Issue 4231: predefined tag for marking attributes to act as qualifier in classifier (mof-rtf) Source: France Telecom R&D (Mr. Mariano Belaunde, mariano.belaunde(at)orange.com) Nature: Uncategorized Issue Severity: Summary: Need a predefined tag for marking attributes that will 'act as a qualifier' in a classifier. In a metamodelling tool it may be useful to derive automatically an object identifier from one or more relevant attributes of the classifier (for instance a 'name' attribute). Note: The attribute 'acting as a qualifier' is owned by the classifier not by an association. Suggestion: pre-define a Tag named 'actAsQualifier : Boolean' applicable to any attribute of a classifier. Resolution: Revised Text: Actions taken: March 21, 2001: received issue Discussion: End of Annotations:===== From: BELAUNDE Mariano FTRD/DTL/LAN To: "'Juergen Boldt'" Subject: Two Ossues to the MOF 1.4 RTF Date: Wed, 21 Mar 2001 12:25:05 +0100 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain; charset="ISO-8859-1" X-UIDL: EWG!!f;h!!##e!!"]&!! Juergen, I would like to post the following two issues to the MOF 1.4 RTF. Thanks in advance, Mariano ----------------- ISSUE Title: Need a predefined tag for marking attributes that will 'act as a qualifier' in a classifier. In a metamodelling tool it may be useful to derive automatically an object identifier from one or more relevant attributes of the classifier (for instance a 'name' attribute). Note: The attribute 'acting as a qualifier' is owned by the classifier not by an association. Suggestion: pre-define a Tag named 'actAsQualifier : Boolean' applicable to any attribute of a classifier. X-Mailer: exmh version 2.2 06/23/2000 with nmh-1.0.4 To: mof-rtf@omg.org, mariano.belaunde@rd.francetelecom.fr Subject: MOF RTF issues 4231 Mime-Version: 1.0 Date: Thu, 22 Mar 2001 09:50:16 +1000 From: Stephen Crawley Content-Type: text/plain; charset=us-ascii X-UIDL: >"5e9#mZd9>Y#e9!6G!! > Issue: "predefined tag for marking attributes to act as qualifier in > classifier" > Description: > Need a predefined tag for marking attributes that will 'act as a > qualifier' in a classifier. In a metamodelling tool it may be useful > to derive automatically an object identifier from one or more > relevant > attributes of the classifier (for instance a 'name' > attribute). Note: > The attribute 'acting as a qualifier' is owned by the classifier not > by an association. Suggestion: pre-define a Tag named > 'actAsQualifier > : Boolean' applicable to any attribute of a classifier. I don't think that this belongs in MOF 1.x. It only makes sense if a standard technology mapping (i.e. MOF -> IDL or MOF -> XMI) supports a general notion of identifying objects based on Attribute values. The mappings don't support this kind of thing now, and adding support at this stage is unnecessarily disruptive. I think this is an idea to be considered by submitters against the forthcoming MOF 2.0 RFP. In the meantime, there is nothing stopping Mariano (and others) from inventing a non-standard Tag id to do this sort of thing. -- Steve X-Mailer: exmh version 2.2 06/23/2000 with nmh-1.0.4 To: mariano.belaunde@rd.francetelecom.fr cc: mof-rtf@omg.org Subject: Re: Issue 4231: predefined tag for marking attributes to act as qualifier Mime-Version: 1.0 Date: Fri, 06 Apr 2001 15:43:29 +1000 From: Stephen Crawley X-Scanned-By: MIMEDefang 1.0 (http://www.roaringpenguin.com/mimedefang/) Content-Type: text/plain; charset=us-ascii X-UIDL: Jjod92X%!!$>*!!5*,!! Mariano, The RTF members at today's teleconference agreed with your suggestion in principle. Can you supply a suitable definition of what it means for an Attribute to "act as a qualifier". We will consider your definition, and if it is appropriate, we will define a new TagId. Bear in mind that it is not the intention of the RTF that this new TagId should modify the IDL mapping generated by the IDL mapping or the behaviour of meta-objects as prescribed by the MOF spec. Any use of the new TagId would be beyond the scope of the MOF spec. Thanks, -- Steve From: BELAUNDE Mariano FTRD/DTL/LAN To: "'Stephen Crawley'" Cc: mof-rtf Subject: RE: Issue 4231: predefined tag for marking attributes to act as q ualifier Date: Fri, 6 Apr 2001 11:29:29 +0200 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain X-UIDL: &FS!![Y7e9'?pd9#HT!! Hi Steve, >> The RTF members at today's teleconference agreed with your suggestion >> in principle. Good decision. This gives the chance for two distinct implementors to share the same convention for capturing this information. Here's the proposed meaning definition: ActAsQualifier tag id: "org.omg.MOF:act_as_qualifier" attaches to: Model::Attribute values: a Boolean meaning: The values assigned to the attributes that are marked with this tag can be used to provide an identity to the class instance owning the attributes. idl generation: no impact Regards, Mariano -----Message d'origine----- De : Stephen Crawley [mailto:crawley@dstc.edu.au] Envoye : vendredi 6 avril 2001 07:43 A : mariano.belaunde@rd.francetelecom.fr Cc : mof-rtf@omg.org Objet : Re: Issue 4231: predefined tag for marking attributes to act as qualifier Mariano, The RTF members at today's teleconference agreed with your suggestion in principle. Can you supply a suitable definition of what it means for an Attribute to "act as a qualifier". We will consider your definition, and if it is appropriate, we will define a new TagId. Bear in mind that it is not the intention of the RTF that this new TagId should modify the IDL mapping generated by the IDL mapping or the behaviour of meta-objects as prescribed by the MOF spec. Any use of the new TagId would be beyond the scope of the MOF spec. Thanks, -- Steve From: "Iyengar, Sridhar" To: BELAUNDE Mariano FTRD/DTL/LAN , "'Stephen Crawley'" Cc: mof-rtf Subject: RE: Issue 4231: predefined tag for marking attributes to act as q ualifier Date: Fri, 6 Apr 2001 11:26:38 -0500 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain; charset="iso-8859-1" X-UIDL: p-1!!@9-!!V -----Original Message----- > From: BELAUNDE Mariano FTRD/DTL/LAN > [mailto:mariano.belaunde@rd.francetelecom.fr] > Sent: Friday, April 06, 2001 2:29 AM > To: 'Stephen Crawley' > Cc: mof-rtf > Subject: RE: Issue 4231: predefined tag for marking > attributes to act as > q ualifier > > > Hi Steve, > > >> The RTF members at today's teleconference agreed with your > suggestion > >> in principle. > Good decision. This gives the chance for two distinct > implementors to share > the same convention for capturing this information. > Here's the proposed meaning definition: > > ActAsQualifier > tag id: "org.omg.MOF:act_as_qualifier" > attaches to: Model::Attribute > values: a Boolean > meaning: The values assigned to the attributes that are > marked with this > tag > can be used to provide an identity to the class > instance owning > the > attributes. > idl generation: no impact > > Regards, > > Mariano > > -----Message d'origine----- > De : Stephen Crawley [mailto:crawley@dstc.edu.au] > Envoye : vendredi 6 avril 2001 07:43 > A : mariano.belaunde@rd.francetelecom.fr > Cc : mof-rtf@omg.org > Objet : Re: Issue 4231: predefined tag for marking attributes > to act as > qualifier > > > > Mariano, > > The RTF members at today's teleconference agreed with your > suggestion in > principle. Can you supply a suitable definition of what it > means for an > Attribute to "act as a qualifier". We will consider your > definition, and > if it is appropriate, we will define a new TagId. > > Bear in mind that it is not the intention of the RTF that > this new TagId > should modify the IDL mapping generated by the IDL mapping or the > behaviour of meta-objects as prescribed by the MOF spec. Any > use of the > new TagId would be beyond the scope of the MOF spec. > > Thanks, > > -- Steve > Date: Fri, 06 Apr 2001 13:59:49 -0400 From: Peter Denno X-Mailer: Mozilla 4.73 [en] (WinNT; U) X-Accept-Language: en To: BELAUNDE Mariano FTRD/DTL/LAN CC: "'Stephen Crawley'" , mof-rtf Subject: Re: Issue 4231: predefined tag for marking attributes to act as qualifier Hi, Mariano's solution is fine where a single attribute provides a unique identifier, but maybe we could also address situations where a combination of attributes provide a unique identifier (i.e. in relational terms, a 'concatenated' or 'compound' key) The Express information modeling language (ISO 10303-11), which I hope could be supported, has something similar called a uniqueness constraint. It states that no two instances shall have the same combination of values for the attributes composing the constraint. BELAUNDE Mariano FTRD/DTL/LAN wrote: > Hi Steve, > > >> The RTF members at today's teleconference agreed with your >suggestion > >> in principle. > Good decision. This gives the chance for two distinct implementors >to share > the same convention for capturing this information. > Here's the proposed meaning definition: > > ActAsQualifier > tag id: "org.omg.MOF:act_as_qualifier" > attaches to: Model::Attribute > values: a Boolean > meaning: The values assigned to the attributes that are marked >with this > tag > can be used to provide an identity to the class >instance owning > the > attributes. > idl generation: no impact > > Regards, > > Mariano > > -----Message d'origine----- > De : Stephen Crawley [mailto:crawley@dstc.edu.au] > Envoye : vendredi 6 avril 2001 07:43 > A : mariano.belaunde@rd.francetelecom.fr > Cc : mof-rtf@omg.org > Objet : Re: Issue 4231: predefined tag for marking attributes to act >as > qualifier > > Mariano, > > The RTF members at today's teleconference agreed with your >suggestion in > principle. Can you supply a suitable definition of what it means >for an > Attribute to "act as a qualifier". We will consider your >definition, and > if it is appropriate, we will define a new TagId. > > Bear in mind that it is not the intention of the RTF that this new >TagId > should modify the IDL mapping generated by the IDL mapping or the > behaviour of meta-objects as prescribed by the MOF spec. Any use of >the > new TagId would be beyond the scope of the MOF spec. > > Thanks, > > -- Steve -- Best Regards, - Peter Peter Denno National Institute of Standards and Technology, Manufacturing System Integration Division, 100 Bureau Drive, Mail Stop 8260 Tel: +1 301-975-3595 Gaithersburg, MD, USA 20899-8260 FAX: +1 301-975-4694 From: "Iyengar, Sridhar" To: Peter Denno , BELAUNDE Mariano FTRD/DTL/LAN Cc: "'Stephen Crawley'" , mof-rtf Subject: RE: Issue 4231: predefined tag for marking attributes to act as qualifier Date: Fri, 6 Apr 2001 14:16:52 -0500 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain; charset="iso-8859-1" X-UIDL: W:"!!U0*e9X9=!!);l!! Agreed : Compound keys (ala set of attributes/references) also need to be supported. ------------------------------------------------------------------------ ---- Sridhar Iyengar Unisys Fellow, Director of Advanced Technology 25725, Jeronimo Road Mission Viejo, CA 92691 E-mail : Sridhar.iyengar2@unisys.com Phone : 949-380-5692 Fax : 949-380-6632 ------------------------------------------------------------------------ -------------- > -----Original Message----- > From: Peter Denno [mailto:peter.denno@nist.gov] > Sent: Friday, April 06, 2001 11:00 AM > To: BELAUNDE Mariano FTRD/DTL/LAN > Cc: 'Stephen Crawley'; mof-rtf > Subject: Re: Issue 4231: predefined tag for marking > attributes to act as > qualifier > > > Hi, > > Mariano's solution is fine where a single attribute provides a unique > identifier, but maybe we could also address situations where > a combination of > attributes provide a unique identifier (i.e. in relational terms, a > 'concatenated' or 'compound' key) The Express information > modeling language (ISO > 10303-11), which I hope could be supported, has something > similar called a > uniqueness constraint. It states that no two instances shall > have the same > combination of values for the attributes composing the constraint. > > > BELAUNDE Mariano FTRD/DTL/LAN wrote: > > > Hi Steve, > > > > >> The RTF members at today's teleconference agreed with > your suggestion > > >> in principle. > > Good decision. This gives the chance for two distinct > implementors to share > > the same convention for capturing this information. > > Here's the proposed meaning definition: > > > > ActAsQualifier > > tag id: "org.omg.MOF:act_as_qualifier" > > attaches to: Model::Attribute > > values: a Boolean > > meaning: The values assigned to the attributes that are > marked with this > > tag > > can be used to provide an identity to the > class instance owning > > the > > attributes. > > idl generation: no impact > > > > Regards, > > > > Mariano > > > > -----Message d'origine----- > > De : Stephen Crawley [mailto:crawley@dstc.edu.au] > > Envoye : vendredi 6 avril 2001 07:43 > > A : mariano.belaunde@rd.francetelecom.fr > > Cc : mof-rtf@omg.org > > Objet : Re: Issue 4231: predefined tag for marking > attributes to act as > > qualifier > > > > Mariano, > > > > The RTF members at today's teleconference agreed with your > suggestion in > > principle. Can you supply a suitable definition of what it > means for an > > Attribute to "act as a qualifier". We will consider your > definition, and > > if it is appropriate, we will define a new TagId. > > > > Bear in mind that it is not the intention of the RTF that > this new TagId > > should modify the IDL mapping generated by the IDL mapping or the > > behaviour of meta-objects as prescribed by the MOF spec. > Any use of the > > new TagId would be beyond the scope of the MOF spec. > > > > Thanks, > > > > -- Steve > > -- > Best Regards, > - Peter > > Peter Denno > National Institute of Standards and Technology, > Manufacturing System Integration Division, > 100 Bureau Drive, Mail Stop 8260 Tel: +1 301-975-3595 > Gaithersburg, MD, USA 20899-8260 FAX: +1 301-975-4694 > > Date: Fri, 06 Apr 2001 15:31:54 -0400 From: Peter Denno X-Mailer: Mozilla 4.73 [en] (WinNT; U) X-Accept-Language: en MIME-Version: 1.0 To: "Iyengar, Sridhar" CC: BELAUNDE Mariano FTRD/DTL/LAN , "'Stephen Crawley'" , mof-rtf Subject: Re: Issue 4231: predefined tag for marking attributes to act as qualifier References: Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=us-ascii X-UIDL: FOM!!jSMe9Z)nd9,S?e9 Unless Mariano, Stephen or anyone else would like to write some text for this, I am willing to give it a try. I am not as familiar with the spec as other here though, so it might be a few days before I have something. "Iyengar, Sridhar" wrote: > Agreed : Compound keys (ala set of attributes/references) also need to be > supported. > > ------------------------------------------------------------------------ > ---- > Sridhar Iyengar > Unisys Fellow, Director of Advanced Technology > 25725, Jeronimo Road > Mission Viejo, CA 92691 > > E-mail : Sridhar.iyengar2@unisys.com > Phone : 949-380-5692 > Fax : 949-380-6632 > ------------------------------------------------------------------------ > -------------- > > > -----Original Message----- > > From: Peter Denno [mailto:peter.denno@nist.gov] > > Sent: Friday, April 06, 2001 11:00 AM > > To: BELAUNDE Mariano FTRD/DTL/LAN > > Cc: 'Stephen Crawley'; mof-rtf > > Subject: Re: Issue 4231: predefined tag for marking > > attributes to act as > > qualifier > > > > > > Hi, > > > > Mariano's solution is fine where a single attribute provides a unique > > identifier, but maybe we could also address situations where > > a combination of > > attributes provide a unique identifier (i.e. in relational terms, a > > 'concatenated' or 'compound' key) The Express information > > modeling language (ISO > > 10303-11), which I hope could be supported, has something > > similar called a > > uniqueness constraint. It states that no two instances shall > > have the same > > combination of values for the attributes composing the constraint. > > > > > > BELAUNDE Mariano FTRD/DTL/LAN wrote: > > > > > Hi Steve, > > > > > > >> The RTF members at today's teleconference agreed with > > your suggestion > > > >> in principle. > > > Good decision. This gives the chance for two distinct > > implementors to share > > > the same convention for capturing this information. > > > Here's the proposed meaning definition: > > > > > > ActAsQualifier > > > tag id: "org.omg.MOF:act_as_qualifier" > > > attaches to: Model::Attribute > > > values: a Boolean > > > meaning: The values assigned to the attributes that are > > marked with this > > > tag > > > can be used to provide an identity to the > > class instance owning > > > the > > > attributes. > > > idl generation: no impact > > > > > > Regards, > > > > > > Mariano > > > > > > -----Message d'origine----- > > > De : Stephen Crawley [mailto:crawley@dstc.edu.au] > > > Envoye : vendredi 6 avril 2001 07:43 > > > A : mariano.belaunde@rd.francetelecom.fr > > > Cc : mof-rtf@omg.org > > > Objet : Re: Issue 4231: predefined tag for marking > > attributes to act as > > > qualifier > > > > > > Mariano, > > > > > > The RTF members at today's teleconference agreed with your > > suggestion in > > > principle. Can you supply a suitable definition of what it > > means for an > > > Attribute to "act as a qualifier". We will consider your > > definition, and > > > if it is appropriate, we will define a new TagId. > > > > > > Bear in mind that it is not the intention of the RTF that > > this new TagId > > > should modify the IDL mapping generated by the IDL mapping or the > > > behaviour of meta-objects as prescribed by the MOF spec. > > Any use of the > > > new TagId would be beyond the scope of the MOF spec. > > > > > > Thanks, > > > > > > -- Steve > > > > -- > > Best Regards, > > - Peter > > > > Peter Denno > > National Institute of Standards and Technology, > > Manufacturing System Integration Division, > > 100 Bureau Drive, Mail Stop 8260 Tel: +1 301-975-3595 > > Gaithersburg, MD, USA 20899-8260 FAX: +1 301-975-4694 > > > > -- Best Regards, - Peter Peter Denno National Institute of Standards and Technology, Manufacturing System Integration Division, 100 Bureau Drive, Mail Stop 8260 Tel: +1 301-975-3595 Gaithersburg, MD, USA 20899-8260 FAX: +1 301-975-4694 X-Mailer: exmh version 2.2 06/23/2000 with nmh-1.0.4 To: Peter Denno cc: "Iyengar, Sridhar" , BELAUNDE Mariano FTRD/DTL/LAN , "'Stephen Crawley'" , mof-rtf , crawley@dstc.edu.au Subject: Re: Issue 4231: predefined tag for marking attributes to act as qualifier In-Reply-To: Message from Peter Denno of "Fri, 06 Apr 2001 15:31:54 -0400." <3ACE19AA.64C8DB86@nist.gov> Mime-Version: 1.0 Date: Sun, 08 Apr 2001 11:17:30 +1000 From: Stephen Crawley X-Scanned-By: MIMEDefang 1.0 (http://www.roaringpenguin.com/mimedefang/) Content-Type: text/plain; charset=us-ascii X-UIDL: UnSd9V)0e9<15!!2A^d9 > > Unless Mariano, Stephen or anyone else would like to write some text > for > this, I am willing to give it a try. I am not as familiar with the > spec as > other here though, so it might be a few days before I have > something. > I can do the editorial aspect of this change; i.e. working up the boiler plate for a new subsection for this (new) kind of TagId etc. But I'd prefer it if someone else with a clearer understanding of this particular TagId could provide the description. -- Steve Date: Mon, 09 Apr 2001 13:23:37 -0400 From: Peter Denno X-Mailer: Mozilla 4.73 [en] (WinNT; U) X-Accept-Language: en MIME-Version: 1.0 To: BELAUNDE Mariano FTRD/DTL/LAN CC: "'Stephen Crawley'" , mof-rtf Subject: Re: Issue 4231: predefined tag for marking attributes to act as qualifier References: Content-Transfer-Encoding: 8bit Content-Type: text/plain; charset=iso-8859-1 X-UIDL: jh@e98RZd9Tlcd9!R]d9 Mariano, I agree that we need not discuss uniqueness specifically. It is well enough to say that instances having the particular combination of values are "qualified". However, regardless of whether or not this qualifier is an unique identifier, I think the issue of implied context (population over which the qualification applies) remains. However, considering what I understand to be the purpose of tags, I guess we need not address this. One problem remains however: I don't think a boolean on the attribute or reference is sufficient for 'compound qualifiers' (combinations of values that qualify instances). If 1 or more tagged attributes are inherited, how, with only a boolean, do we determine what collection of tagged attributes refer to what instance of qualification? In order to address compound qualifiers, I suggest that the tag should be one that attaches to classes and has as a value a set of names referring to attributes or references. Something like this: ActAsQualifier: TAG ID: "omg.org.mof:act_as_qualifier ATTACHES TO: Model::Class VALUES: set [1..*] of NameType MEANING: The values of the set refer to attributes and references of the class to which the tag is attached. The combination of the values of these attributes and references of a given instance serve as a qualifier or identifier to the instance. - Peter BELAUNDE Mariano FTRD/DTL/LAN wrote: > I agree that the tag can apply to references as well as to attributes. > Taking in account this, the revised meaning definition could be: > > ActAsQualifier > tag id: "org.omg.MOF:act_as_qualifier" > attaches to: Model::Attribute, Model::Reference > values: a Boolean > meaning: The combination of values assigned to the attributes and > references > that are marked with this tag can be used to provide an identity to > the class > instance owning the attributes and the references. > idl generation: no impact > > Peter, do not hesitate to propose another definition if you feel that this > one is not > clear enough or if an extra explanation is needed. > > Note that the current proposal is not formulated in terms of any uniqueness > constraint > in order not to restrict the usage of this tag. Otherwise we will need to > know what > is the implied context (what are the set of instances that need to conform > to the > constraint). The context to use may depend on the naming facility provided > by a tool > (and it does not necessarily depends on an existing composite association > defined in the > metamodel). > > In the case we really want a way to express that in a composite association > all > the referenced parts need to have distinct combinations of values for all > the > slots (attributes/references) "acting as qualifiers", I believe that this > can be > achieved by adding an extra 'isQualified' tag to the composite association. > Of course, if you agree on the latter suggestion to take into account the > Peter's > "uniqueness constraint", we will also need to provide a meaning definition > for the 'isQualified' tag. > > Mariano > > -----Message d'origine----- > De : Peter Denno [mailto:peter.denno@nist.gov] > Envoy : vendredi 6 avril 2001 20:00 > : BELAUNDE Mariano FTRD/DTL/LAN > Cc : 'Stephen Crawley'; mof-rtf > Objet : Re: Issue 4231: predefined tag for marking attributes to act as > qualifier > > Hi, > > Mariano's solution is fine where a single attribute provides a unique > identifier, but maybe we could also address situations where a combination > of > attributes provide a unique identifier (i.e. in relational terms, a > 'concatenated' or 'compound' key) The Express information modeling language > (ISO > 10303-11), which I hope could be supported, has something similar called a > uniqueness constraint. It states that no two instances shall have the same > combination of values for the attributes composing the constraint. > > BELAUNDE Mariano FTRD/DTL/LAN wrote: > > > Hi Steve, > > > > >> The RTF members at today's teleconference agreed with your suggestion > > >> in principle. > > Good decision. This gives the chance for two distinct implementors to > share > > the same convention for capturing this information. > > Here's the proposed meaning definition: > > > > ActAsQualifier > > tag id: "org.omg.MOF:act_as_qualifier" > > attaches to: Model::Attribute > > values: a Boolean > > meaning: The values assigned to the attributes that are marked with > this > > tag > > can be used to provide an identity to the class instance > owning > > the > > attributes. > > idl generation: no impact > > > > Regards, > > > > Mariano > > > > -----Message d'origine----- > > De : Stephen Crawley [mailto:crawley@dstc.edu.au] > > Envoye : vendredi 6 avril 2001 07:43 > > A : mariano.belaunde@rd.francetelecom.fr > > Cc : mof-rtf@omg.org > > Objet : Re: Issue 4231: predefined tag for marking attributes to act as > > qualifier > > > > Mariano, > > > > The RTF members at today's teleconference agreed with your suggestion in > > principle. Can you supply a suitable definition of what it means for an > > Attribute to "act as a qualifier". We will consider your definition, and > > if it is appropriate, we will define a new TagId. > > > > Bear in mind that it is not the intention of the RTF that this new TagId > > should modify the IDL mapping generated by the IDL mapping or the > > behaviour of meta-objects as prescribed by the MOF spec. Any use of the > > new TagId would be beyond the scope of the MOF spec. > > > > Thanks, > > > > -- Steve > > -- > Best Regards, > - Peter > > Peter Denno > National Institute of Standards and Technology, > Manufacturing System Integration Division, > 100 Bureau Drive, Mail Stop 8260 Tel: +1 301-975-3595 > Gaithersburg, MD, USA 20899-8260 FAX: +1 301-975-4694 -- Best Regards, - Peter Peter Denno National Institute of Standards and Technology, Manufacturing System Integration Division, 100 Bureau Drive, Mail Stop 8260 Tel: +1 301-975-3595 Gaithersburg, MD, USA 20899-8260 FAX: +1 301-975-4694 From: "Iyengar, Sridhar" To: BELAUNDE Mariano FTRD/DTL/LAN , "'Peter Denno'" , "'Stephen Crawley'" Cc: mof-rtf Subject: RE: Issue 4231: predefined tag for marking attributes to act as qualifier Date: Mon, 9 Apr 2001 12:27:35 -0500 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by emerald.omg.org id f39HNYv22811 Content-Type: text/plain; charset="iso-8859-1" X-UIDL: X!@!!/Vm!!QPOd9&%`!! See [Sri] ------------------------------------------------------------------------ ---- Sridhar Iyengar Unisys Fellow, Director of Advanced Technology 25725, Jeronimo Road Mission Viejo, CA 92691 E-mail : Sridhar.iyengar2@unisys.com Phone : 949-380-5692 Fax : 949-380-6632 ------------------------------------------------------------------------ -------------- > -----Original Message----- > From: BELAUNDE Mariano FTRD/DTL/LAN > [mailto:mariano.belaunde@rd.francetelecom.fr] > Sent: Monday, April 09, 2001 4:16 AM > To: 'Peter Denno'; 'Stephen Crawley' > Cc: mof-rtf > Subject: RE: Issue 4231: predefined tag for marking > attributes to act as > qualifier > > > I agree that the tag can apply to references as well as to attributes. > Taking in account this, the revised meaning definition could be: > > ActAsQualifier > tag id: "org.omg.MOF:act_as_qualifier" > attaches to: Model::Attribute, Model::Reference [Sri] You can attach it to Model::StructuralFeature which is a superclass of reference and Attribute. Also if this is being used as a composite key, the ordering of the features becomes important qualifier (a,b,c) is NOT the same as (b,c,a) for purposes of searching/sorting etc. > values: a Boolean > meaning: The combination of values assigned to the attributes > and > references > that are marked with this tag can be used to provide > an identity to > the class > instance owning the attributes and the references. > idl generation: no impact > > Peter, do not hesitate to propose another definition if you > feel that this > one is not > clear enough or if an extra explanation is needed. > > Note that the current proposal is not formulated in terms of > any uniqueness > constraint > in order not to restrict the usage of this tag. Otherwise we > will need to > know what > is the implied context (what are the set of instances that > need to conform > to the > constraint). The context to use may depend on the naming > facility provided > by a tool > (and it does not necessarily depends on an existing composite > association > defined in the > metamodel). > > In the case we really want a way to express that in a > composite association > all > the referenced parts need to have distinct combinations of > values for all > the > slots (attributes/references) "acting as qualifiers", I > believe that this > can be > achieved by adding an extra 'isQualified' tag to the > composite association. > Of course, if you agree on the latter suggestion to take into > account the > Peter's > "uniqueness constraint", we will also need to provide a > meaning definition > for the 'isQualified' tag. > > Mariano > > -----Message d'origine----- > De : Peter Denno [mailto:peter.denno@nist.gov] > Envoy : vendredi 6 avril 2001 20:00 > : BELAUNDE Mariano FTRD/DTL/LAN > Cc : 'Stephen Crawley'; mof-rtf > Objet : Re: Issue 4231: predefined tag for marking attributes > to act as > qualifier > > > Hi, > > Mariano's solution is fine where a single attribute provides a > unique > identifier, but maybe we could also address situations where > a combination > of > attributes provide a unique identifier (i.e. in relational terms, a > 'concatenated' or 'compound' key) The Express information > modeling language > (ISO > 10303-11), which I hope could be supported, has something > similar called a > uniqueness constraint. It states that no two instances shall > have the same > combination of values for the attributes composing the constraint. > > > BELAUNDE Mariano FTRD/DTL/LAN wrote: > > > Hi Steve, > > > > >> The RTF members at today's teleconference agreed with > your suggestion > > >> in principle. > > Good decision. This gives the chance for two distinct > implementors to > share > > the same convention for capturing this information. > > Here's the proposed meaning definition: > > > > ActAsQualifier > > tag id: "org.omg.MOF:act_as_qualifier" > > attaches to: Model::Attribute > > values: a Boolean > > meaning: The values assigned to the attributes that are > marked with > this > > tag > > can be used to provide an identity to the > class instance > owning > > the > > attributes. > > idl generation: no impact > > > > Regards, > > > > Mariano > > > > -----Message d'origine----- > > De : Stephen Crawley [mailto:crawley@dstc.edu.au] > > Envoye : vendredi 6 avril 2001 07:43 > > A : mariano.belaunde@rd.francetelecom.fr > > Cc : mof-rtf@omg.org > > Objet : Re: Issue 4231: predefined tag for marking > attributes to act as > > qualifier > > > > Mariano, > > > > The RTF members at today's teleconference agreed with your > suggestion in > > principle. Can you supply a suitable definition of what it > means for an > > Attribute to "act as a qualifier". We will consider your > definition, and > > if it is appropriate, we will define a new TagId. > > > > Bear in mind that it is not the intention of the RTF that > this new TagId > > should modify the IDL mapping generated by the IDL mapping or the > > behaviour of meta-objects as prescribed by the MOF spec. > Any use of the > > new TagId would be beyond the scope of the MOF spec. > > > > Thanks, > > > > -- Steve > > -- > Best Regards, > - Peter > > Peter Denno > National Institute of Standards and Technology, > Manufacturing System Integration Division, > 100 Bureau Drive, Mail Stop 8260 Tel: +1 301-975-3595 > Gaithersburg, MD, USA 20899-8260 FAX: +1 301-975-4694 > From: "Iyengar, Sridhar" To: Peter Denno , BELAUNDE Mariano FTRD/DTL/LAN Cc: "'Stephen Crawley'" , mof-rtf Subject: RE: Issue 4231: predefined tag for marking attributes to act as qualifier Date: Mon, 9 Apr 2001 12:33:11 -0500 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain; charset="iso-8859-1" X-UIDL: `WL!!(31!!l?"!!n^a!! ActAsQualifier: TAG ID: "omg.org.mof:act_as_qualifier ATTACHES TO: Model::Class VALUES: set [1..*] of NameType [sri] should be list [1..*] of StructuralFeature otherwise there is ambiguity. The ordering of the features could be significant. This looks more and more like the specification of an index (as in relational and other databases) to a class (table, data set etc.) Need to give this some more thought. [Sri end] Importance: Normal Subject: RE: Issue 4231: predefined tag for marking attributes to act as qualifier To: "Iyengar, Sridhar" Cc: BELAUNDE Mariano FTRD/DTL/LAN , "'Peter Denno'" , "'Stephen Crawley'" , mof-rtf X-Mailer: Lotus Notes Release 5.0.4a July 24, 2000 Message-ID: From: "Stephen Brodsky" Date: Mon, 9 Apr 2001 12:01:57 -0700 X-MIMETrack: Serialize by Router on D03NM039/03/M/IBM(Release 5.0.6 |December 14, 2000) at 04/09/2001 01:09:34 PM MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by emerald.omg.org id f39J5Qv26508 Content-Type: text/plain; charset=iso-8859-1 X-UIDL: lo`d980&e9J!Ie9%6l!! This is becoming a very heavy tag. What we have in our implementations is an abstract Key class that represents an index of certain properties (ie qualifiers) over a set of objects. If MOF includes indexing capabilities, they should be modeled. The next question is "if". Thanks, -Steve Stephen A. Brodsky, Ph.D. Software Architect Notes Address: Stephen Brodsky/Santa Teresa/IBM@IBMUS Internet Address: sbrodsky@us.ibm.com Phone: 408.463.5659 "Iyengar, Sridhar" on 04/09/2001 10:27:35 AM To: BELAUNDE Mariano FTRD/DTL/LAN , "'Peter Denno'" , "'Stephen Crawley'" cc: mof-rtf Subject: RE: Issue 4231: predefined tag for marking attributes to act as qualifier See [Sri] ------------------------------------------------------------------------ ---- Sridhar Iyengar Unisys Fellow, Director of Advanced Technology 25725, Jeronimo Road Mission Viejo, CA 92691 E-mail : Sridhar.iyengar2@unisys.com Phone : 949-380-5692 Fax : 949-380-6632 ------------------------------------------------------------------------ -------------- > -----Original Message----- > From: BELAUNDE Mariano FTRD/DTL/LAN > [mailto:mariano.belaunde@rd.francetelecom.fr] > Sent: Monday, April 09, 2001 4:16 AM > To: 'Peter Denno'; 'Stephen Crawley' > Cc: mof-rtf > Subject: RE: Issue 4231: predefined tag for marking > attributes to act as > qualifier > > > I agree that the tag can apply to references as well as to attributes. > Taking in account this, the revised meaning definition could be: > > ActAsQualifier > tag id: "org.omg.MOF:act_as_qualifier" > attaches to: Model::Attribute, Model::Reference [Sri] You can attach it to Model::StructuralFeature which is a superclass of reference and Attribute. Also if this is being used as a composite key, the ordering of the features becomes important qualifier (a,b,c) is NOT the same as (b,c,a) for purposes of searching/sorting etc. > values: a Boolean > meaning: The combination of values assigned to the attributes > and > references > that are marked with this tag can be used to provide > an identity to > the class > instance owning the attributes and the references. > idl generation: no impact > > Peter, do not hesitate to propose another definition if you > feel that this > one is not > clear enough or if an extra explanation is needed. > > Note that the current proposal is not formulated in terms of > any uniqueness > constraint > in order not to restrict the usage of this tag. Otherwise we > will need to > know what > is the implied context (what are the set of instances that > need to conform > to the > constraint). The context to use may depend on the naming > facility provided > by a tool > (and it does not necessarily depends on an existing composite > association > defined in the > metamodel). > > In the case we really want a way to express that in a > composite association > all > the referenced parts need to have distinct combinations of > values for all > the > slots (attributes/references) "acting as qualifiers", I > believe that this > can be > achieved by adding an extra 'isQualified' tag to the > composite association. > Of course, if you agree on the latter suggestion to take into > account the > Peter's > "uniqueness constraint", we will also need to provide a > meaning definition > for the 'isQualified' tag. > > Mariano > > -----Message d'origine----- > De : Peter Denno [mailto:peter.denno@nist.gov] > Envoy : vendredi 6 avril 2001 20:00 > : BELAUNDE Mariano FTRD/DTL/LAN > Cc : 'Stephen Crawley'; mof-rtf > Objet : Re: Issue 4231: predefined tag for marking attributes > to act as > qualifier > > > Hi, > > Mariano's solution is fine where a single attribute provides a > unique > identifier, but maybe we could also address situations where > a combination > of > attributes provide a unique identifier (i.e. in relational terms, a > 'concatenated' or 'compound' key) The Express information > modeling language > (ISO > 10303-11), which I hope could be supported, has something > similar called a > uniqueness constraint. It states that no two instances shall > have the same > combination of values for the attributes composing the constraint. > > > BELAUNDE Mariano FTRD/DTL/LAN wrote: > > > Hi Steve, > > > > >> The RTF members at today's teleconference agreed with > your suggestion > > >> in principle. > > Good decision. This gives the chance for two distinct > implementors to > share > > the same convention for capturing this information. > > Here's the proposed meaning definition: > > > > ActAsQualifier > > tag id: "org.omg.MOF:act_as_qualifier" > > attaches to: Model::Attribute > > values: a Boolean > > meaning: The values assigned to the attributes that are > marked with > this > > tag > > can be used to provide an identity to the > class instance > owning > > the > > attributes. > > idl generation: no impact > > > > Regards, > > > > Mariano > > > > -----Message d'origine----- > > De : Stephen Crawley [mailto:crawley@dstc.edu.au] > > Envoye : vendredi 6 avril 2001 07:43 > > A : mariano.belaunde@rd.francetelecom.fr > > Cc : mof-rtf@omg.org > > Objet : Re: Issue 4231: predefined tag for marking > attributes to act as > > qualifier > > > > Mariano, > > > > The RTF members at today's teleconference agreed with your > suggestion in > > principle. Can you supply a suitable definition of what it > means for an > > Attribute to "act as a qualifier". We will consider your > definition, and > > if it is appropriate, we will define a new TagId. > > > > Bear in mind that it is not the intention of the RTF that > this new TagId > > should modify the IDL mapping generated by the IDL mapping or the > > behaviour of meta-objects as prescribed by the MOF spec. > Any use of the > > new TagId would be beyond the scope of the MOF spec. > > > > Thanks, > > > > -- Steve > > -- > Best Regards, > - Peter > > Peter Denno > National Institute of Standards and Technology, > Manufacturing System Integration Division, > 100 Bureau Drive, Mail Stop 8260 Tel: +1 301-975-3595 > Gaithersburg, MD, USA 20899-8260 FAX: +1 301-975-4694 > Date: Mon, 09 Apr 2001 16:07:43 -0400 From: Peter Denno X-Mailer: Mozilla 4.73 [en] (WinNT; U) X-Accept-Language: en MIME-Version: 1.0 To: Stephen Brodsky CC: "Iyengar, Sridhar" , BELAUNDE Mariano FTRD/DTL/LAN , "'Stephen Crawley'" , mof-rtf Subject: Re: Issue 4231: predefined tag for marking attributes to act as qualifier References: Content-Transfer-Encoding: 8bit Content-Type: text/plain; charset=iso-8859-1 X-UIDL: UGpd9G&b!!mh>e9n[,!! I agree that this tag sounds pretty heavy. OTOH, things like the CWM model their relational table object with a class object. This would make that modeling choice more credible. It would be nice for mapping to certain information modeling languages too (e.g. ISO 10303-11, aka Express). Stephen Brodsky wrote: > This is becoming a very heavy tag. What we have in our implementations is > an abstract Key class that represents an index of certain properties (ie > qualifiers) over a set of objects. If MOF includes indexing capabilities, > they should be modeled. The next question is "if". > > Thanks, > > -Steve > > Stephen A. Brodsky, Ph.D. > Software Architect > Notes Address: Stephen Brodsky/Santa Teresa/IBM@IBMUS > Internet Address: sbrodsky@us.ibm.com > Phone: 408.463.5659 > > "Iyengar, Sridhar" on 04/09/2001 10:27:35 AM > > To: BELAUNDE Mariano FTRD/DTL/LAN , > "'Peter Denno'" , "'Stephen Crawley'" > > cc: mof-rtf > Subject: RE: Issue 4231: predefined tag for marking attributes to act as > qualifier > > See [Sri] > > ------------------------------------------------------------------------ > ---- > Sridhar Iyengar > Unisys Fellow, Director of Advanced Technology > 25725, Jeronimo Road > Mission Viejo, CA 92691 > > E-mail : Sridhar.iyengar2@unisys.com > Phone : 949-380-5692 > Fax : 949-380-6632 > ------------------------------------------------------------------------ > -------------- > > > -----Original Message----- > > From: BELAUNDE Mariano FTRD/DTL/LAN > > [mailto:mariano.belaunde@rd.francetelecom.fr] > > Sent: Monday, April 09, 2001 4:16 AM > > To: 'Peter Denno'; 'Stephen Crawley' > > Cc: mof-rtf > > Subject: RE: Issue 4231: predefined tag for marking > > attributes to act as > > qualifier > > > > > > I agree that the tag can apply to references as well as to attributes. > > Taking in account this, the revised meaning definition could be: > > > > ActAsQualifier > > tag id: "org.omg.MOF:act_as_qualifier" > > attaches to: Model::Attribute, Model::Reference > > [Sri] You can attach it to Model::StructuralFeature which is a superclass > of > reference and Attribute. > > Also if this is being used as a composite key, the ordering of the > features becomes important > > qualifier (a,b,c) is NOT the same as (b,c,a) for purposes of > searching/sorting etc. > > > values: a Boolean > > meaning: The combination of values assigned to the attributes and > > references > > that are marked with this tag can be used to provide > > an identity to > > the class > > instance owning the attributes and the references. > > idl generation: no impact > > > > Peter, do not hesitate to propose another definition if you > > feel that this > > one is not > > clear enough or if an extra explanation is needed. > > > > Note that the current proposal is not formulated in terms of > > any uniqueness > > constraint > > in order not to restrict the usage of this tag. Otherwise we > > will need to > > know what > > is the implied context (what are the set of instances that > > need to conform > > to the > > constraint). The context to use may depend on the naming > > facility provided > > by a tool > > (and it does not necessarily depends on an existing composite > > association > > defined in the > > metamodel). > > > > In the case we really want a way to express that in a > > composite association > > all > > the referenced parts need to have distinct combinations of > > values for all > > the > > slots (attributes/references) "acting as qualifiers", I > > believe that this > > can be > > achieved by adding an extra 'isQualified' tag to the > > composite association. > > Of course, if you agree on the latter suggestion to take into > > account the > > Peter's > > "uniqueness constraint", we will also need to provide a > > meaning definition > > for the 'isQualified' tag. > > > > Mariano > > > > -----Message d'origine----- > > De : Peter Denno [mailto:peter.denno@nist.gov] > > Envoy : vendredi 6 avril 2001 20:00 > > : BELAUNDE Mariano FTRD/DTL/LAN > > Cc : 'Stephen Crawley'; mof-rtf > > Objet : Re: Issue 4231: predefined tag for marking attributes > > to act as > > qualifier > > > > > > Hi, > > > > Mariano's solution is fine where a single attribute provides a unique > > identifier, but maybe we could also address situations where > > a combination > > of > > attributes provide a unique identifier (i.e. in relational terms, a > > 'concatenated' or 'compound' key) The Express information > > modeling language > > (ISO > > 10303-11), which I hope could be supported, has something > > similar called a > > uniqueness constraint. It states that no two instances shall > > have the same > > combination of values for the attributes composing the constraint. > > > > > > BELAUNDE Mariano FTRD/DTL/LAN wrote: > > > > > Hi Steve, > > > > > > >> The RTF members at today's teleconference agreed with > > your suggestion > > > >> in principle. > > > Good decision. This gives the chance for two distinct > > implementors to > > share > > > the same convention for capturing this information. > > > Here's the proposed meaning definition: > > > > > > ActAsQualifier > > > tag id: "org.omg.MOF:act_as_qualifier" > > > attaches to: Model::Attribute > > > values: a Boolean > > > meaning: The values assigned to the attributes that are > > marked with > > this > > > tag > > > can be used to provide an identity to the > > class instance > > owning > > > the > > > attributes. > > > idl generation: no impact > > > > > > Regards, > > > > > > Mariano > > > > > > -----Message d'origine----- > > > De : Stephen Crawley [mailto:crawley@dstc.edu.au] > > > Envoye : vendredi 6 avril 2001 07:43 > > > A : mariano.belaunde@rd.francetelecom.fr > > > Cc : mof-rtf@omg.org > > > Objet : Re: Issue 4231: predefined tag for marking > > attributes to act as > > > qualifier > > > > > > Mariano, > > > > > > The RTF members at today's teleconference agreed with your > > suggestion in > > > principle. Can you supply a suitable definition of what it > > means for an > > > Attribute to "act as a qualifier". We will consider your > > definition, and > > > if it is appropriate, we will define a new TagId. > > > > > > Bear in mind that it is not the intention of the RTF that > > this new TagId > > > should modify the IDL mapping generated by the IDL mapping or the > > > behaviour of meta-objects as prescribed by the MOF spec. > > Any use of the > > > new TagId would be beyond the scope of the MOF spec. > > > > > > Thanks, > > > > > > -- Steve > > > > -- > > Best Regards, > > - Peter > > > > Peter Denno > > National Institute of Standards and Technology, > > Manufacturing System Integration Division, > > 100 Bureau Drive, Mail Stop 8260 Tel: +1 301-975-3595 > > Gaithersburg, MD, USA 20899-8260 FAX: +1 301-975-4694 > > -- Best Regards, - Peter Peter Denno National Institute of Standards and Technology, Manufacturing System Integration Division, 100 Bureau Drive, Mail Stop 8260 Tel: +1 301-975-3595 Gaithersburg, MD, USA 20899-8260 FAX: +1 301-975-4694 From: Martin Matula To: "'Peter Denno'" , BELAUNDE Mariano FTRD/DTL/LAN Cc: "'Stephen Crawley'" , mof-rtf Subject: RE: Issue 4231: predefined tag for marking attributes to act as qualifier Date: Tue, 10 Apr 2001 11:28:19 +0200 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain; charset="iso-8859-1" X-UIDL: l47!!Wabd9 In order to address compound qualifiers, I suggest that the > tag should be one > that attaches to classes and has as a value a set of names > referring to > attributes or references. > > Something like this: > > ActAsQualifier: > TAG ID: "omg.org.mof:act_as_qualifier > ATTACHES TO: Model::Class > VALUES: set [1..*] of NameType > MEANING: The values of the set refer to attributes and > references of the > class to which the tag is attached. The combination of the > values of these > attributes and references of a given instance serve as a qualifier > or > identifier to the instance. If I may step into this discussion, I would like to point out that any tag can be attached to 1 or MORE ModelElements. In addition, the AttachesTo association is ordered. That's why I prefer tag of type boolean attached to all the references and attributes within a class that serve as a qualifier (in a proper order - as AttachesTo is ordered) than having this kind of heavy tag with some kind of duplicate mechanism for AttachesTo association. So I rather agree with Mariano. Tag should be: ActAsQualifier: TAG ID: omg.org.mof:act_as_qualifier ATTACHES TO: Model::StructuralFeature (one or more in a proper order) VALUES: true/false MEANING: StructuralFeatures which this tag is attached to serve as an indentifier of the parent class instances. Regards, Martin Date: Tue, 10 Apr 2001 11:48:32 -0400 From: Peter Denno X-Mailer: Mozilla 4.73 [en] (WinNT; U) X-Accept-Language: en MIME-Version: 1.0 To: Martin Matula CC: BELAUNDE Mariano FTRD/DTL/LAN , "'Stephen Crawley'" , mof-rtf Subject: Re: Issue 4231: predefined tag for marking attributes to act as qualifier References: <10F82FC788E0D311AB820060083EAE1E02290B47@mail.netbeans.com> Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=us-ascii X-UIDL: \O#e9Vdg!!?^Ie9LT/!! Hi Martin, I think your solution would be a better one. It would be 'lightweight'. BUT I think the ordering is on the wrong side of the AttachesTo association for it to work. If I understand this correctly, the tags of a ModelElement are ordered. The ModelElements of a tag are not. I think a tag needs to order the ModelElements of a 'compound qualifier' in order for qualification to work. Is this correct? So I have a more general question for the RTF crew: Why is the ordering on the tags and not on the ModelElements? At the moment I can't think of any use for ordering on tags. Note also that ModelElements do not have a Reference to their tags. Another note on your approach. The Values attribute (having values true/false) doesn't really serve any purpose here, correct? Martin Matula wrote: > [Peter's text removed. It referred to an older description of the tag] > > If I may step into this discussion, I would like to point out that any tag > can be attached to 1 or MORE ModelElements. In addition, the AttachesTo > association is ordered. That's why I prefer tag of type boolean attached to > all the references and attributes within a class that serve as a qualifier > (in a proper order - as AttachesTo is ordered) than having this kind of > heavy tag with some kind of duplicate mechanism for AttachesTo association. > > So I rather agree with Mariano. Tag should be: > > ActAsQualifier: > TAG ID: omg.org.mof:act_as_qualifier > ATTACHES TO: Model::StructuralFeature (one or more in a proper > order) > VALUES: true/false > MEANING: StructuralFeatures which this tag is attached to serve as > an indentifier of the parent class instances. > > Regards, > Martin -- Best Regards, - Peter Peter Denno National Institute of Standards and Technology, Manufacturing System Integration Division, 100 Bureau Drive, Mail Stop 8260 Tel: +1 301-975-3595 Gaithersburg, MD, USA 20899-8260 FAX: +1 301-975-4694 From: Martin Matula To: "'Peter Denno'" Cc: mof-rtf Subject: RE: Issue 4231: predefined tag for marking attributes to act as qualifier Date: Tue, 10 Apr 2001 18:36:43 +0200 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain X-UIDL: ]NB!!UhT!!LIn!!;Lkd9 > The ModelElements of a tag are not. I think a tag needs to order the > ModelElements of a 'compound qualifier' in order for > qualification to work. Is > this correct? Oops! You are right. That's bad. I would create a new issue, however, this would affect the IDL interfaces, which I guess we don't want to change in 1.4 :( so this will be probably defered. > ordering on tags. Note also that ModelElements do not have a > Reference to their > tags. Reference for tags is not needed. > Another note on your approach. The Values attribute (having > values true/false) > doesn't really serve any purpose here, correct? Correct :). I doubt that someone would create this tag with value set to "false"... Martin From: Martin Matula To: "'Peter Denno'" Cc: mof-rtf Subject: RE: Issue 4231: predefined tag for marking attributes to act as qualifier Date: Wed, 11 Apr 2001 16:39:28 +0200 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain X-UIDL: K0D!!HNB!!W?+e9;gQd9 One more idea :) > Another note on your approach. The Values attribute (having > values true/false) > doesn't really serve any purpose here, correct? I think it does make sense to give each qualifier a name (e.g. in case > you wanted to have more than one qualifier in a class). So what about > having Value of this tag contain the name of the unique identifier defined by > this tag? So the tag could look like: ActAsQualifier: TAG ID: omg.org.mof:act_as_qualifier ATTACHES TO: Model::StructuralFeature VALUES: (String) name of this qualifier MEANING: StructuralFeatures which this tag is attached to > serve as an indentifier of the parent class instances. Martin Date: Wed, 11 Apr 2001 15:32:20 -0400 From: Peter Denno X-Mailer: Mozilla 4.73 [en] (WinNT; U) X-Accept-Language: en MIME-Version: 1.0 To: Martin Matula CC: mof-rtf Subject: Re: Issue 4231: predefined tag for marking attributes to act as qualifier References: <10F82FC788E0D311AB820060083EAE1E02290B53@mail.netbeans.com> Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=us-ascii X-UIDL: M:Ie920d!!"\L!!-HQd9 Martin, I think a string value in Values makes sense. In fact, it may be necessary to signifiy to what parent class this is a qualifier for, since the attributes may be inherited. For those who aren't following this thread too closely, the solution Martin offers requires that the model_element end of the AttachesTo association is ordered. Martin wrote an email regarding this issue earlier today. Martin Matula wrote: > One more idea :) > > > Another note on your approach. The Values attribute (having > > values true/false) > > doesn't really serve any purpose here, correct? > I think it does make sense to give each qualifier a name (e.g. in >case you > wanted to have more than one qualifier in a class). So what about >having > Value of this tag contain the name of the unique identifier defined >by this > tag? > So the tag could look like: > ActAsQualifier: > TAG ID: omg.org.mof:act_as_qualifier > ATTACHES TO: Model::StructuralFeature > VALUES: (String) name of this qualifier > MEANING: StructuralFeatures which this tag is attached to >serve as > an indentifier of the parent class instances. > > Martin -- Best Regards, - Peter Peter Denno National Institute of Standards and Technology, Manufacturing System Integration Division, 100 Bureau Drive, Mail Stop 8260 Tel: +1 301-975-3595 Gaithersburg, MD, USA 20899-8260 FAX: +1 301-975-4694 From: Martin Matula To: "'Peter Denno'" Cc: mof-rtf , "Stephen Crawley (E-mail)" Subject: RE: Issue 4231: predefined tag for marking attributes to act as qualifier Date: Thu, 12 Apr 2001 11:09:22 +0200 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain X-UIDL: \:`d9c%\d9@-?e9b1'"! Peter, as you can see from the Steve's reply to the "AttachesTo" issue, it won't be possible to change the modelElement end of this association to ordered. Why is it needed to explicitly specify order of attributes comprising the qualifier? Can you give some examples? I don't see a reason why we need ordering here. Martin > -----Original Message----- > From: Peter Denno [mailto:peter.denno@nist.gov] > Sent: Wednesday, April 11, 2001 9:32 PM > To: Martin Matula > Cc: mof-rtf > Subject: Re: Issue 4231: predefined tag for marking > attributes to act as > qualifier > > > Martin, > > I think a string value in Values makes sense. In fact, it may > be necessary to > signifiy to what parent class this is a qualifier for, since > the attributes may > be inherited. > > For those who aren't following this thread too closely, the > solution Martin > offers requires that the model_element end of the AttachesTo > association is > ordered. Martin wrote an email regarding this issue earlier today. > > Martin Matula wrote: > > > One more idea :) > > > > > Another note on your approach. The Values attribute (having > > > values true/false) > > > doesn't really serve any purpose here, correct? > > I think it does make sense to give each qualifier a name > (e.g. in case you > > wanted to have more than one qualifier in a class). So what > about having > > Value of this tag contain the name of the unique identifier > defined by this > > tag? > > So the tag could look like: > > ActAsQualifier: > > TAG ID: omg.org.mof:act_as_qualifier > > ATTACHES TO: Model::StructuralFeature > > VALUES: (String) name of this qualifier > > MEANING: StructuralFeatures which this tag is > attached to serve as > > an indentifier of the parent class instances. > > > > Martin > > -- > Best Regards, > - Peter > > Peter Denno > National Institute of Standards and Technology, > Manufacturing System Integration Division, > 100 Bureau Drive, Mail Stop 8260 Tel: +1 301-975-3595 > Gaithersburg, MD, USA 20899-8260 FAX: +1 301-975-4694 > > Date: Wed, 13 Jun 2001 12:31:05 -0400 From: Peter Denno X-Mailer: Mozilla 4.73 [en] (WinNT; U) X-Accept-Language: en MIME-Version: 1.0 To: mof-rtf Subject: Summary and recommendation for issue 4231 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=us-ascii X-UIDL: X0Je9L$?e9>PQe9:Dkd9 Here is a summary and recommendation for the resolution of issue 4231. There were three approaches discussed. OPTIONS: 1) Go with Mariano's solution and forego the compound key. 2) Either allow ordering on both sides of an association or switch the ordering (ordered ModelElements of a tag, not tags of a ModelElement) and Go with Martin's solution 3) Allow tags to have non-primitive type values and go with Peter's solution. RECOMMENDATION: Go with Mariano's original solution for 1.4. But long term, if we are going to continue to shun ordering on both sides of an association, we ought to say why. Must the MOF map neatly to a relational DB? ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; DETAILS (from various emails): MARIANO'S SOLUTION: ActAsQualifier TAG ID: "org.omg.MOF:act_as_qualifier" ATTACHES TO: Model::Attribute VALUES: a Boolean MEANING: The values assigned to the attributes that are marked with this tag can be used to provide an identity to the class instance owning the attributes. Mariano's solution is fine where a single attribute provides a unique identifier, but maybe we could also address situations where a combination of attributes provide a unique identifier (i.e. in relational terms, a 'concatenated' or 'compound' key) The Express information modeling language, which I hope could be supported, has something similar called a uniqueness constraint which states that no two instances shall have the same combination of values for the attributes composing the constraint. MARTIN'S SOLUTION: ActAsQualifier: TAG ID: omg.org.mof:act_as_qualifier ATTACHES TO: Model::StructuralFeature (one or more in a proper order) VALUES: true/false MEANING: StructuralFeatures which this tag is attached to serve as an indentifier of the parent class instances. Martin's solution would be best, EXCEPT I think the ordering is on the wrong side of the AttachesTo association for it to work. If I understand this correctly, the tags of a ModelElement are ordered. The ModelElements of a tag are not. I think a tag needs to order the ModelElements of a 'compound qualifier' in order for qualification to work. Is this correct? So I have a more general question for the RTF crew: Why is the ordering on the tags and not on the ModelElements? At the moment I can't think of any use for ordering on tags. Note also that ModelElements do not have a Reference to their tags. PETER'S SOLUTION: ActAsQualifier: TAG ID: "omg.org.mof:act_as_qualifier ATTACHES TO: Model::Class VALUES: list [1..*] of StructuralFeature MEANING: The values of the list refer to attributes and references of the class to which the tag is attached. The combination of the values of these attributes and references of a given instance serve as a qualifier or identifier to the instance. It is "heavy" in the sense that the values are not primitive types. -- Best Regards, - Peter Peter Denno National Institute of Standards and Technology, Manufacturing System Integration Division, 100 Bureau Drive, Mail Stop 8260 Tel: +1 301-975-3595 Gaithersburg, MD, USA 20899-8260 FAX: +1 301-975-4694 X-Mailer: exmh version 2.2 06/23/2000 with nmh-1.0.4 To: Peter Denno cc: mof-rtf , crawley@dstc.edu.au Subject: Re: Summary and recommendation for issue 4231 In-Reply-To: Message from Peter Denno of "Wed, 13 Jun 2001 12:31:05 -0400." <3B279549.565597E7@nist.gov> Mime-Version: 1.0 Date: Thu, 14 Jun 2001 10:00:44 +1000 From: Stephen Crawley X-Scanned-By: MIMEDefang 1.0 (http://www.roaringpenguin.com/mimedefang/) Content-Type: text/plain; charset=us-ascii X-UIDL: #KHe9L2'e9kF/e913&e9 Status: RO > RECOMMENDATION: Go with Mariano's original solution for 1.4. But long > term, if we are going to continue to shun ordering on both sides of an > association, we ought to say why. Must the MOF map neatly to a > relational DB? For the record, the main reasons we forbade Associations ordered at both end were: * a doubly ordered Association is difficult to visualise and explain * APIs for updating doubly ordered Associations would be incredibly complex and unwieldy MOF singly ordered Associations do NOT map neatly onto RDMBS tables, because the rows in a classical RDBMs table are not ordered. -- Steve X-Mailer: exmh version 2.2 06/23/2000 with nmh-1.0.4 To: mof-rtf@omg.org Subject: Re: Issue 4231: predefined tag for marking attributes to act as qualifier in classifier Mime-Version: 1.0 Date: Thu, 05 Jul 2001 23:22:19 +1000 From: Stephen Crawley X-Scanned-By: MIMEDefang 1.0 (http://www.roaringpenguin.com/mimedefang/) Content-Type: text/plain; charset=us-ascii X-UIDL: ^E\!!5>X!!2`md9jj^!! > Source: France Telecom R&D > (Mr. Mariano Belaunde, mariano.belaunde@rd.francetelecom.fr) > Nature: Uncategorized Issue > Severity: > Summary: Need a predefined tag for marking attributes that will 'act > as a > qualifier' in a classifier. In a metamodelling tool it may be useful > to derive automatically an object identifier from one or more > relevant > attributes of the classifier (for instance a 'name' attribute). > > Note: The attribute 'acting as a qualifier' is owned by the > classifier not > by an association. > > Suggestion: pre-define a Tag named 'actAsQualifier : Boolean' > applicable > to any attribute of a classifier. Proposed Resolution: Create a new section of Chapter 3 called Technology Neutral Tags with a suitable description. Add the "org.omg.MOF:act_as_qualifier" tag as described below. Proposed Text: ActAsQualifier TAG ID: "org.omg.MOF:act_as_qualifier" ATTACHES TO: Model::Attribute VALUES: a Boolean MEANING: The values assigned to the attributes that are marked with this tag can be used to provide an identity to the class instance owning the attributes. From: "Baisley, Donald E" To: mof-rtf@omg.org Subject: RE: Issue 4231: predefined tag for marking attributes to act as q ualifier in classifier Date: Thu, 5 Jul 2001 19:07:20 -0500 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain; charset="iso-8859-1" X-UIDL: 55N!!Z2U!!d?o!!YO3!! Some questions: 1. If two attributes of a class have org.omg.MOF:act_as_qualifier set to true, does either attribute by itself uniquely identify an object, or do the two attributes together uniquely identify an object? 2. Why can't this property apply to references as well as attributes? 3. Why not model this properly? Why mess around with a tag? Regards, Don -----Original Message----- From: Stephen Crawley [mailto:crawley@dstc.edu.au] Sent: Thursday, July 05, 2001 6:22 AM To: mof-rtf@omg.org Subject: Re: Issue 4231: predefined tag for marking attributes to act as qualifier in classifier > Source: France Telecom R&D > (Mr. Mariano Belaunde, mariano.belaunde@rd.francetelecom.fr) > Nature: Uncategorized Issue > Severity: > Summary: Need a predefined tag for marking attributes that will 'act > as a > qualifier' in a classifier. In a metamodelling tool it may be useful > to derive automatically an object identifier from one or more > relevant > attributes of the classifier (for instance a 'name' attribute). > > Note: The attribute 'acting as a qualifier' is owned by the > classifier not > by an association. > > Suggestion: pre-define a Tag named 'actAsQualifier : Boolean' > applicable > to any attribute of a classifier. Proposed Resolution: Create a new section of Chapter 3 called Technology Neutral Tags with a suitable description. Add the "org.omg.MOF:act_as_qualifier" tag as described below. Proposed Text: ActAsQualifier TAG ID: "org.omg.MOF:act_as_qualifier" ATTACHES TO: Model::Attribute VALUES: a Boolean MEANING: The values assigned to the attributes that are marked with this tag can be used to provide an identity to the class instance owning the attributes. From: "Pete Rivett" To: Subject: Adaptive votes and 4231 Date: Sun, 8 Jul 2001 06:27:33 +0100 Message-ID: <00aa01c10773$c45b0f20$4f7c7bd5@CHIMAY> MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook CWS, Build 9.0.2416 (9.0.2911.0) In-Reply-To: <200107052253.f65Mr0400340@piglet.dstc.edu.au> Importance: Normal X-MimeOLE: Produced By Microsoft MimeOLE V5.00.3018.1300 Content-Type: text/plain; charset="us-ascii" X-UIDL: :(_d9GC9e9["2e9^EM!! Adaptive votes YES to Blocks 2 and 3 for all issues except 4231 (act_as_qualifier) Adaptive votes NO on the current form of resolution to 4231 The reason for 'NO' on 4231 is that I think it's underspecified (e.g. as per Don's email, is one of a set of attributes sufficient? is there any notion of ordering the attributes to form a single string identifier which would seem to be the original requirement?) and the fact that the concept of 'identifier' itself does not seem well-defined (e.g. what can one do with it or use it for? Does it have any significance (e.g. as element name) in an XMI file? Is there any requirement for uniqueness in some context? (I'd say no)). Pete Pete Rivett (pete.rivett@adaptive.com) Chief Technology Officer, Adaptive Ltd Dean Park House, 8-10 Dean Park Crescent, Bournemouth, BH1 1HL, UK Tel: +44 (0)1202 449419 Fax: +44 (0)1202 449448 http://www.adaptive.com > -----Original Message----- > From: Stephen Crawley [mailto:crawley@dstc.edu.au] > Sent: 05 July 2001 23:53 > To: mof-rtf@omg.org > Subject: Final MOF RTF Issue resolutions ready for votes. > > > > Folks, > > Following today's teleconference, I've added the third (and > final) Block > of issue resolutions to the Votes page at: > > http://www.dstc.edu.au/Research/Projects/MOF/rtf1.4/votes.html > > Officially, I can't call the vote or set the deadline -- > Sridhar must do > this according to the P&P. > > Unofficially, since Sridhar's teleconference agenda item implies we > should do this: > > * please consider the vote to have been called, and > * please consider the deadline to be the MOF RTF meeting at the > Danvers OMG meeting (Tuesday 10th July). > > [If Sridhar decides more time is needed, he will set the > official voting > deadline accordingly.] > > PLEASE EMAIL YOUR BLOCK 3 VOTES ... stating clearly which > issues / corrigenda > you are voting for. > > DSTC and Sun voted YES to all issues and corrigenda in Block 3 at the > last teleconference. I have also received YES votes from Sprint, > Telelogic and Oracle to issues 3444 and 4384. > > -- Steve > > 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. X-Mailer: exmh version 2.2 06/23/2000 with nmh-1.0.4 To: mof-rtf@omg.org Subject: Re: Issue 4231 Mime-Version: 1.0 Date: Sun, 08 Jul 2001 23:47:22 +1000 From: Stephen Crawley X-Scanned-By: MIMEDefang 1.0 (http://www.roaringpenguin.com/mimedefang/) Content-Type: text/plain; charset=us-ascii X-UIDL: "_$e94