Issue 11068: Section: 14 Interactions: Lifeline representing an actor (uml2-rtf) Source: oose Innovative Informatik GmbH (Mr. Tim Weilkiens, tim.weilkiens(at)oose.de) Nature: Enhancement Severity: Significant Summary: It is common usage to model a lifeline in a interaction that represents an actor. I can't see how that could be done formally correct. A lifeline represents a connectable element, e.g. a property. It is not allowed to define a property that is typed by an actor. Resolution: Revised Text: Actions taken: May 25, 2007: received issue Discussion: End of Annotations:===== m: webmaster@omg.org Date: 25 May 2007 05:25:23 -0400 To: Subject: Issue/Bug Report -------------------------------------------------------------------------------- Name: Tim Weilkiens Company: oose Innovative Informatik GmbH mailFrom: tim.weilkiens@oose.de Notification: Yes Specification: UML Superstructure Section: 14 Interactions: Lifeline representing an actor FormalNumber: formal/2007-02-03 Version: 2.1.1 RevisionDate: 02/03/07 Page: 457ff Nature: Enhancement Severity: Significant HTTP User Agent: Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 5.1; .NET CLR 2.0.50727; .NET CLR 1.1.4322) Description It is common usage to model a lifeline in a interaction that represents an actor. I can't see how that could be done formally correct. A lifeline represents a connectable element, e.g. a property. It is not allowed to define a property that is typed by an actor. Subject: [Issue 11068] Lifeline representing an actor From: Florian Schneider Date: Fri, 28 May 2010 12:39:10 +0200 To: uml2-rtf@omg.org X-Mailer: Apple Mail (2.1078) X-Virus-Scanned: ClamAV using ClamSMTP Dear Ladies and Gentlemen, I have a comment regarding issue 11068, reported by Mr. Tim Weilkiens. Mr. Weilkiens claims the following: "A lifeline represents a connectable element, e.g. a property". I believe that this is not correct because a ConnectableElement is a subclass of TypedElement so it can have any type. I don't see why it is equivalent to a property. Concerning this issue, I am interested in a clarification whether there are constraints on what Type subclasses may be placed on a Sequence Diagram (Exemplary questions: May an Actor be placed on a sequence diagram? May a use case be placed on a sequence diagram? May a package be placed on a sequence diagram?). With best regards, Florian Schneider -------------------------------- Dipl.-Inf. Univ. Florian Schneider Research Assistant Chair for Applied Software Engineering Technische UniversitäMü Tel.: +49 (89) 289-18233 Fax.: +49 (89) 289-18207 Web http://www1.in.tum.de/FlorianSchneider Address: Lehrstuhl füewandte Softwaretechnik, I1 Technische UniversitäMü Institut füormatik Boltzmannstraß 3 D-85748 Garching smime.p7s Date: Fri, 28 May 2010 11:52:20 +0100 From: Dave Hawkins User-Agent: Thunderbird 2.0.0.21 (X11/20090302) To: Florian Schneider CC: uml2-rtf@omg.org Subject: Re: [Issue 11068] Lifeline representing an actor X-Auth-Type: Internal IP X-Source-IP: acsinet15.oracle.com [141.146.126.227] X-CT-RefId: str=0001.0A090207.4BFFA0A5.003A:SCFMA922111,ss=1,fgs=0 The quoted sentence doesn't say it's equivalent, it's simply giving an example. Property is a subclass of ConnectableElement and therefore it is valid for Lifeline::represents to be a Property. Cheers, Dave Florian Schneider wrote: Dear Ladies and Gentlemen, I have a comment regarding issue 11068, reported by Mr. Tim Weilkiens. Mr. Weilkiens claims the following: "A lifeline represents a connectable element, e.g. a property". I believe that this is not correct because a ConnectableElement is a subclass of TypedElement so it can have any type. I don't see why it is equivalent to a property. Concerning this issue, I am interested in a clarification whether there are constraints on what Type subclasses may be placed on a Sequence Diagram (Exemplary questions: May an Actor be placed on a sequence diagram? May a use case be placed on a sequence diagram? May a package be placed on a sequence diagram?). With best regards, Florian Schneider -------------------------------- Dipl.-Inf. Univ. Florian Schneider Research Assistant Chair for Applied Software Engineering Technische UniversitäMü Tel.: +49 (89) 289-18233 Fax.: +49 (89) 289-18207 Web http://www1.in.tum.de/FlorianSchneider Address: Lehrstuhl füewandte Softwaretechnik, I1 Technische UniversitäMü Institut füormatik Boltzmannstraß 3 D-85748 Garching -- Dave Hawkins | Principal Software Engineer | +44 118 924 0022 Oracle JDeveloper Development Oracle Corporation UK Ltd is a company incorporated in England & Wales. Company Reg. No. 1782505. Reg. office: Oracle Parkway, Thames Valley Park, Reading RG6 1RA. From: Østein Haugen To: Dave Hawkins , Florian Schneider CC: "uml2-rtf@omg.org" , Østein Haugen Date: Fri, 28 May 2010 13:59:49 +0200 Subject: RE: [Issue 11068] Lifeline representing an actor Thread-Topic: [Issue 11068] Lifeline representing an actor Thread-Index: Acr+VE/9coUr3oyLSYO10b/llwN/GAACCiaQ Accept-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: acceptlanguage: en-US X-MIME-Autoconverted: from quoted-printable to 8bit by amethyst.omg.org id o4SBm9Lc000892 I agree with Dave's point, but let me provide the motivation for why a Lifeline respresents a ConnectableElement and not merely a Property. The reason is that there was a need to let sequence diagrams describe communications with Parameters and Parameter is not a specialization of Property. About the Type subclasses that you are asking about, I get a little confused, since Actor is not Property, and neither are UseCase nor Package, so these things cannot be represented by Lifelines. /Oystein ---- Dr. Oystein Haugen Senior Researcher SINTEF -----Original Message----- From: Dave Hawkins [mailto:dave.hawkins@oracle.com] Sent: 28. mai 2010 12:52 To: Florian Schneider Cc: uml2-rtf@omg.org Subject: Re: [Issue 11068] Lifeline representing an actor The quoted sentence doesn't say it's equivalent, it's simply giving an example. Property is a subclass of ConnectableElement and therefore it is valid for Lifeline::represents to be a Property. Cheers, Dave Florian Schneider wrote: > Dear Ladies and Gentlemen, > > I have a comment regarding issue 11068, reported by Mr. Tim Weilkiens. > > Mr. Weilkiens claims the following: "A lifeline represents a connectable element, e.g. a property". I believe that this is not correct because a ConnectableElement is a subclass of TypedElement so it can have any type. I don't see why it is equivalent to a property. > > Concerning this issue, I am interested in a clarification whether there are constraints on what Type subclasses may be placed on a Sequence Diagram (Exemplary questions: May an Actor be placed on a sequence diagram? May a use case be placed on a sequence diagram? May a package be placed on a sequence diagram?). > > With best regards, > Florian Schneider > > > > -------------------------------- > Dipl.-Inf. Univ. Florian Schneider > > Research Assistant > Chair for Applied Software Engineering > Technische UniversitäMü > > Tel.: +49 (89) 289-18233 > Fax.: +49 (89) 289-18207 > Web http://www1.in.tum.de/FlorianSchneider > > Address: > Lehrstuhl füewandte Softwaretechnik, I1 > Technische UniversitäMü > Institut füormatik > Boltzmannstraß 3 > D-85748 Garching > > > > > -- Dave Hawkins | Principal Software Engineer | +44 118 924 0022 Oracle JDeveloper Development Oracle Corporation UK Ltd is a company incorporated in England & Wales. Company Reg. No. 1782505. Reg. office: Oracle Parkway, Thames Valley Park, Reading RG6 1RA. Cc: Dave Hawkins , "uml2-rtf@omg.org" From: Florian Schneider Subject: Re: [Issue 11068] Lifeline representing an actor Date: Fri, 28 May 2010 14:47:39 +0200 To: Østein Haugen X-Mailer: Apple Mail (2.1078) X-Virus-Scanned: ClamAV using ClamSMTP OK so maybe I got something wrong here. Here is how I understood the metamodel: A lifeline represents a ConnectableElement, which is a TypedElement. A TypedElement has a Type. Subclasses of Type are, among others, Classifier and BehavioredClassifier. Subclasses of Classifier are, among others, Class and Interface. Subclasses of BehavioredClassifier are, among others, Actor and UseCase. So as I understood it would be (syntactically) possible to place e.g. Classes, Interfaces, Actors and UseCases could be placed on a Sequence Diagram. Did I get that right up to here? Cheers, Florian On 28.05.2010, at 13:59, Østein Haugen wrote: > I agree with Dave's point, but let me provide the motivation for why a Lifeline respresents a ConnectableElement and not merely a Property. The reason is that there was a need to let sequence diagrams describe communications with Parameters and Parameter is not a specialization of Property. > > About the Type subclasses that you are asking about, I get a little confused, since Actor is not Property, and neither are UseCase nor Package, so these things cannot be represented by Lifelines. > > /Oystein > > > ---- > Dr. Oystein Haugen > Senior Researcher > SINTEF > > -----Original Message----- > From: Dave Hawkins [mailto:dave.hawkins@oracle.com] > Sent: 28. mai 2010 12:52 > To: Florian Schneider > Cc: uml2-rtf@omg.org > Subject: Re: [Issue 11068] Lifeline representing an actor > > The quoted sentence doesn't say it's equivalent, it's simply giving an example. Property > is a subclass of ConnectableElement and therefore it is valid for Lifeline::represents to > be a Property. > > Cheers, > > Dave > > Florian Schneider wrote: >> Dear Ladies and Gentlemen, >> >> I have a comment regarding issue 11068, reported by Mr. Tim Weilkiens. >> >> Mr. Weilkiens claims the following: "A lifeline represents a connectable element, e.g. a property". I believe that this is not correct because a ConnectableElement is a subclass of TypedElement so it can have any type. I don't see why it is equivalent to a property. >> >> Concerning this issue, I am interested in a clarification whether there are constraints on what Type subclasses may be placed on a Sequence Diagram (Exemplary questions: May an Actor be placed on a sequence diagram? May a use case be placed on a sequence diagram? May a package be placed on a sequence diagram?). >> >> With best regards, >> Florian Schneider >> >> >> >> -------------------------------- >> Dipl.-Inf. Univ. Florian Schneider >> >> Research Assistant >> Chair for Applied Software Engineering >> Technische UniversitäMü >> >> Tel.: +49 (89) 289-18233 >> Fax.: +49 (89) 289-18207 >> Web http://www1.in.tum.de/FlorianSchneider >> >> Address: >> Lehrstuhl füewandte Softwaretechnik, I1 >> Technische UniversitäMü >> Institut füormatik >> Boltzmannstraß 3 >> D-85748 Garching >> >> >> >> >> > > -- > Dave Hawkins | Principal Software Engineer | +44 118 924 0022 > Oracle JDeveloper Development > Oracle Corporation UK Ltd is a company incorporated in England & Wales. > Company Reg. No. 1782505. > Reg. office: Oracle Parkway, Thames Valley Park, Reading RG6 1RA. smime1.p7s From: Østein Haugen To: Florian Schneider CC: Dave Hawkins , "uml2-rtf@omg.org" , Østein Haugen Date: Fri, 28 May 2010 14:58:49 +0200 Subject: RE: [Issue 11068] Lifeline representing an actor Thread-Topic: [Issue 11068] Lifeline representing an actor Thread-Index: Acr+Y/eOpRW4wEgWRMKDUPWWToGI+gAAHYbw Accept-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: acceptlanguage: en-US X-MIME-Autoconverted: from quoted-printable to 8bit by amethyst.omg.org id o4SCl8BR005779 Florian Model Elements that themselves are Types (such as Actor) are not "placed" in a sequence diagram (there is no concrete syntax for that), but since Interaction is a Class, such Types can be included in Interaction through the model browser. But, I do not think this is what you are asking. I sense a slight ambiguity in what you are writing since you must distinguish clearly between the elements that _have_ a type, and those that _are_ a type. Actor _is_ a type, but you do not find diagrams where there are connectable elements having a specific Actor as type. If you found such a context, you could also describe its behavior through an Interaction and let that connectable element be represented by a Lifeline. (Actors are found in use case diagrams, but the elements there are Actors, not connectable elements.) /Oystein ---- Dr. Oystein Haugen Senior Researcher SINTEF -----Original Message----- From: Florian Schneider [mailto:schneidf@in.tum.de] Sent: 28. mai 2010 14:48 To: Østein Haugen Cc: Dave Hawkins; uml2-rtf@omg.org Subject: Re: [Issue 11068] Lifeline representing an actor OK so maybe I got something wrong here. Here is how I understood the metamodel: A lifeline represents a ConnectableElement, which is a TypedElement. A TypedElement has a Type. Subclasses of Type are, among others, Classifier and BehavioredClassifier. Subclasses of Classifier are, among others, Class and Interface. Subclasses of BehavioredClassifier are, among others, Actor and UseCase. So as I understood it would be (syntactically) possible to place e.g. Classes, Interfaces, Actors and UseCases could be placed on a Sequence Diagram. Did I get that right up to here? Cheers, Florian On 28.05.2010, at 13:59, Østein Haugen wrote: > I agree with Dave's point, but let me provide the motivation for why a Lifeline respresents a ConnectableElement and not merely a Property. The reason is that there was a need to let sequence diagrams describe communications with Parameters and Parameter is not a specialization of Property. > > About the Type subclasses that you are asking about, I get a little confused, since Actor is not Property, and neither are UseCase nor Package, so these things cannot be represented by Lifelines. > > /Oystein > > > ---- > Dr. Oystein Haugen > Senior Researcher > SINTEF > > -----Original Message----- > From: Dave Hawkins [mailto:dave.hawkins@oracle.com] > Sent: 28. mai 2010 12:52 > To: Florian Schneider > Cc: uml2-rtf@omg.org > Subject: Re: [Issue 11068] Lifeline representing an actor > > The quoted sentence doesn't say it's equivalent, it's simply giving an example. Property > is a subclass of ConnectableElement and therefore it is valid for Lifeline::represents to > be a Property. > > Cheers, > > Dave > > Florian Schneider wrote: >> Dear Ladies and Gentlemen, >> >> I have a comment regarding issue 11068, reported by Mr. Tim Weilkiens. >> >> Mr. Weilkiens claims the following: "A lifeline represents a connectable element, e.g. a property". I believe that this is not correct because a ConnectableElement is a subclass of TypedElement so it can have any type. I don't see why it is equivalent to a property. >> >> Concerning this issue, I am interested in a clarification whether there are constraints on what Type subclasses may be placed on a Sequence Diagram (Exemplary questions: May an Actor be placed on a sequence diagram? May a use case be placed on a sequence diagram? May a package be placed on a sequence diagram?). >> >> With best regards, >> Florian Schneider >> >> >> >> -------------------------------- >> Dipl.-Inf. Univ. Florian Schneider >> >> Research Assistant >> Chair for Applied Software Engineering >> Technische UniversitäMü >> >> Tel.: +49 (89) 289-18233 >> Fax.: +49 (89) 289-18207 >> Web http://www1.in.tum.de/FlorianSchneider >> >> Address: >> Lehrstuhl füewandte Softwaretechnik, I1 >> Technische UniversitäMü >> Institut füormatik >> Boltzmannstraß 3 >> D-85748 Garching >> >> >> >> >> > > -- > Dave Hawkins | Principal Software Engineer | +44 118 924 0022 > Oracle JDeveloper Development > Oracle Corporation UK Ltd is a company incorporated in England & Wales. > Company Reg. No. 1782505. > Reg. office: Oracle Parkway, Thames Valley Park, Reading RG6 1RA. Cc: Dave Hawkins , "uml2-rtf@omg.org" From: Florian Schneider Subject: Re: [Issue 11068] Lifeline representing an actor Date: Fri, 28 May 2010 15:20:18 +0200 To: Østein Haugen X-Mailer: Apple Mail (2.1078) X-Virus-Scanned: ClamAV using ClamSMTP I am sorry for any ambiguity! I agree that an instance of the class Actor is not in a sequence diagram (I also agree that it can be part of a use case diagram). But ConnectableElement, through its superclass TypedElement has an association (which is named 'type') to the class Type. So when talking about a "ConnectableElement of type Actor", I mean that the association named type of the ConnectableElement instance points to an instance of the class Actor. I found no constraint that would make that impossible. Let's say we have a Lifeline that represents a ConnectableElement of type Actor. How is that notationally reflected on the according sequence diagram? Cheers, Florian On 28.05.2010, at 14:58, Østein Haugen wrote: > Florian > Model Elements that themselves are Types (such as Actor) are not "placed" in a sequence diagram (there is no concrete syntax for that), but since Interaction is a Class, such Types can be included in Interaction through the model browser. > > But, I do not think this is what you are asking. I sense a slight ambiguity in what you are writing since you must distinguish clearly between the elements that _have_ a type, and those that _are_ a type. Actor _is_ a type, but you do not find diagrams where there are connectable elements having a specific Actor as type. If you found such a context, you could also describe its behavior through an Interaction and let that connectable element be represented by a Lifeline. > > (Actors are found in use case diagrams, but the elements there are Actors, not connectable elements.) > > /Oystein > > > ---- > Dr. Oystein Haugen > Senior Researcher > SINTEF > > -----Original Message----- > From: Florian Schneider [mailto:schneidf@in.tum.de] > Sent: 28. mai 2010 14:48 > To: Østein Haugen > Cc: Dave Hawkins; uml2-rtf@omg.org > Subject: Re: [Issue 11068] Lifeline representing an actor > > OK so maybe I got something wrong here. > > Here is how I understood the metamodel: A lifeline represents a ConnectableElement, which is a TypedElement. A TypedElement has a Type. Subclasses of Type are, among others, Classifier and BehavioredClassifier. Subclasses of Classifier are, among others, Class and Interface. Subclasses of BehavioredClassifier are, among others, Actor and UseCase. > So as I understood it would be (syntactically) possible to place e.g. Classes, Interfaces, Actors and UseCases could be placed on a Sequence Diagram. Did I get that right up to here? > > Cheers, > Florian > > On 28.05.2010, at 13:59, Østein Haugen wrote: > >> I agree with Dave's point, but let me provide the motivation for why a Lifeline respresents a ConnectableElement and not merely a Property. The reason is that there was a need to let sequence diagrams describe communications with Parameters and Parameter is not a specialization of Property. >> >> About the Type subclasses that you are asking about, I get a little confused, since Actor is not Property, and neither are UseCase nor Package, so these things cannot be represented by Lifelines. >> >> /Oystein >> >> >> ---- >> Dr. Oystein Haugen >> Senior Researcher >> SINTEF >> >> -----Original Message----- >> From: Dave Hawkins [mailto:dave.hawkins@oracle.com] >> Sent: 28. mai 2010 12:52 >> To: Florian Schneider >> Cc: uml2-rtf@omg.org >> Subject: Re: [Issue 11068] Lifeline representing an actor >> >> The quoted sentence doesn't say it's equivalent, it's simply giving an example. Property >> is a subclass of ConnectableElement and therefore it is valid for Lifeline::represents to >> be a Property. >> >> Cheers, >> >> Dave >> >> Florian Schneider wrote: >>> Dear Ladies and Gentlemen, >>> >>> I have a comment regarding issue 11068, reported by Mr. Tim Weilkiens. >>> >>> Mr. Weilkiens claims the following: "A lifeline represents a connectable element, e.g. a property". I believe that this is not correct because a ConnectableElement is a subclass of TypedElement so it can have any type. I don't see why it is equivalent to a property. >>> >>> Concerning this issue, I am interested in a clarification whether there are constraints on what Type subclasses may be placed on a Sequence Diagram (Exemplary questions: May an Actor be placed on a sequence diagram? May a use case be placed on a sequence diagram? May a package be placed on a sequence diagram?). >>> >>> With best regards, >>> Florian Schneider >>> >>> >>> >>> -------------------------------- >>> Dipl.-Inf. Univ. Florian Schneider >>> >>> Research Assistant >>> Chair for Applied Software Engineering >>> Technische UniversitäMü >>> >>> Tel.: +49 (89) 289-18233 >>> Fax.: +49 (89) 289-18207 >>> Web http://www1.in.tum.de/FlorianSchneider >>> >>> Address: >>> Lehrstuhl füewandte Softwaretechnik, I1 >>> Technische UniversitäMü >>> Institut füormatik >>> Boltzmannstraß 3 >>> D-85748 Garching >>> >>> >>> >>> >>> >> >> -- >> Dave Hawkins | Principal Software Engineer | +44 118 924 0022 >> Oracle JDeveloper Development >> Oracle Corporation UK Ltd is a company incorporated in England & Wales. >> Company Reg. No. 1782505. >> Reg. office: Oracle Parkway, Thames Valley Park, Reading RG6 1RA. > smime2.p7s From: Østein Haugen To: Florian Schneider CC: Dave Hawkins , "uml2-rtf@omg.org" , Østein Haugen Date: Fri, 28 May 2010 15:26:39 +0200 Subject: RE: [Issue 11068] Lifeline representing an actor Thread-Topic: [Issue 11068] Lifeline representing an actor Thread-Index: Acr+aIY/+AE1lGMhTWmgftqNYK/RLQAAGrWw Accept-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: acceptlanguage: en-US X-MIME-Autoconverted: from quoted-printable to 8bit by amethyst.omg.org id o4SDEsTV009275 Florian If you find such a connectable element, then a corresponding Lifeline to represent it looks just like any other lifeline with the head containing "name of connectable element : name of Actor" You cannot legally use the Actor symbol on that lifeline even though I know several tools have implemented this option probably because they are still hung up on UML1 stuff. /Oystein ---- Dr. Oystein Haugen Senior Researcher SINTEF -----Original Message----- From: Florian Schneider [mailto:schneidf@in.tum.de] Sent: 28. mai 2010 15:20 To: Østein Haugen Cc: Dave Hawkins; uml2-rtf@omg.org Subject: Re: [Issue 11068] Lifeline representing an actor I am sorry for any ambiguity! I agree that an instance of the class Actor is not in a sequence diagram (I also agree that it can be part of a use case diagram). But ConnectableElement, through its superclass TypedElement has an association (which is named 'type') to the class Type. So when talking about a "ConnectableElement of type Actor", I mean that the association named type of the ConnectableElement instance points to an instance of the class Actor. I found no constraint that would make that impossible. Let's say we have a Lifeline that represents a ConnectableElement of type Actor. How is that notationally reflected on the according sequence diagram? Cheers, Florian On 28.05.2010, at 14:58, Østein Haugen wrote: > Florian > Model Elements that themselves are Types (such as Actor) are not "placed" in a sequence diagram (there is no concrete syntax for that), but since Interaction is a Class, such Types can be included in Interaction through the model browser. > > But, I do not think this is what you are asking. I sense a slight ambiguity in what you are writing since you must distinguish clearly between the elements that _have_ a type, and those that _are_ a type. Actor _is_ a type, but you do not find diagrams where there are connectable elements having a specific Actor as type. If you found such a context, you could also describe its behavior through an Interaction and let that connectable element be represented by a Lifeline. > > (Actors are found in use case diagrams, but the elements there are Actors, not connectable elements.) > > /Oystein > > > ---- > Dr. Oystein Haugen > Senior Researcher > SINTEF > > -----Original Message----- > From: Florian Schneider [mailto:schneidf@in.tum.de] > Sent: 28. mai 2010 14:48 > To: Østein Haugen > Cc: Dave Hawkins; uml2-rtf@omg.org > Subject: Re: [Issue 11068] Lifeline representing an actor > > OK so maybe I got something wrong here. > > Here is how I understood the metamodel: A lifeline represents a ConnectableElement, which is a TypedElement. A TypedElement has a Type. Subclasses of Type are, among others, Classifier and BehavioredClassifier. Subclasses of Classifier are, among others, Class and Interface. Subclasses of BehavioredClassifier are, among others, Actor and UseCase. > So as I understood it would be (syntactically) possible to place e.g. Classes, Interfaces, Actors and UseCases could be placed on a Sequence Diagram. Did I get that right up to here? > > Cheers, > Florian > > On 28.05.2010, at 13:59, Østein Haugen wrote: > >> I agree with Dave's point, but let me provide the motivation for why a Lifeline respresents a ConnectableElement and not merely a Property. The reason is that there was a need to let sequence diagrams describe communications with Parameters and Parameter is not a specialization of Property. >> >> About the Type subclasses that you are asking about, I get a little confused, since Actor is not Property, and neither are UseCase nor Package, so these things cannot be represented by Lifelines. >> >> /Oystein >> >> >> ---- >> Dr. Oystein Haugen >> Senior Researcher >> SINTEF >> >> -----Original Message----- >> From: Dave Hawkins [mailto:dave.hawkins@oracle.com] >> Sent: 28. mai 2010 12:52 >> To: Florian Schneider >> Cc: uml2-rtf@omg.org >> Subject: Re: [Issue 11068] Lifeline representing an actor >> >> The quoted sentence doesn't say it's equivalent, it's simply giving an example. Property >> is a subclass of ConnectableElement and therefore it is valid for Lifeline::represents to >> be a Property. >> >> Cheers, >> >> Dave >> >> Florian Schneider wrote: >>> Dear Ladies and Gentlemen, >>> >>> I have a comment regarding issue 11068, reported by Mr. Tim Weilkiens. >>> >>> Mr. Weilkiens claims the following: "A lifeline represents a connectable element, e.g. a property". I believe that this is not correct because a ConnectableElement is a subclass of TypedElement so it can have any type. I don't see why it is equivalent to a property. >>> >>> Concerning this issue, I am interested in a clarification whether there are constraints on what Type subclasses may be placed on a Sequence Diagram (Exemplary questions: May an Actor be placed on a sequence diagram? May a use case be placed on a sequence diagram? May a package be placed on a sequence diagram?). >>> >>> With best regards, >>> Florian Schneider >>> >>> >>> >>> -------------------------------- >>> Dipl.-Inf. Univ. Florian Schneider >>> >>> Research Assistant >>> Chair for Applied Software Engineering >>> Technische UniversitäMü >>> >>> Tel.: +49 (89) 289-18233 >>> Fax.: +49 (89) 289-18207 >>> Web http://www1.in.tum.de/FlorianSchneider >>> >>> Address: >>> Lehrstuhl füewandte Softwaretechnik, I1 >>> Technische UniversitäMü >>> Institut füormatik >>> Boltzmannstraß 3 >>> D-85748 Garching >>> >>> >>> >>> >>> >> >> -- >> Dave Hawkins | Principal Software Engineer | +44 118 924 0022 >> Oracle JDeveloper Development >> Oracle Corporation UK Ltd is a company incorporated in England & Wales. >> Company Reg. No. 1782505. >> Reg. office: Oracle Parkway, Thames Valley Park, Reading RG6 1RA. > Cc: Dave Hawkins , "uml2-rtf@omg.org" From: Florian Schneider Subject: Re: [Issue 11068] Lifeline representing an actor Date: Fri, 28 May 2010 15:37:50 +0200 To: Østein Haugen X-Mailer: Apple Mail (2.1078) X-Virus-Scanned: ClamAV using ClamSMTP Thanks Oystein for the clarification! To sum things up: The actor symbol is not allowed in sequence and communication diagrams, but in fact elements on these diagrams can represent actor instances. Same goes for instances of other Type subclasses. Is that correct? In the situation I described: Is it correct to say the lifeline represents an actor instance? Regards, Florian On 28.05.2010, at 15:26, Østein Haugen wrote: > Florian > If you find such a connectable element, then a corresponding Lifeline to represent it looks just like any other lifeline with the head containing "name of connectable element : name of Actor" > You cannot legally use the Actor symbol on that lifeline even though I know several tools have implemented this option probably because they are still hung up on UML1 stuff. > > /Oystein > > > ---- > Dr. Oystein Haugen > Senior Researcher > SINTEF > > -----Original Message----- > From: Florian Schneider [mailto:schneidf@in.tum.de] > Sent: 28. mai 2010 15:20 > To: Østein Haugen > Cc: Dave Hawkins; uml2-rtf@omg.org > Subject: Re: [Issue 11068] Lifeline representing an actor > > I am sorry for any ambiguity! > > I agree that an instance of the class Actor is not in a sequence diagram (I also agree that it can be part of a use case diagram). But ConnectableElement, through its superclass TypedElement has an association (which is named 'type') to the class Type. > So when talking about a "ConnectableElement of type Actor", I mean that the association named type of the ConnectableElement instance points to an instance of the class Actor. I found no constraint that would make that impossible. > > Let's say we have a Lifeline that represents a ConnectableElement of type Actor. How is that notationally reflected on the according sequence diagram? > > Cheers, > Florian > > On 28.05.2010, at 14:58, Østein Haugen wrote: > >> Florian >> Model Elements that themselves are Types (such as Actor) are not "placed" in a sequence diagram (there is no concrete syntax for that), but since Interaction is a Class, such Types can be included in Interaction through the model browser. >> >> But, I do not think this is what you are asking. I sense a slight ambiguity in what you are writing since you must distinguish clearly between the elements that _have_ a type, and those that _are_ a type. Actor _is_ a type, but you do not find diagrams where there are connectable elements having a specific Actor as type. If you found such a context, you could also describe its behavior through an Interaction and let that connectable element be represented by a Lifeline. >> >> (Actors are found in use case diagrams, but the elements there are Actors, not connectable elements.) >> >> /Oystein >> >> >> ---- >> Dr. Oystein Haugen >> Senior Researcher >> SINTEF >> >> -----Original Message----- >> From: Florian Schneider [mailto:schneidf@in.tum.de] >> Sent: 28. mai 2010 14:48 >> To: Østein Haugen >> Cc: Dave Hawkins; uml2-rtf@omg.org >> Subject: Re: [Issue 11068] Lifeline representing an actor >> >> OK so maybe I got something wrong here. >> >> Here is how I understood the metamodel: A lifeline represents a ConnectableElement, which is a TypedElement. A TypedElement has a Type. Subclasses of Type are, among others, Classifier and BehavioredClassifier. Subclasses of Classifier are, among others, Class and Interface. Subclasses of BehavioredClassifier are, among others, Actor and UseCase. >> So as I understood it would be (syntactically) possible to place e.g. Classes, Interfaces, Actors and UseCases could be placed on a Sequence Diagram. Did I get that right up to here? >> >> Cheers, >> Florian >> >> On 28.05.2010, at 13:59, Østein Haugen wrote: >> >>> I agree with Dave's point, but let me provide the motivation for why a Lifeline respresents a ConnectableElement and not merely a Property. The reason is that there was a need to let sequence diagrams describe communications with Parameters and Parameter is not a specialization of Property. >>> >>> About the Type subclasses that you are asking about, I get a little confused, since Actor is not Property, and neither are UseCase nor Package, so these things cannot be represented by Lifelines. >>> >>> /Oystein >>> >>> >>> ---- >>> Dr. Oystein Haugen >>> Senior Researcher >>> SINTEF >>> >>> -----Original Message----- >>> From: Dave Hawkins [mailto:dave.hawkins@oracle.com] >>> Sent: 28. mai 2010 12:52 >>> To: Florian Schneider >>> Cc: uml2-rtf@omg.org >>> Subject: Re: [Issue 11068] Lifeline representing an actor >>> >>> The quoted sentence doesn't say it's equivalent, it's simply giving an example. Property >>> is a subclass of ConnectableElement and therefore it is valid for Lifeline::represents to >>> be a Property. >>> >>> Cheers, >>> >>> Dave >>> >>> Florian Schneider wrote: >>>> Dear Ladies and Gentlemen, >>>> >>>> I have a comment regarding issue 11068, reported by Mr. Tim Weilkiens. >>>> >>>> Mr. Weilkiens claims the following: "A lifeline represents a connectable element, e.g. a property". I believe that this is not correct because a ConnectableElement is a subclass of TypedElement so it can have any type. I don't see why it is equivalent to a property. >>>> >>>> Concerning this issue, I am interested in a clarification whether there are constraints on what Type subclasses may be placed on a Sequence Diagram (Exemplary questions: May an Actor be placed on a sequence diagram? May a use case be placed on a sequence diagram? May a package be placed on a sequence diagram?). >>>> >>>> With best regards, >>>> Florian Schneider >>>> >>>> >>>> >>>> -------------------------------- >>>> Dipl.-Inf. Univ. Florian Schneider >>>> >>>> Research Assistant >>>> Chair for Applied Software Engineering >>>> Technische UniversitäMü >>>> >>>> Tel.: +49 (89) 289-18233 >>>> Fax.: +49 (89) 289-18207 >>>> Web http://www1.in.tum.de/FlorianSchneider >>>> >>>> Address: >>>> Lehrstuhl füewandte Softwaretechnik, I1 >>>> Technische UniversitäMü >>>> Institut füormatik >>>> Boltzmannstraß 3 >>>> D-85748 Garching >>>> >>>> >>>> >>>> >>>> >>> >>> -- >>> Dave Hawkins | Principal Software Engineer | +44 118 924 0022 >>> Oracle JDeveloper Development >>> Oracle Corporation UK Ltd is a company incorporated in England & Wales. >>> Company Reg. No. 1782505. >>> Reg. office: Oracle Parkway, Thames Valley Park, Reading RG6 1RA. >> > smime3.p7s From: Østein Haugen To: Florian Schneider CC: Dave Hawkins , "uml2-rtf@omg.org" , Østein Haugen Date: Fri, 28 May 2010 15:50:28 +0200 Subject: RE: [Issue 11068] Lifeline representing an actor Thread-Topic: [Issue 11068] Lifeline representing an actor Thread-Index: Acr+avkEvDsLXd/PTrOD4QC0rFjZewAAIZ/g Accept-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: acceptlanguage: en-US X-MIME-Autoconverted: from quoted-printable to 8bit by amethyst.omg.org id o4SDcrC8017757 Florian We are converging ( The only slight problem with how you summarize the situation is your use of the word "instance". We are talking here about connectable elements. Instance specifications are something else, and I will _not_ open that can of worms now. Still, I would also point out that it is not so simple to find meaningful contexts where these very special connectable elements appear, and there is a constraint demanding that all lifelines in an Interaction represent connectable elements of the same context/container/composite structure. "The classifier containing the referenced ConnectableElement must be the same classifier, or an ancestor, of the classifier that contains the interaction enclosing this lifeline." /Oystein ---- Dr. Oystein Haugen Senior Researcher SINTEF -----Original Message----- From: Florian Schneider [mailto:schneidf@in.tum.de] Sent: 28. mai 2010 15:38 To: Østein Haugen Cc: Dave Hawkins; uml2-rtf@omg.org Subject: Re: [Issue 11068] Lifeline representing an actor Thanks Oystein for the clarification! To sum things up: The actor symbol is not allowed in sequence and communication diagrams, but in fact elements on these diagrams can represent actor instances. Same goes for instances of other Type subclasses. Is that correct? In the situation I described: Is it correct to say the lifeline represents an actor instance? Regards, Florian On 28.05.2010, at 15:26, Østein Haugen wrote: > Florian > If you find such a connectable element, then a corresponding Lifeline to represent it looks just like any other lifeline with the head containing "name of connectable element : name of Actor" > You cannot legally use the Actor symbol on that lifeline even though I know several tools have implemented this option probably because they are still hung up on UML1 stuff. > > /Oystein > > > ---- > Dr. Oystein Haugen > Senior Researcher > SINTEF > > -----Original Message----- > From: Florian Schneider [mailto:schneidf@in.tum.de] > Sent: 28. mai 2010 15:20 > To: Østein Haugen > Cc: Dave Hawkins; uml2-rtf@omg.org > Subject: Re: [Issue 11068] Lifeline representing an actor > > I am sorry for any ambiguity! > > I agree that an instance of the class Actor is not in a sequence diagram (I also agree that it can be part of a use case diagram). But ConnectableElement, through its superclass TypedElement has an association (which is named 'type') to the class Type. > So when talking about a "ConnectableElement of type Actor", I mean that the association named type of the ConnectableElement instance points to an instance of the class Actor. I found no constraint that would make that impossible. > > Let's say we have a Lifeline that represents a ConnectableElement of type Actor. How is that notationally reflected on the according sequence diagram? > > Cheers, > Florian > > On 28.05.2010, at 14:58, Østein Haugen wrote: > >> Florian >> Model Elements that themselves are Types (such as Actor) are not "placed" in a sequence diagram (there is no concrete syntax for that), but since Interaction is a Class, such Types can be included in Interaction through the model browser. >> >> But, I do not think this is what you are asking. I sense a slight ambiguity in what you are writing since you must distinguish clearly between the elements that _have_ a type, and those that _are_ a type. Actor _is_ a type, but you do not find diagrams where there are connectable elements having a specific Actor as type. If you found such a context, you could also describe its behavior through an Interaction and let that connectable element be represented by a Lifeline. >> >> (Actors are found in use case diagrams, but the elements there are Actors, not connectable elements.) >> >> /Oystein >> >> >> ---- >> Dr. Oystein Haugen >> Senior Researcher >> SINTEF >> >> -----Original Message----- >> From: Florian Schneider [mailto:schneidf@in.tum.de] >> Sent: 28. mai 2010 14:48 >> To: Østein Haugen >> Cc: Dave Hawkins; uml2-rtf@omg.org >> Subject: Re: [Issue 11068] Lifeline representing an actor >> >> OK so maybe I got something wrong here. >> >> Here is how I understood the metamodel: A lifeline represents a ConnectableElement, which is a TypedElement. A TypedElement has a Type. Subclasses of Type are, among others, Classifier and BehavioredClassifier. Subclasses of Classifier are, among others, Class and Interface. Subclasses of BehavioredClassifier are, among others, Actor and UseCase. >> So as I understood it would be (syntactically) possible to place e.g. Classes, Interfaces, Actors and UseCases could be placed on a Sequence Diagram. Did I get that right up to here? >> >> Cheers, >> Florian >> >> On 28.05.2010, at 13:59, Østein Haugen wrote: >> >>> I agree with Dave's point, but let me provide the motivation for why a Lifeline respresents a ConnectableElement and not merely a Property. The reason is that there was a need to let sequence diagrams describe communications with Parameters and Parameter is not a specialization of Property. >>> >>> About the Type subclasses that you are asking about, I get a little confused, since Actor is not Property, and neither are UseCase nor Package, so these things cannot be represented by Lifelines. >>> >>> /Oystein >>> >>> >>> ---- >>> Dr. Oystein Haugen >>> Senior Researcher >>> SINTEF >>> >>> -----Original Message----- >>> From: Dave Hawkins [mailto:dave.hawkins@oracle.com] >>> Sent: 28. mai 2010 12:52 >>> To: Florian Schneider >>> Cc: uml2-rtf@omg.org >>> Subject: Re: [Issue 11068] Lifeline representing an actor >>> >>> The quoted sentence doesn't say it's equivalent, it's simply giving an example. Property >>> is a subclass of ConnectableElement and therefore it is valid for Lifeline::represents to >>> be a Property. >>> >>> Cheers, >>> >>> Dave >>> >>> Florian Schneider wrote: >>>> Dear Ladies and Gentlemen, >>>> >>>> I have a comment regarding issue 11068, reported by Mr. Tim Weilkiens. >>>> >>>> Mr. Weilkiens claims the following: "A lifeline represents a connectable element, e.g. a property". I believe that this is not correct because a ConnectableElement is a subclass of TypedElement so it can have any type. I don't see why it is equivalent to a property. >>>> >>>> Concerning this issue, I am interested in a clarification whether there are constraints on what Type subclasses may be placed on a Sequence Diagram (Exemplary questions: May an Actor be placed on a sequence diagram? May a use case be placed on a sequence diagram? May a package be placed on a sequence diagram?). >>>> >>>> With best regards, >>>> Florian Schneider >>>> >>>> >>>> >>>> -------------------------------- >>>> Dipl.-Inf. Univ. Florian Schneider >>>> >>>> Research Assistant >>>> Chair for Applied Software Engineering >>>> Technische UniversitäMü >>>> >>>> Tel.: +49 (89) 289-18233 >>>> Fax.: +49 (89) 289-18207 >>>> Web http://www1.in.tum.de/FlorianSchneider >>>> >>>> Address: >>>> Lehrstuhl füewandte Softwaretechnik, I1 >>>> Technische UniversitäMü >>>> Institut füormatik >>>> Boltzmannstraß 3 >>>> D-85748 Garching >>>> >>>> >>>> >>>> >>>> >>> >>> -- >>> Dave Hawkins | Principal Software Engineer | +44 118 924 0022 >>> Oracle JDeveloper Development >>> Oracle Corporation UK Ltd is a company incorporated in England & Wales. >>> Company Reg. No. 1782505. >>> Reg. office: Oracle Parkway, Thames Valley Park, Reading RG6 1RA. >> > Date: Fri, 28 May 2010 09:56:33 -0400 From: "Chonoles, Michael J" Subject: RE: [Issue 11068] Lifeline representing an actor To: Østein Haugen , Florian Schneider Cc: Dave Hawkins , "uml2-rtf@omg.org" Thread-Topic: [Issue 11068] Lifeline representing an actor Thread-Index: Acr+avkEvDsLXd/PTrOD4QC0rFjZewAAIZ/gAABgwJA= Accept-Language: en-US acceptlanguage: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: X-MIME-Autoconverted: from quoted-printable to 8bit by amethyst.omg.org id o4SDj8Xp021035 However, since an interaction is really more of a conceptual thing, it is easy to just declare all the connectable elements to be part of the classifier, for example by placing them on a class diagram where the class being described is the interaction. This is exactly an area of UML 2 that needs to be revisited as we move in UML 3. The limitations imposed here are interfering with practical use. -----Original Message----- From: Østein Haugen [mailto:Oystein.Haugen@sintef.no] Sent: Friday, May 28, 2010 9:50 AM To: Florian Schneider Cc: Dave Hawkins; uml2-rtf@omg.org; Østein Haugen Subject: RE: [Issue 11068] Lifeline representing an actor Florian We are converging ( The only slight problem with how you summarize the situation is your use of the word "instance". We are talking here about connectable elements. Instance specifications are something else, and I will _not_ open that can of worms now. Still, I would also point out that it is not so simple to find meaningful contexts where these very special connectable elements appear, and there is a constraint demanding that all lifelines in an Interaction represent connectable elements of the same context/container/composite structure. "The classifier containing the referenced ConnectableElement must be the same classifier, or an ancestor, of the classifier that contains the interaction enclosing this lifeline." /Oystein ---- Dr. Oystein Haugen Senior Researcher SINTEF -----Original Message----- From: Florian Schneider [mailto:schneidf@in.tum.de] Sent: 28. mai 2010 15:38 To: Østein Haugen Cc: Dave Hawkins; uml2-rtf@omg.org Subject: Re: [Issue 11068] Lifeline representing an actor Thanks Oystein for the clarification! To sum things up: The actor symbol is not allowed in sequence and communication diagrams, but in fact elements on these diagrams can represent actor instances. Same goes for instances of other Type subclasses. Is that correct? In the situation I described: Is it correct to say the lifeline represents an actor instance? Regards, Florian On 28.05.2010, at 15:26, Østein Haugen wrote: > Florian > If you find such a connectable element, then a corresponding Lifeline to represent it looks just like any other lifeline with the head containing "name of connectable element : name of Actor" > You cannot legally use the Actor symbol on that lifeline even though I know several tools have implemented this option probably because they are still hung up on UML1 stuff. > > /Oystein > > > ---- > Dr. Oystein Haugen > Senior Researcher > SINTEF > > -----Original Message----- > From: Florian Schneider [mailto:schneidf@in.tum.de] > Sent: 28. mai 2010 15:20 > To: Østein Haugen > Cc: Dave Hawkins; uml2-rtf@omg.org > Subject: Re: [Issue 11068] Lifeline representing an actor > > I am sorry for any ambiguity! > > I agree that an instance of the class Actor is not in a sequence diagram (I also agree that it can be part of a use case diagram). But ConnectableElement, through its superclass TypedElement has an association (which is named 'type') to the class Type. > So when talking about a "ConnectableElement of type Actor", I mean that the association named type of the ConnectableElement instance points to an instance of the class Actor. I found no constraint that would make that impossible. > > Let's say we have a Lifeline that represents a ConnectableElement of type Actor. How is that notationally reflected on the according sequence diagram? > > Cheers, > Florian > > On 28.05.2010, at 14:58, Østein Haugen wrote: > >> Florian >> Model Elements that themselves are Types (such as Actor) are not "placed" in a sequence diagram (there is no concrete syntax for that), but since Interaction is a Class, such Types can be included in Interaction through the model browser. >> >> But, I do not think this is what you are asking. I sense a slight ambiguity in what you are writing since you must distinguish clearly between the elements that _have_ a type, and those that _are_ a type. Actor _is_ a type, but you do not find diagrams where there are connectable elements having a specific Actor as type. If you found such a context, you could also describe its behavior through an Interaction and let that connectable element be represented by a Lifeline. >> >> (Actors are found in use case diagrams, but the elements there are Actors, not connectable elements.) >> >> /Oystein >> >> >> ---- >> Dr. Oystein Haugen >> Senior Researcher >> SINTEF >> >> -----Original Message----- >> From: Florian Schneider [mailto:schneidf@in.tum.de] >> Sent: 28. mai 2010 14:48 >> To: Østein Haugen >> Cc: Dave Hawkins; uml2-rtf@omg.org >> Subject: Re: [Issue 11068] Lifeline representing an actor >> >> OK so maybe I got something wrong here. >> >> Here is how I understood the metamodel: A lifeline represents a ConnectableElement, which is a TypedElement. A TypedElement has a Type. Subclasses of Type are, among others, Classifier and BehavioredClassifier. Subclasses of Classifier are, among others, Class and Interface. Subclasses of BehavioredClassifier are, among others, Actor and UseCase. >> So as I understood it would be (syntactically) possible to place e.g. Classes, Interfaces, Actors and UseCases could be placed on a Sequence Diagram. Did I get that right up to here? >> >> Cheers, >> Florian >> >> On 28.05.2010, at 13:59, Østein Haugen wrote: >> >>> I agree with Dave's point, but let me provide the motivation for why a Lifeline respresents a ConnectableElement and not merely a Property. The reason is that there was a need to let sequence diagrams describe communications with Parameters and Parameter is not a specialization of Property. >>> >>> About the Type subclasses that you are asking about, I get a little confused, since Actor is not Property, and neither are UseCase nor Package, so these things cannot be represented by Lifelines. >>> >>> /Oystein >>> >>> >>> ---- >>> Dr. Oystein Haugen >>> Senior Researcher >>> SINTEF >>> >>> -----Original Message----- >>> From: Dave Hawkins [mailto:dave.hawkins@oracle.com] >>> Sent: 28. mai 2010 12:52 >>> To: Florian Schneider >>> Cc: uml2-rtf@omg.org >>> Subject: Re: [Issue 11068] Lifeline representing an actor >>> >>> The quoted sentence doesn't say it's equivalent, it's simply giving an example. Property >>> is a subclass of ConnectableElement and therefore it is valid for Lifeline::represents to >>> be a Property. >>> >>> Cheers, >>> >>> Dave >>> >>> Florian Schneider wrote: >>>> Dear Ladies and Gentlemen, >>>> >>>> I have a comment regarding issue 11068, reported by Mr. Tim Weilkiens. >>>> >>>> Mr. Weilkiens claims the following: "A lifeline represents a connectable element, e.g. a property". I believe that this is not correct because a ConnectableElement is a subclass of TypedElement so it can have any type. I don't see why it is equivalent to a property. >>>> >>>> Concerning this issue, I am interested in a clarification whether there are constraints on what Type subclasses may be placed on a Sequence Diagram (Exemplary questions: May an Actor be placed on a sequence diagram? May a use case be placed on a sequence diagram? May a package be placed on a sequence diagram?). >>>> >>>> With best regards, >>>> Florian Schneider >>>> >>>> >>>> >>>> -------------------------------- >>>> Dipl.-Inf. Univ. Florian Schneider >>>> >>>> Research Assistant >>>> Chair for Applied Software Engineering >>>> Technische UniversitäMü >>>> >>>> Tel.: +49 (89) 289-18233 >>>> Fax.: +49 (89) 289-18207 >>>> Web http://www1.in.tum.de/FlorianSchneider >>>> >>>> Address: >>>> Lehrstuhl füewandte Softwaretechnik, I1 >>>> Technische UniversitäMü >>>> Institut füormatik >>>> Boltzmannstraß 3 >>>> D-85748 Garching >>>> >>>> >>>> >>>> >>>> >>> >>> -- >>> Dave Hawkins | Principal Software Engineer | +44 118 924 0022 >>> Oracle JDeveloper Development >>> Oracle Corporation UK Ltd is a company incorporated in England & Wales. >>> Company Reg. No. 1782505. >>> Reg. office: Oracle Parkway, Thames Valley Park, Reading RG6 1RA. >> > From: "Bock, Conrad" To: Florian Schneider , "uml2-rtf@omg.org" Date: Fri, 28 May 2010 09:58:42 -0400 Subject: RE: [Issue 11068] Lifeline representing an actor Thread-Topic: [Issue 11068] Lifeline representing an actor Thread-Index: Acr+UkQojg2KoIQxQL2RGawKIWqqxwAG2gRg Accept-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: acceptlanguage: en-US X-NIST-MailScanner: Found to be clean X-NIST-MailScanner-From: conrad.bock@nist.gov Florian, > Mr. Weilkiens claims the following: "A lifeline represents > a connectable element, e.g. a property". I believe that > this is not correct because a ConnectableElement is a > subclass of TypedElement so it can have any type. I don't > see why it is equivalent to a property. The abbreviation "e.g." means "for example", so Tim's sentence is correct. A property is an example of a connectable element. Conrad Cc: Dave Hawkins , "uml2-rtf@omg.org" From: Florian Schneider Subject: Re: [Issue 11068] Lifeline representing an actor Date: Fri, 28 May 2010 16:25:05 +0200 To: Østein Haugen X-Mailer: Apple Mail (2.1078) X-Virus-Scanned: ClamAV using ClamSMTP Oystein, I also agree that connectable elements are no instance specifications. (we are definitely converging..) I just do not know how to name it tersely when talking of linelines that represent connectable elements which have a type. I know of some textbooks (not only CASE tools) that use actor symbols on sequence diagrams to show how an actor interacts with the system under development. This has been used conventionally and before this conversation, I was not sure whether this is syntactically correct. Conrad and Dave, of course I was wrong. Properties are connectable elements. I did not see that 7.3.44 Property (from Kernel, AssociationClasses, Interfaces) is merged incremented by 9.3.12 Property (from InternalStructures) from which ConnectableElement is a generalization. My excuses go to Tim What I originally wanted to contradict is that it is not possible to put actor symbols on sequence diagrams. I thought the metamodel would allow for ConnectableElements of any type to be represented by Lifelines. Which is impossible as I just learned. I will have another look at Properties and the constraint that Oystein mentioned. Cheers and thanks for the feedback, Florian On 28.05.2010, at 15:50, Østein Haugen wrote: > Florian > We are converging ( > The only slight problem with how you summarize the situation is your use of the word "instance". We are talking here about connectable elements. Instance specifications are something else, and I will _not_ open that can of worms now. > > Still, I would also point out that it is not so simple to find meaningful contexts where these very special connectable elements appear, and there is a constraint demanding that all lifelines in an Interaction represent connectable elements of the same context/container/composite structure. > > "The classifier containing the referenced ConnectableElement must be the same classifier, or an ancestor, of the classifier > that contains the interaction enclosing this lifeline." > > /Oystein > > [...] >> Mr. Weilkiens claims the following: "A lifeline represents >> a connectable element, e.g. a property". I believe that >> this is not correct because a ConnectableElement is a >> subclass of TypedElement so it can have any type. I don't >> see why it is equivalent to a property. > > The abbreviation "e.g." means "for example", so Tim's sentence > is correct. A property is an example of a connectable element. > > Conrad > The quoted sentence doesn't say it's equivalent, it's simply giving an example. Property > is a subclass of ConnectableElement and therefore it is valid for Lifeline::represents to > be a Property. > > Cheers, > > Dave smime4.p7s DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:received:mime-version:sender:received :in-reply-to:references:from:date:x-google-sender-auth:message-id :subject:to:cc:content-type; bh=WynpGRFR6dP69uuLrRxR/N9f0nkLSiOYaE+J8D9ATAI=; b=UzAkbr8lscaaQCYLpYFRdnWfX2BnlAXPRqj14bjGUnCEo4axE2wx4D6BIDoIxWjAFp MEzRWnnmM+ZN+3WK4baWVw/ZBKmYLkUE6Os/jT71ZOXS46ssy4kptAL2C43L7nJHoAYH euDFM1Cbe5dLZmF4TPB5a8x7QxYzbqCrxBBGQ= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:in-reply-to:references:from:date :x-google-sender-auth:message-id:subject:to:cc:content-type; b=ATSD6etIsE0a/RNefDiWAs4uBKPomK+EETWMDkzX8sjqlyCoBn36C4ZBZ0AudPhUvB I3L1djygED9zKYqg2IbJq0S4pk7YK6WlzKQ33vCk4xgVD9BQXPISx3iPe6mUs+8m629u /n9JZYKD2yUxe6Vydz6JUJNMJQyOaZbWZU8nw= Sender: bran.selic@gmail.com From: Bran Selic Date: Fri, 28 May 2010 15:56:46 -0400 X-Google-Sender-Auth: ZeEnY-aOnjsmHaWal33V2ZWVms4 Subject: Re: [Issue 11068] Lifeline representing an actor To: Florian Schneider Cc: Østein Haugen , Dave Hawkins , "uml2-rtf@omg.org" Actually, there is nothing in the UML 2 spec preventing an actor SYMBOL from being used as a lifeline header (UML graphcial syntax only specifies what is allowed and is mostly silent on what is not allowed -- which gives you a lot of freedom). The only constraint is that the actor symbol has to represent a connectable element. What is not allowed is to use an Actor (which is a kind of Classifier) as a lifeline. This is not the same as saying that you cannot use the actor symbol. Cheers...Bran 2010/5/28 Florian Schneider Thanks Oystein for the clarification! To sum things up: The actor symbol is not allowed in sequence and communication diagrams, but in fact elements on these diagrams can represent actor instances. Same goes for instances of other Type subclasses. Is that correct? In the situation I described: Is it correct to say the lifeline represents an actor instance? Regards, Florian On 28.05.2010, at 15:26, Østein Haugen wrote: > Florian > If you find such a connectable element, then a corresponding Lifeline to represent it looks just like any other lifeline with the head containing "name of connectable element : name of Actor" > You cannot legally use the Actor symbol on that lifeline even though I know several tools have implemented this option probably because they are still hung up on UML1 stuff. > > /Oystein > > > ---- > Dr. Oystein Haugen > Senior Researcher > SINTEF > > -----Original Message----- > From: Florian Schneider [mailto:schneidf@in.tum.de] > Sent: 28. mai 2010 15:20 > To: Østein Haugen > Cc: Dave Hawkins; uml2-rtf@omg.org > Subject: Re: [Issue 11068] Lifeline representing an actor > > I am sorry for any ambiguity! > > I agree that an instance of the class Actor is not in a sequence diagram (I also agree that it can be part of a use case diagram). But ConnectableElement, through its superclass TypedElement has an association (which is named 'type') to the class Type. > So when talking about a "ConnectableElement of type Actor", I mean that the association named type of the ConnectableElement instance points to an instance of the class Actor. I found no constraint that would make that impossible. > > Let's say we have a Lifeline that represents a ConnectableElement of type Actor. How is that notationally reflected on the according sequence diagram? > > Cheers, > Florian > > On 28.05.2010, at 14:58, Østein Haugen wrote: > >> Florian >> Model Elements that themselves are Types (such as Actor) are not "placed" in a sequence diagram (there is no concrete syntax for that), but since Interaction is a Class, such Types can be included in Interaction through the model browser. >> >> But, I do not think this is what you are asking. I sense a slight ambiguity in what you are writing since you must distinguish clearly between the elements that _have_ a type, and those that _are_ a type. Actor _is_ a type, but you do not find diagrams where there are connectable elements having a specific Actor as type. If you found such a context, you could also describe its behavior through an Interaction and let that connectable element be represented by a Lifeline. >> >> (Actors are found in use case diagrams, but the elements there are Actors, not connectable elements.) >> >> /Oystein >> >> >> ---- >> Dr. Oystein Haugen >> Senior Researcher >> SINTEF >> >> -----Original Message----- >> From: Florian Schneider [mailto:schneidf@in.tum.de] >> Sent: 28. mai 2010 14:48 >> To: Østein Haugen >> Cc: Dave Hawkins; uml2-rtf@omg.org >> Subject: Re: [Issue 11068] Lifeline representing an actor >> >> OK so maybe I got something wrong here. >> >> Here is how I understood the metamodel: A lifeline represents a ConnectableElement, which is a TypedElement. A TypedElement has a Type. Subclasses of Type are, among others, Classifier and BehavioredClassifier. Subclasses of Classifier are, among others, Class and Interface. Subclasses of BehavioredClassifier are, among others, Actor and UseCase. >> So as I understood it would be (syntactically) possible to place e.g. Classes, Interfaces, Actors and UseCases could be placed on a Sequence Diagram. Did I get that right up to here? >> >> Cheers, >> Florian >> >> On 28.05.2010, at 13:59, Østein Haugen wrote: >> >>> I agree with Dave's point, but let me provide the motivation for why a Lifeline respresents a ConnectableElement and not merely a Property. The reason is that there was a need to let sequence diagrams describe communications with Parameters and Parameter is not a specialization of Property. >>> >>> About the Type subclasses that you are asking about, I get a little confused, since Actor is not Property, and neither are UseCase nor Package, so these things cannot be represented by Lifelines. >>> >>> /Oystein >>> >>> >>> ---- >>> Dr. Oystein Haugen >>> Senior Researcher >>> SINTEF >>> >>> -----Original Message----- >>> From: Dave Hawkins [mailto:dave.hawkins@oracle.com] >>> Sent: 28. mai 2010 12:52 >>> To: Florian Schneider >>> Cc: uml2-rtf@omg.org >>> Subject: Re: [Issue 11068] Lifeline representing an actor >>> >>> The quoted sentence doesn't say it's equivalent, it's simply giving an example. Property >>> is a subclass of ConnectableElement and therefore it is valid for Lifeline::represents to >>> be a Property. >>> >>> Cheers, >>> >>> Dave >>> >>> Florian Schneider wrote: >>>> Dear Ladies and Gentlemen, >>>> >>>> I have a comment regarding issue 11068, reported by Mr. Tim Weilkiens. >>>> >>>> Mr. Weilkiens claims the following: "A lifeline represents a connectable element, e.g. a property". I believe that this is not correct because a ConnectableElement is a subclass of TypedElement so it can have any type. I don't see why it is equivalent to a property. >>>> >>>> Concerning this issue, I am interested in a clarification whether there are constraints on what Type subclasses may be placed on a Sequence Diagram (Exemplary questions: May an Actor be placed on a sequence diagram? May a use case be placed on a sequence diagram? May a package be placed on a sequence diagram?). >>>> >>>> With best regards, >>>> Florian Schneider >>>> >>>> >>>> >>>> -------------------------------- >>>> Dipl.-Inf. Univ. Florian Schneider >>>> >>>> Research Assistant >>>> Chair for Applied Software Engineering >>>> Technische UniversitäMü >>>> >>>> Tel.: +49 (89) 289-18233 >>>> Fax.: +49 (89) 289-18207 >>>> Web http://www1.in.tum.de/FlorianSchneider >>>> >>>> Address: >>>> Lehrstuhl füewandte Softwaretechnik, I1 >>>> Technische UniversitäMü >>>> Institut füormatik >>>> Boltzmannstraß 3 >>>> D-85748 Garching >>>> >>>> >>>> >>>> >>>> >>> >>> -- >>> Dave Hawkins | Principal Software Engineer | +44 118 924 0022 >>> Oracle JDeveloper Development >>> Oracle Corporation UK Ltd is a company incorporated in England & Wales. >>> Company Reg. No. 1782505. >>> Reg. office: Oracle Parkway, Thames Valley Park, Reading RG6 1RA. >> > From: "Rouquette, Nicolas F (316A)" To: Bran Selic CC: Florian Schneider , Østein Haugen , Dave Hawkins , "uml2-rtf@omg.org" Date: Fri, 28 May 2010 14:31:56 -0700 Subject: Re: [Issue 11068] Lifeline representing an actor Thread-Topic: [Issue 11068] Lifeline representing an actor Thread-Index: Acr+rTMVVhKTv6egSGeAIdMylpp6Cg== Accept-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: acceptlanguage: en-US X-Source-IP: altvirehtstap01.jpl.nasa.gov [128.149.137.72] X-Source-Sender: nicolas.f.rouquette@jpl.nasa.gov X-AUTH: Authorized Bran, I'm surprised that you'd go as far as suggesting to use symbolic boxology; what happened to UML dogfooding? Wouldn't it be better here to use an idiom associating a sequence diagram to a collaboration whose parameters (a kind of connectable element) can be typed by actor classifiers? With judicious idioms, one can be precise with the UML, much more than most people are giving UML credit for. Without this, the whole metamodeling architecture would have collapsed a long time ago. - Nicolas. On May 28, 2010, at 12:56 PM, Bran Selic wrote: Actually, there is nothing in the UML 2 spec preventing an actor SYMBOL from being used as a lifeline header (UML graphcial syntax only specifies what is allowed and is mostly silent on what is not allowed -- which gives you a lot of freedom). The only constraint is that the actor symbol has to represent a connectable element. What is not allowed is to use an Actor (which is a kind of Classifier) as a lifeline. This is not the same as saying that you cannot use the actor symbol. Cheers...Bran 2010/5/28 Florian Schneider Thanks Oystein for the clarification! To sum things up: The actor symbol is not allowed in sequence and communication diagrams, but in fact elements on these diagrams can represent actor instances. Same goes for instances of other Type subclasses. Is that correct? In the situation I described: Is it correct to say the lifeline represents an actor instance? Regards, Florian On 28.05.2010, at 15:26, Østein Haugen wrote: > Florian > If you find such a connectable element, then a corresponding Lifeline to represent it looks just like any other lifeline with the head containing "name of connectable element : name of Actor" > You cannot legally use the Actor symbol on that lifeline even though I know several tools have implemented this option probably because they are still hung up on UML1 stuff. > > /Oystein > > > ---- > Dr. Oystein Haugen > Senior Researcher > SINTEF > > -----Original Message----- > From: Florian Schneider [mailto:schneidf@in.tum.de] > Sent: 28. mai 2010 15:20 > To: Østein Haugen > Cc: Dave Hawkins; uml2-rtf@omg.org > Subject: Re: [Issue 11068] Lifeline representing an actor > > I am sorry for any ambiguity! > > I agree that an instance of the class Actor is not in a sequence diagram (I also agree that it can be part of a use case diagram). But ConnectableElement, through its superclass TypedElement has an association (which is named 'type') to the class Type. > So when talking about a "ConnectableElement of type Actor", I mean that the association named type of the ConnectableElement instance points to an instance of the class Actor. I found no constraint that would make that impossible. > > Let's say we have a Lifeline that represents a ConnectableElement of type Actor. How is that notationally reflected on the according sequence diagram? > > Cheers, > Florian > > On 28.05.2010, at 14:58, Østein Haugen wrote: > >> Florian >> Model Elements that themselves are Types (such as Actor) are not "placed" in a sequence diagram (there is no concrete syntax for that), but since Interaction is a Class, such Types can be included in Interaction through the model browser. >> >> But, I do not think this is what you are asking. I sense a slight ambiguity in what you are writing since you must distinguish clearly between the elements that _have_ a type, and those that _are_ a type. Actor _is_ a type, but you do not find diagrams where there are connectable elements having a specific Actor as type. If you found such a context, you could also describe its behavior through an Interaction and let that connectable element be represented by a Lifeline. >> >> (Actors are found in use case diagrams, but the elements there are Actors, not connectable elements.) >> >> /Oystein >> >> >> ---- >> Dr. Oystein Haugen >> Senior Researcher >> SINTEF >> >> -----Original Message----- >> From: Florian Schneider [mailto:schneidf@in.tum.de] >> Sent: 28. mai 2010 14:48 >> To: Østein Haugen >> Cc: Dave Hawkins; uml2-rtf@omg.org >> Subject: Re: [Issue 11068] Lifeline representing an actor >> >> OK so maybe I got something wrong here. >> >> Here is how I understood the metamodel: A lifeline represents a ConnectableElement, which is a TypedElement. A TypedElement has a Type. Subclasses of Type are, among others, Classifier and BehavioredClassifier. Subclasses of Classifier are, among others, Class and Interface. Subclasses of BehavioredClassifier are, among others, Actor and UseCase. >> So as I understood it would be (syntactically) possible to place e.g. Classes, Interfaces, Actors and UseCases could be placed on a Sequence Diagram. Did I get that right up to here? >> >> Cheers, >> Florian >> >> On 28.05.2010, at 13:59, Østein Haugen wrote: >> >>> I agree with Dave's point, but let me provide the motivation for why a Lifeline respresents a ConnectableElement and not merely a Property. The reason is that there was a need to let sequence diagrams describe communications with Parameters and Parameter is not a specialization of Property. >>> >>> About the Type subclasses that you are asking about, I get a little confused, since Actor is not Property, and neither are UseCase nor Package, so these things cannot be represented by Lifelines. >>> >>> /Oystein >>> >>> >>> ---- >>> Dr. Oystein Haugen >>> Senior Researcher >>> SINTEF >>> >>> -----Original Message----- >>> From: Dave Hawkins [mailto:dave.hawkins@oracle.com] >>> Sent: 28. mai 2010 12:52 >>> To: Florian Schneider >>> Cc: uml2-rtf@omg.org >>> Subject: Re: [Issue 11068] Lifeline representing an actor >>> >>> The quoted sentence doesn't say it's equivalent, it's simply giving an example. Property >>> is a subclass of ConnectableElement and therefore it is valid for Lifeline::represents to >>> be a Property. >>> >>> Cheers, >>> >>> Dave >>> >>> Florian Schneider wrote: >>>> Dear Ladies and Gentlemen, >>>> >>>> I have a comment regarding issue 11068, reported by Mr. Tim Weilkiens. >>>> >>>> Mr. Weilkiens claims the following: "A lifeline represents a connectable element, e.g. a property". I believe that this is not correct because a ConnectableElement is a subclass of TypedElement so it can have any type. I don't see why it is equivalent to a property. >>>> >>>> Concerning this issue, I am interested in a clarification whether there are constraints on what Type subclasses may be placed on a Sequence Diagram (Exemplary questions: May an Actor be placed on a sequence diagram? May a use case be placed on a sequence diagram? May a package be placed on a sequence diagram?). >>>> >>>> With best regards, >>>> Florian Schneider >>>> >>>> >>>> >>>> -------------------------------- >>>> Dipl.-Inf. Univ. Florian Schneider >>>> >>>> Research Assistant >>>> Chair for Applied Software Engineering >>>> Technische UniversitäMü >>>> >>>> Tel.: +49 (89) 289-18233 >>>> Fax.: +49 (89) 289-18207 >>>> Web http://www1.in.tum.de/FlorianSchneider >>>> >>>> Address: >>>> Lehrstuhl füewandte Softwaretechnik, I1 >>>> Technische UniversitäMü >>>> Institut füormatik >>>> Boltzmannstraß 3 >>>> D-85748 Garching >>>> >>>> >>>> >>>> >>>> >>> >>> -- >>> Dave Hawkins | Principal Software Engineer | +44 118 924 0022 >>> Oracle JDeveloper Development >>> Oracle Corporation UK Ltd is a company incorporated in England & Wales. >>> Company Reg. No. 1782505. >>> Reg. office: Oracle Parkway, Thames Valley Park, Reading RG6 1RA. >> > DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:received:mime-version:sender:received :in-reply-to:references:from:date:x-google-sender-auth:message-id :subject:to:cc:content-type; bh=+HX4CyWnY1aJwBoZ+7mGiGaUKHclgTyFnK7gWwnzjOs=; b=rBnwPTXjoizjhcGamXCC7RYvBshkXM9Jz9ZCfaJVMbSwjzO9soAnnqUjxSHABp5JYc cdHu78llxlNz/E6BFVYRl21TE9t1dcodXwfdmUkh26c9S6xisuN5xqel9B46qs2tdxuj IZPJxfs104S4j5QATKVGgSLeVVgWm2PxXUz7c= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:in-reply-to:references:from:date :x-google-sender-auth:message-id:subject:to:cc:content-type; b=MERLiSqBQ5RApdkQAlOCecejcB+LfdhSWsrgmD/sSh8LoV1k2nj0UfmVa33HGN04vu ShmAh9QiI3KZeS0KjLcDiK44lu3ZuInzJY8rcIzN8z8+JVA6SWSU9h+3LEsigETJHwsU micahFSJlCvhrywr0szS8Yux0vXC0KyRQmcWw= Sender: bran.selic@gmail.com From: Bran Selic Date: Mon, 31 May 2010 08:27:40 -0400 X-Google-Sender-Auth: I_3MoD-l7eJvk1dl4xgN6eH2OZc Subject: Re: [Issue 11068] Lifeline representing an actor To: "Rouquette, Nicolas F (316A)" Cc: Florian Schneider , Østein Haugen , Dave Hawkins , "uml2-rtf@omg.org" Hi Nicolas, I am not sure what you are saying, so I may be responding to the wrong view. On Fri, May 28, 2010 at 5:31 PM, Rouquette, Nicolas F (316A) wrote: Bran, I'm surprised that you'd go as far as suggesting to use symbolic boxology; what happened to UML dogfooding? [bvs] Not sure what you mean by "symbolic boxology". I was simply saying that it was quite within the current bounds of UML concrete syntax to choose your own iconic representation for something like a lifeline element, including, if so desired, the use of a human stick figure in a sequence diagram to represent the head of a lifeline. The fact that the same symbol is used to represent an actor in a use case diagram is not a formal barrier. I am not suggesting any changes to the metamodel or a relaxation of its rules. Wouldn't it be better here to use an idiom associating a sequence diagram to a collaboration whose parameters (a kind of connectable element) can be typed by actor classifiers? [bvs] Of course. But, I didn't think that was the question. And, BTW, you could use the stick figure to represent a role in such a collaboration diagram. With judicious idioms, one can be precise with the UML, much more than most people are giving UML credit for. Without this, the whole metamodeling architecture would have collapsed a long time ago. [bvs] I believe that you misunderstood my point. Cheers...Bran From: "Rouquette, Nicolas F (316A)" To: Bran Selic CC: Florian Schneider , Østein Haugen , Dave Hawkins , "uml2-rtf@omg.org" Date: Mon, 31 May 2010 15:34:38 -0700 Subject: Re: [Issue 11068] Lifeline representing an actor Thread-Topic: [Issue 11068] Lifeline representing an actor Thread-Index: AcsBEXRY6e1CQ0rYQm2S/5HRnuu2uw== Accept-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: acceptlanguage: en-US X-Source-IP: altvirehtstap02.jpl.nasa.gov [128.149.137.73] X-Source-Sender: nicolas.f.rouquette@jpl.nasa.gov X-AUTH: Authorized Bran, On May 31, 2010, at 5:27 AM, Bran Selic wrote: Hi Nicolas, I am not sure what you are saying, so I may be responding to the wrong view. Now it's me who's confused. On Fri, May 28, 2010 at 5:31 PM, Rouquette, Nicolas F (316A) wrote: Bran, I'm surprised that you'd go as far as suggesting to use symbolic boxology; what happened to UML dogfooding? [bvs] Not sure what you mean by "symbolic boxology". It's a term that I heard from Ralph Hodgson who uses it to refer to diagrams involving boxes & lines. Sure these diagrams say a 1,000 words but the words may be different to different people. I was simply saying that it was quite within the current bounds of UML concrete syntax to choose your own iconic representation for something like a lifeline element, including, if so desired, the use of a human stick figure in a sequence diagram to represent the head of a lifeline. Ok. The fact that the same symbol is used to represent an actor in a use case diagram is not a formal barrier. This is the part that gets a bit slippery. What do you mean by "same symbol"? The original question came about the fact that a lifeline is a connectable element but an actor isn't. How could the symbol of an actor then show up as the symbol that's meant to represent a connectable element? - Nicolas. I am not suggesting any changes to the metamodel or a relaxation of its rules. Wouldn't it be better here to use an idiom associating a sequence diagram to a collaboration whose parameters (a kind of connectable element) can be typed by actor classifiers? [bvs] Of course. But, I didn't think that was the question. And, BTW, you could use the stick figure to represent a role in such a collaboration diagram. With judicious idioms, one can be precise with the UML, much more than most people are giving UML credit for. Without this, the whole metamodeling architecture would have collapsed a long time ago. [bvs] I believe that you misunderstood my point. Cheers...Bran Cc: Bran Selic , Østein Haugen , Dave Hawkins , "uml2-rtf@omg.org" From: Florian Schneider Subject: Re: [Issue 11068] Lifeline representing an actor Date: Fri, 15 Apr 2011 10:42:57 +0200 To: "Rouquette, Nicolas F (316A)" X-Mailer: Apple Mail (2.1084) X-MIME-Autoconverted: from quoted-printable to 8bit by amethyst.omg.org id p3F8cp5o003145 A somewhat "late" reply, but that issue did not let go of me In essence I wanted to know if it was syntactically legal to place an actor on a sequence diagram. If I understood your answers correctly, it is not legal. Below are some points of discussion that I started formulating back in June 2010. I am not sure if this will lead us anywhere, but I thought maybe you would like to comment. What I would be very much interested in is the reason why in UML 2.x , actors are not allowed on sequence diagrams any more (opposed to UML with version <= 1.5 ). Best Regards, Florian Schneider On 01.06.2010, at 00:34, Rouquette, Nicolas F (316A) wrote: > [....] > >> The fact that the same symbol is used to represent an actor in a use case diagram is not a formal barrier. > > This is the part that gets a bit slippery. What do you mean by "same symbol"? > > The original question came about the fact that a lifeline is a connectable element but an actor isn't. > How could the symbol of an actor then show up as the symbol that's meant to represent a connectable element? [....] So a lifeline has an association to a connectable element (that is called "represents") but it isn't one. I'd like to comment on that because I think Nicolas last question touches something essential. A kind of naive view of mine was that elements on diagrams are a symbolic representation for instances of metaclasses of the UML. Like this: A rectangle on a class diagram means that I am talking about an instance of the metaclass Class or Interface. A circle on a use case diagram means that I am talking about an instance of the metaclass UseCase. These are well-known terms to me: UseCase, Class, Interface.... Now Lifeline feels like something more artificial (and I am not talking of connectable elements yet....). I will not go to a customer and deliver a Lifeline to him. I don't even wanna talk to him about lifelines but maybe about classes that are interacting. I see one problem there: The notation for lifeline (especially in communication diagrams) is very similar to that of an instance specification, and because of the rectangular head a little bit similar to that of classes and interfaces. This might lead us to the observation that the lifeline reuses the notation of the 'thing' it represents. I think the reasons for this might be the following 1) the vagueness of the specification (it does not say 'you have to use it that way in order to be UML 2.3 compliant') 2) I suppose that the standard user of UML doesn't want to delve into the details of the metamodel. So the naive conclusion might be that a lifeline representing actors would reuse the actor symbol. Now please don't jump in and say "well you have to read the spec thoroughly, a lifeline never represents an actor, but a connectable element". I know about that:-) Let's consider UML as a framework. Would you as a framework user be interested in the inner workings of the framework? No, you would just ask about the API and its documentation. If you call a function you expect that it returns the behavior that was specified Let's say I am a standard user of UML and want to draw a sequence diagram. I am not interested in connectable elements. I want to show how two objects (that are instances of a certain class) interact. Date: Fri, 15 Apr 2011 10:22:18 +0100 From: Dave Hawkins User-Agent: Thunderbird 2.0.0.21 (X11/20090302) To: Florian Schneider CC: "Rouquette, Nicolas F (316A)" , Bran Selic , Østein Haugen , "uml2-rtf@omg.org" Subject: Re: [Issue 11068] Lifeline representing an actor X-Source-IP: acsmt356.oracle.com [141.146.40.156] X-Auth-Type: Internal IP X-CT-RefId: str=0001.0A090209.4DA80E8B.0029:SCFSTAT5015188,ss=1,fgs=0 Florian, I only vaguely remember the previous discussion. However from what you've quoted, "...a lifeline is a connectable element but an actor isn't." A lifeline isn't a connectable element either. The only things that are connectable elements are variables, parameters, properties, ports and extension ends. The lifeline represents a particular instance, the type of which is the type of the connectable element. One way to use interactions is to create a collaboration where the parts are properties typed by the interacting classifiers. I don't believe there's a restriction on what those classifiers can be, so actors are perfectly acceptable. In fact the use case description contains the following: "It may...be described indirectly through a Collaboration that uses the use case and its actors as the classifiers that type its parts." Which effectively means that lifelines can represent instances of actors, when you have an interaction owned by that collaboration. Cheers, Dave Florian Schneider wrote: A somewhat "late" reply, but that issue did not let go of me In essence I wanted to know if it was syntactically legal to place an actor on a sequence diagram. If I understood your answers correctly, it is not legal. Below are some points of discussion that I started formulating back in June 2010. I am not sure if this will lead us anywhere, but I thought maybe you would like to comment. What I would be very much interested in is the reason why in UML 2.x , actors are not allowed on sequence diagrams any more (opposed to UML with version <= 1.5 ). Best Regards, Florian Schneider On 01.06.2010, at 00:34, Rouquette, Nicolas F (316A) wrote: [....] The fact that the same symbol is used to represent an actor in a use case diagram is not a formal barrier. This is the part that gets a bit slippery. What do you mean by "same symbol"? The original question came about the fact that a lifeline is a connectable element but an actor isn't. How could the symbol of an actor then show up as the symbol that's meant to represent a connectable element? [....] So a lifeline has an association to a connectable element (that is called "represents") but it isn't one. I'd like to comment on that because I think Nicolas last question touches something essential. A kind of naive view of mine was that elements on diagrams are a symbolic representation for instances of metaclasses of the UML. Like this: A rectangle on a class diagram means that I am talking about an instance of the metaclass Class or Interface. A circle on a use case diagram means that I am talking about an instance of the metaclass UseCase. These are well-known terms to me: UseCase, Class, Interface.... Now Lifeline feels like something more artificial (and I am not talking of connectable elements yet....). I will not go to a customer and deliver a Lifeline to him. I don't even wanna talk to him about lifelines but maybe about classes that are interacting. I see one problem there: The notation for lifeline (especially in communication diagrams) is very similar to that of an instance specification, and because of the rectangular head a little bit similar to that of classes and interfaces. This might lead us to the observation that the lifeline reuses the notation of the 'thing' it represents. I think the reasons for this might be the following 1) the vagueness of the specification (it does not say 'you have to use it that way in order to be UML 2.3 compliant') 2) I suppose that the standard user of UML doesn't want to delve into the details of the metamodel. So the naive conclusion might be that a lifeline representing actors would reuse the actor symbol. Now please don't jump in and say "well you have to read the spec thoroughly, a lifeline never represents an actor, but a connectable element". I know about that:-) Let's consider UML as a framework. Would you as a framework user be interested in the inner workings of the framework? No, you would just ask about the API and its documentation. If you call a function you expect that it returns the behavior that was specified Let's say I am a standard user of UML and want to draw a sequence diagram. I am not interested in connectable elements. I want to show how two objects (that are instances of a certain class) interact. -- Dave Hawkins | Principal Software Engineer | +44 118 924 0022 Oracle JDeveloper Development Oracle Corporation UK Ltd is a company incorporated in England & Wales. Company Reg. No. 1782505. Reg. office: Oracle Parkway, Thames Valley Park, Reading RG6 1RA. Cc: "Rouquette, Nicolas F (316A)" , Bran Selic , Østein Haugen , "uml2-rtf@omg.org" From: Florian Schneider Subject: Re: [Issue 11068] Lifeline representing an actor Date: Fri, 15 Apr 2011 11:38:21 +0200 To: Dave Hawkins X-Mailer: Apple Mail (2.1084) X-MIME-Autoconverted: from quoted-printable to 8bit by amethyst.omg.org id p3F9Y6se007557 So when lifelines actually can represent instances of actors (given the circumstances that you just explained) and it is acceptable to use a different icon for the lifeline head - it also would be syntactically correct to use the actor symbol in sequence diagrams and this would in fact mean that the lifeline represents an instance of an actor. Correct ? In that case I do not understand why this issue is still open. I quote the original summary of the issue here: "It is common usage to model a lifeline in a interaction that represents an actor. I can't see how that could be done formally correct. A lifeline represents a connectable element, e.g. a property. It is not allowed to define a property that is typed by an actor." Cheers and thanks a lot, Florian On 15.04.2011, at 11:22, Dave Hawkins wrote: > [...] > > The lifeline represents a particular instance, the type of which is > the type of the connectable element. One way to use interactions is > to create a collaboration where the parts are properties typed by > the interacting classifiers. I don't believe there's a restriction > on what those classifiers can be, so actors are perfectly acceptable. > In fact the use case description contains the following: > > "It may...be described indirectly through a Collaboration that uses > the use case and its actors as the classifiers that type its parts." > > Which effectively means that lifelines can represent instances of > actors, when you have an interaction owned by that collaboration. > [...] Date: Fri, 15 Apr 2011 10:54:12 +0100 From: Dave Hawkins User-Agent: Thunderbird 2.0.0.21 (X11/20090302) To: Florian Schneider CC: "Rouquette, Nicolas F (316A)" , Bran Selic , Østein Haugen , "uml2-rtf@omg.org" Subject: Re: [Issue 11068] Lifeline representing an actor X-Source-IP: acsmt356.oracle.com [141.146.40.156] X-Auth-Type: Internal IP X-CT-RefId: str=0001.0A090208.4DA81604.0060:SCFSTAT5015188,ss=1,fgs=0 Tim would have to comment on why he thinks a property cannot be typed by an actor. I'd suggest the reason the issue is open is simply because it hasn't been closed as not an issue. The lifeline description contains the text: "The Lifeline head has a shape that is based on the classifier for the part that this lifeline represents." So, yes, the specification not only explicitly allows the lifeline shape to be an actor shape, it's actually the correct notation. Florian Schneider wrote: So when lifelines actually can represent instances of actors (given the circumstances that you just explained) and it is acceptable to use a different icon for the lifeline head - it also would be syntactically correct to use the actor symbol in sequence diagrams and this would in fact mean that the lifeline represents an instance of an actor. Correct ? In that case I do not understand why this issue is still open. I quote the original summary of the issue here: "It is common usage to model a lifeline in a interaction that represents an actor. I can't see how that could be done formally correct. A lifeline represents a connectable element, e.g. a property. It is not allowed to define a property that is typed by an actor." Cheers and thanks a lot, Florian On 15.04.2011, at 11:22, Dave Hawkins wrote: [...] The lifeline represents a particular instance, the type of which is the type of the connectable element. One way to use interactions is to create a collaboration where the parts are properties typed by the interacting classifiers. I don't believe there's a restriction on what those classifiers can be, so actors are perfectly acceptable. In fact the use case description contains the following: "It may...be described indirectly through a Collaboration that uses the use case and its actors as the classifiers that type its parts." Which effectively means that lifelines can represent instances of actors, when you have an interaction owned by that collaboration. [...] -- Dave Hawkins | Principal Software Engineer | +44 118 924 0022 Oracle JDeveloper Development Oracle Corporation UK Ltd is a company incorporated in England & Wales. Company Reg. No. 1782505. Reg. office: Oracle Parkway, Thames Valley Park, Reading RG6 1RA. From: Østein Haugen To: Dave Hawkins , Florian Schneider CC: "Rouquette, Nicolas F (316A)" , "Bran Selic" , "uml2-rtf@omg.org" , Østein Haugen Date: Fri, 15 Apr 2011 13:05:23 +0200 Subject: RE: [Issue 11068] Lifeline representing an actor Thread-Topic: [Issue 11068] Lifeline representing an actor Thread-Index: Acv7Uz/3K4wE3/Z/Qziu/4vQhrcgQAAB7ujA Accept-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: acceptlanguage: en-US X-MIME-Autoconverted: from quoted-printable to 8bit by amethyst.omg.org id p3FB1BVw007457 There seem to be several separate issues here all tangled together. 1) Actors as such cannot be represented by a lifeline. Lifelines represent ConnectableElements. This is important because there have been discussions over the years whether Lifelines should also be allowed (like they were misused to be in UML 1) to represent to Classifiers, not only Actor, but also Class or Collaborations. It is possible to enhance the semantics of lifelines and interactions to achieve this, but it has been decided until now that this should not be done. 2) There may be ConnectableElements (e.g. Properties in Collaborations) that are typed by Actors, and they may be represented by Lifelines. Yes, this is true, but I am not sure how frequent such Collaborations are used. The example in UML 2.4 Superstructure standard Figure 16.8 is as far as I can see hardly a Collaboration diagram. 3) May the Actor symbol be used on a lifeline? The original intention was that it was not allowed to change the shape of the lifeline head. The liberal interpretation quoted below allows it, but I am not so sure of the positive effect. As a summary I believe this issue originated long time ago as an issue in category 1 above. Please enlighten me if Use Case users have started using Collaborations with Actor properties. I still believe this is normally not the case, but I may be mistaken in this. Finally, we are not talking about "instances" here, we are talking about "properties" and "connectable elements". Lifelines cannot represent Instance Specifications. /Oystein -----Original Message----- From: Dave Hawkins [mailto:dave.hawkins@oracle.com] Sent: 15. april 2011 11:54 To: Florian Schneider Cc: Rouquette, Nicolas F (316A); Bran Selic; Østein Haugen; uml2-rtf@omg.org Subject: Re: [Issue 11068] Lifeline representing an actor Tim would have to comment on why he thinks a property cannot be typed by an actor. I'd suggest the reason the issue is open is simply because it hasn't been closed as not an issue. The lifeline description contains the text: "The Lifeline head has a shape that is based on the classifier for the part that this lifeline represents." So, yes, the specification not only explicitly allows the lifeline shape to be an actor shape, it's actually the correct notation. Florian Schneider wrote: > So when lifelines actually can represent instances of actors (given the circumstances that you just explained) and it is acceptable to use a different icon for the lifeline head - it also would be syntactically correct to use the actor symbol in sequence diagrams and this would in fact mean that the lifeline represents an instance of an actor. Correct ? > > In that case I do not understand why this issue is still open. > > I quote the original summary of the issue here: > "It is common usage to model a lifeline in a interaction that represents an actor. I can't see how that could be done formally correct. A lifeline represents a connectable element, e.g. a property. It is not allowed to define a property that is typed by an actor." > > Cheers and thanks a lot, > Florian > > On 15.04.2011, at 11:22, Dave Hawkins wrote: > >> [...] >> >> The lifeline represents a particular instance, the type of which is >> the type of the connectable element. One way to use interactions is >> to create a collaboration where the parts are properties typed by >> the interacting classifiers. I don't believe there's a restriction >> on what those classifiers can be, so actors are perfectly acceptable. >> In fact the use case description contains the following: >> >> "It may...be described indirectly through a Collaboration that uses >> the use case and its actors as the classifiers that type its parts." >> >> Which effectively means that lifelines can represent instances of >> actors, when you have an interaction owned by that collaboration. >> [...] > -- Dave Hawkins | Principal Software Engineer | +44 118 924 0022 Oracle JDeveloper Development Oracle Corporation UK Ltd is a company incorporated in England & Wales. Company Reg. No. 1782505. Reg. office: Oracle Parkway, Thames Valley Park, Reading RG6 1RA. Subject: RE: [Issue 11068] Lifeline representing an actor Date: Fri, 15 Apr 2011 13:14:32 +0200 X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: [Issue 11068] Lifeline representing an actor thread-index: Acv7U4EOBHs35BCbSMiQbqgXiYgdUAACeAqg From: "Tim Weilkiens" To: "Dave Hawkins" , "Florian Schneider" Cc: "Rouquette, Nicolas F (316A)" , "Bran Selic" , Østein Haugen , X-MIME-Autoconverted: from quoted-printable to 8bit by amethyst.omg.org id p3FBAIKO008232 Some time passed since I filed that issue. I can't remember if I found a constraint or some other thing in the specification that caused the issue. Today I know that a lifeline could represent a connectable element that is a property typed by an actor. Typically I model a class that has association relationships to actors and other elements to be able to - model them in an interaction diagram - model their structure in a composite structure diagram. I know that collaboration would be the better choice according to the UML formalism. However in (my) practice collaborations are not so common and often unknown. For communication purposes it is easier for me to use a class. The class isn't for code generation, but a pure model element. I name and stereotype it with context, e.g. a <>. Best regards, Tim -----Original Message----- From: Dave Hawkins [mailto:dave.hawkins@oracle.com] Sent: Friday, April 15, 2011 11:54 AM To: Florian Schneider Cc: Rouquette, Nicolas F (316A); Bran Selic; Østein Haugen; uml2-rtf@omg.org Subject: Re: [Issue 11068] Lifeline representing an actor Tim would have to comment on why he thinks a property cannot be typed by an actor. I'd suggest the reason the issue is open is simply because it hasn't been closed as not an issue. The lifeline description contains the text: "The Lifeline head has a shape that is based on the classifier for the part that this lifeline represents." So, yes, the specification not only explicitly allows the lifeline shape to be an actor shape, it's actually the correct notation. Florian Schneider wrote: > So when lifelines actually can represent instances of actors (given the circumstances that you just explained) and it is acceptable to use a different icon for the lifeline head - it also would be syntactically correct to use the actor symbol in sequence diagrams and this would in fact mean that the lifeline represents an instance of an actor. Correct ? > > In that case I do not understand why this issue is still open. > > I quote the original summary of the issue here: > "It is common usage to model a lifeline in a interaction that represents an actor. I can't see how that could be done formally correct. A lifeline represents a connectable element, e.g. a property. It is not allowed to define a property that is typed by an actor." > > Cheers and thanks a lot, > Florian > > On 15.04.2011, at 11:22, Dave Hawkins wrote: > >> [...] >> >> The lifeline represents a particular instance, the type of which is >> the type of the connectable element. One way to use interactions is >> to create a collaboration where the parts are properties typed by >> the interacting classifiers. I don't believe there's a restriction >> on what those classifiers can be, so actors are perfectly acceptable. >> In fact the use case description contains the following: >> >> "It may...be described indirectly through a Collaboration that uses >> the use case and its actors as the classifiers that type its parts." >> >> Which effectively means that lifelines can represent instances of >> actors, when you have an interaction owned by that collaboration. >> [...] > -- Dave Hawkins | Principal Software Engineer | +44 118 924 0022 Oracle JDeveloper Development Oracle Corporation UK Ltd is a company incorporated in England & Wales. Company Reg. No. 1782505. Reg. office: Oracle Parkway, Thames Valley Park, Reading RG6 1RA.