not the named addressee you may not use, copy or > disclose this information to any other person. If you > received this message in error please notify the sender immediately. > > Any views or opinions presented here may be solely those of > the originator and do not necessarily reflect those of the Company. > > > > > > The information contained in this email and any attached files are confidential and intended solely for the addressee(s). The e-mail may be legally privileged or prohibited from disclosure and unauthorised use. If you are not the named addressee you may not use, copy or disclose this information to any other person. If you received this message in error please notify the sender immediately. Any views or opinions presented here may be solely those of the originator and do not necessarily reflect those of the Company. Subject: RE: issue 5697 -- CWM RTF issue - the UML DTD is wrong To: "Pete Rivett" Cc: "Dan Chang" , cwm-rtf@omg.org X-Mailer: Lotus Notes Release 5.0.8 June 18, 2001 From: "Sridhar Iyengar" Date: Mon, 28 Oct 2002 14:09:44 -0500 X-MIMETrack: Serialize by Router on D03NM120/03/M/IBM(Release 5.0.10 |March 22, 2002) at 10/28/2002 01:10:54 PM Pete said: >UML has a separate 'concrete' or 'interchange' metamodel which does have >explicit references. This is an important point. While 'explicit references' are not in UML 1.x, they are in MOF 1.x and used to define the UML 1.x interchange metamodel. Looking ahead - The concept of references (it is now called NavigableEnd) is very likely going to be in UML 2.0 - and MOF 2.0 (I say 'very likely' only because we have not taken the final adoption vote) - The U2P proposal (see the latest submitted draft) has the concept of 'NavigableEnd' which is modeled as a subtype of 'Attribute' and 'AssociationEnd'. This is intended to provide the functionality of a MOF Reference. In fact the associationend which connects NavigableEnd to Classifier is called 'reference' so that the programming model for MOF will be consistent between 1.x and 2.0 MOF applications. This unification is an important UML2/MOF2 goal. Discussions between UML2 and MOF2 teams continue on this topic. Sridhar --------------------------------------------------------------------------------------------------------- Sridhar Iyengar Distinguished Engineer Application and Integration Middleware IBM : Research Triangle Park, NC E-Mail: siyengar@us.ibm.com Phone: (919) 486-1768 --------------------------------------------------------------------------------------------------------- |---------+----------------------------> | | "Pete Rivett" | | | | | | | | | 10/28/2002 12:52 | | | PM | | | | |---------+----------------------------> >------------------------------------------------------------------------------------------------------------------------------| | | | To: Dan Chang/Santa Teresa/IBM@IBMUS | | cc: | | Subject: RE: issue 5697 -- CWM RTF issue - the UML DTD is wrong | | | | | >------------------------------------------------------------------------------------------------------------------------------| Hi Dan, UML has a separate 'concrete' or 'interchange' metamodel which does have explicit references. The UML issue I raised discussed the pros and cons of 2 possible resolutions: - make the metamodel match the DTD - make the DTD match the metamodel I'm not a member of the UML RTF so don't know if it's been discussed on their telcons: I am on the mailing list and have seen little if no discussion there. The issue was however discussed at the UML XMI Interoperability Workshop at OMG TC in Helsinki, and it was agreed by the vendors present that this was a genuine problem (in fact the 'public enemy number one' standing in the way of UML interoperability). No every composition does not need an explicit reference from the owner: however if not then the owned elements must be held as 'top level' elements in the XMI file which have a reference back to the owner. This is what several metamodel-driven tools implement (e.g. Unisys Rose integration, Interactive Objects, Sun MDR). BTW I have knocked up a XSLT stylesheet to convert XMI files with nested Tags into the correct 'top level' representation. Pete > -----Original Message----- > From: Dan Chang [mailto:dtchang@us.ibm.com] > Sent: 28 October 2002 17:16 > To: Pete Rivett > Cc: cwm-rtf@omg.org > Subject: RE: issue 5697 -- CWM RTF issue - the UML DTD is wrong > > > > Does the UML team agree with your branding of UML 1.3 DTD as > "incorrect"? UML 1.3, unlike MOF, does not have explicit references. > > Also, are you saying every composition must have an explicit > reference defined on the owner ModelElement for XMI to work? > Is this specified in MOF or XMI? > > Regards, Dan > > IBM Almaden Research Center/Silicon Valley Lab > Notes: Dan Chang/Santa Teresa/IBM@IBMUS > Internet: dtchang@us.ibm.com > Phone: (408)-927-1714/(408)-463-2319 > Office: ARC(G1-018)/SVL(D168) > > > > > > "Pete Rivett" > > > Chang/Santa Teresa/IBM@IBMUS > > tive.com> cc: > > > Subject: RE: > issue 5697 -- CWM RTF issue - the UML DTD is wrong > > 10/26/2002 01:19 > > > AM > > > > > > > > > > > > Hi Dan, > Actually the part of the UML DTD you point out is *incorrect* > (due to its inconsistency with the metamodel) and this is > causing a fair amount of hassle for interoperability, since > some people have implementations based on the DTD and others > based on the metamodel and they can't interoperate. See UML > Issue 5525. > > The CWM situation is different since the metamodel *does* > have the reference (with the resolution to 4408). > > Pete > > > -----Original Message----- > > From: Dan Chang [mailto:dtchang@us.ibm.com] > > Sent: 25 October 2002 19:15 > > To: John_Poole@hyperion.com > > Cc: cwm-rtf@omg.org; Pete Rivett > > Subject: RE: issue 5697 -- CWM RTF issue > > > > > > > > John, > > > > I agree with you. The DTD should not have been wrong in the first > > place. Though fixing #4408 will incidentally fix #5697. > > > > If one takes a look at UML 1.3, ModelElement doesn't have a > reference > > to TaggedVaule, but its DTD is correct. > > > > Regards, Dan > > > > IBM Almaden Research Center/Silicon Valley Lab > > Notes: Dan Chang/Santa Teresa/IBM@IBMUS > > Internet: dtchang@us.ibm.com > > Phone: (408)-927-1714/(408)-463-2319 > > Office: ARC(G1-018)/SVL(D168) > > > > > > > > > > > > John_Poole@hyperi > > > > > > on.com To: > > "Pete Rivett" > > > > cc: cwm-rtf@omg.org > > > > 10/25/2002 10:36 Subject: RE: > > issue 5697 -- CWM RTF issue > > > > AM > > > > > > > > > > > > > > > > > > > > > > > > > > But my point is that the DTD should not have been wrong in > the first > > place. The CWM ModelElement represented in the DTD should > still have > > contained an XML Element representing owned TaggedValues, > by virtue of > > the composition relationship from ModelElement to > TaggedValue in the > > CWM model. > > > > It seems like the DTD only got corrected after a "taggedValue" > > reference was added to ModelElement; hence, my conclusion > that perhaps > > the DTD generation process is being driven by the presence (or > > absence) of references, rather than the associations upon which the > > references are based. > > > > Do you agree? > > > > Regards, > > John > > > > > > > > > > > > "Pete Rivett" > > > > > Poole/na/Hyperion@Hyperion, > > tive.com> cc: > > > > Subject: RE: issue > > 5697 -- CWM RTF issue > > 10/25/02 07:26 AM > > > > > > > > > > > > > > > > With respect to the DTD, I don't think there's a generator > > problem: I think it's just that we have not > generated/issued a CWM 1.0 > > DTD with the fix to issue 4408 applied. In fact I don't think we've > > issued ANY updated CWM 1.0 artifacts (mdl, XMI, IDL) with just that > > emergency fix applied. > > > > Is that something we want to do? > > Then there's the issue of how we catalog it and distinguish it from > > the previous CWM 1.0 DTD. > > > > Pete > > > > > -----Original Message----- > > > From: John_Poole@hyperion.com [mailto:John_Poole@hyperion.com] > > > Sent: 24 October 2002 18:47 > > > To: cwm-rtf@omg.org > > > Subject: RE: issue 5697 -- CWM RTF issue > > > > > > > > > > > > Hi all, > > > > > > I think I know (approximately) where the problem lies... > > > > > > By way of comparison, there is an identical situation in > > the CWM OLAP > > > metamodel, in which an instance of DimensionDeployment may > > be owned by > > > at most one instance of HierarchyLevelAssociation or by > at most one > > > instance of ValueBasedHierarchy, but not by both at the > > same time, and > > > this situation is formally expressed by an OCL constraint > > of the form: > > > > > > context DimensionDeployment inv: > > > self.hierarchyLevelAssociation->isEmpty xor > > > self.valueBasedHierarchy->isEmpty > > > > > > (Of course, the constraint is just documentation, is not > > embedded in > > > the Rose model, and could not have possibly influenced the DTD > > > generation in any way). > > > > > > The main thing that is different between this part of the > > OLAP model > > > and the Core classes ModelElement and TaggedValue is the > > absense of an > > > explicit reference on ModelElement corresponding to the > > > TaggedElement::taggedValue association end, where as both > > > HierarchyLevelAssociation and ValueBasedHierarchy both have > > explicit > > > references to DimensionDeployment. There is another > > difference in that > > > both the association ends touching DimenisonDeployment are also > > > ordered, whereas this is not the case with the taggedValue / > > > requiredTag ends. Other than these two points, I see > nothing else in > > > the class properties and/or features that might otherwise > > > distinguish the two situations. > > > > > > So, I suspect that the absense of a TaggedValue reference on > > > ModelElement is the root cause of the problem. Is it > possible that > > > the DTD generation process was driven off the references, > > and not the > > > association ends? Or could this point to a possible error in XMI > > > itself? > > > > > > In any event, the absence of a TaggedValue reference in > > ModelElement > > > had previously been raised by issue 4408 and certainly is indeed > > > problematic for any CWM implementation not using MOF. > > > > > > Also, we will need to update the CWM specification to > > include an XOR > > > constraint in the Core model for ownership of TaggedValue. > > > > > > SO IN SUMMARY: I believe this is either a DTD generator > problem or > > > possibly an XMI issue. It is not a problem in the CWM > > model itself. > > > I think Dan's current open issue should be amended to > > require that an > > > appropriate XOR constraint be added to the spec., and that > > the DTD be > > > fixed. In the meantime, someone needs to investigate the DTD > > > generator that was used (I'd offer to do it myself, but I > > do not own a > > > copy of the XMI tool used to produce the CWM 1.0 DTD). > > > > > > Best regards, > > > John > > > > > > > > > > > > > > > > > > > > > "Pete Rivett" > > > > > > > > > > > tive.com> cc: > > > > > > Subject: > RE: issue > > > 5697 -- CWM RTF issue > > > 10/24/2002 10:05 > > > > > > AM > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > It's fine in MOF for a class to potentially be owned via > > more than one > > > composition: the restriction is that a single instance can > > only have > > > one owner. So that means that the compositions must be > 0..1 rather > > > than 1..1 which they are in this case in CWM. As it > stands it means > > > that a ModelElement can own tags and a Stereotype can own > > tags but the > > > same tag cannot be owned by both. > > > > > > The particular issue of the missing reference from > ModelElement to > > > TaggedValue I thought was fixed by issue 4408 (which was > > dealt with as > > > an urgent Issue). > > > > > > Pete > > > > > > > -----Original Message----- > > > > From: Juergen Boldt [mailto:juergen@omg.org] > > > > Sent: 24 October 2002 14:17 > > > > To: issues@omg.org; cwm-rtf@omg.org > > > > Subject: issue 5697 -- CWM RTF issue > > > > > > > > > > > > This is issue # 5697 Dan Chang [mailto:dtchang@us.ibm.com] > > > > > > > > TaggedValue > > > > > > > > We have encountered a problem in our CWM enablement: > > > CWM_1.0.dtd would > > > > not allow one to attach TaggedValue to ModelElement. I would > > > > appreciate your immediate attention. > > > > > > > > > > > > I believe the error is caused by associations defined on > > > TaggedValue. > > > > In Figure 7-3-3 (CWM 1.0), TaggedValue is owned by both > > > ModelElement > > > > and Stereotype. I don't think this should be allowed. The > > > consequence > > > > is that, arbitrarily, in CWM_1.0.dtd, ModelElement does > > not have a > > > > child element ModelElement.taggedValue but Stereotype > does have a > > > > child element Stereotype.requiredTag. The consequence > is that one > > > > cannot simply "attach" TaggedValue to ModelElement as what > > > > TaggedValue is intended for. I believe the association between > > > > Stereotype and TaggedValue should be a normal association > > > > > > > > ================================= > > > > Juergen Boldt > > > > Director, Member Services > > > > > > > > Object Management Group > > > > 250 First Avenue, Suite 100 > > > > Needham, MA 02494 > > > > > > > > Tel. +1 781 444 0404 ext. 132 > > > > Fax: +1 781 444 0320 > > > > email: juergen@omg.org > > > > www www.omg.org > > > > > > > > ================================ > > > > > > > > > > > > > > The information contained in this email and any attached > files are > > > confidential and intended solely for the addressee(s). The > > e-mail may > > > be legally privileged or prohibited from disclosure and > > unauthorised > > > use. > > > > > > If you are not the named addressee you may not use, copy or > > disclose > > > this information to any other person. If you received this > > message in > > > error please notify the sender immediately. > > > > > > Any views or opinions presented here may be solely those of the > > > originator and do not necessarily reflect those of the Company. > > > > > > > > > > > > > > > > > > > The information contained in this email and any attached files are > > confidential and intended solely for the addressee(s). The > e-mail may > > be legally privileged or prohibited from disclosure and > unauthorised > > use. > > > > If you are not the named addressee you may not use, copy or > disclose > > this information to any other person. If you received this > message in > > error please notify the sender immediately. > > > > Any views or opinions presented here may be solely those of the > > originator and do not necessarily reflect those of the Company. > > > > > > > > > > > > > > > > > > > > > > The information contained in this email and any attached > files are confidential and intended solely for the > addressee(s). The e-mail may be legally privileged or > prohibited from disclosure and unauthorised use. > > If you are not the named addressee you may not use, copy or > disclose this information to any other person. If you > received this message in error please notify the sender immediately. > > Any views or opinions presented here may be solely those of > the originator and do not necessarily reflect those of the Company. > > > > > > The information contained in this email and any attached files are confidential and intended solely for the addressee(s). The e-mail may be legally privileged or prohibited from disclosure and unauthorised use. If you are not the named addressee you may not use, copy or disclose this information to any other person. If you received this message in error please notify the sender immediately. Any views or opinions presented here may be solely those of the originator and do not necessarily reflect those of the Company.