Issue 3999: Attributes obsolete in UML 1.3 (uml2-superstructure-ftf) Source: (, ) Nature: Uncategorized Issue Severity: Summary: the association between StructuralFeature and Classifier should be removed. Attributes can not describe more information than Associations/AssociationEnds can. Therefore it is obsolete and confuses the user of UML, which to choose when modeling. On page 3-40 in the UML 1.3 specification it says: "Note that an attribute is semantically equivalent to a composition association; however, the intent and usage is normally different." If the semantics are equivalent, then it is impossible to distinguish between them. There is no extra layer of meaning above the semantics layer that can distinguish between two things with equal semantics. Semantics is meaning. I think this sentence is contradictory. I have not been able to find out what the difference in "intent and usage" is. If this is defined, it will obviously make the semantics of the two different. To improve the readability of class diagrams when everything is associations, I propose that associations should be possible to represent as text in the compartment where attributes are written today. Resolution: see above Revised Text: Actions taken: October 25, 2000: received issue March 9, 2005: closed issue Discussion: The modeling of attributes and association ends has been unified in UML 2.0, so that this issue is no longer applicable. End of Annotations:===== Sender: maran@d1o67.telia.com Message-ID: <39F723AE.AA145495@telia.com> Date: Wed, 25 Oct 2000 20:17:18 +0200 From: Marcus Andersson <054185595@telia.com> X-Mailer: Mozilla 4.75 [en] (X11; U; Linux 2.2.14 i686) X-Accept-Language: en MIME-Version: 1.0 To: issues@omg.org Subject: Attributes obsolete in UML 1.3 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=us-ascii X-UIDL: :A9!!,`W!!3mZd9 Organization: LAAS-CNRS X-Mailer: Mozilla 4.7 [fr] (WinNT; I) X-Accept-Language: fr MIME-Version: 1.0 To: uml-rtf@emerald.omg.org CC: maran@d1o67.telia.com Subject: Re: issue 3999 -- UML RTF issue References: <4.2.0.58.20001025150349.00cd2ca0@emerald.omg.org> Content-Transfer-Encoding: 8bit Content-Type: text/plain; charset=iso-8859-1 X-UIDL: ^I?e9dFZd9RdLe9@]!"! Hi everyone, Juergen Boldt a icrit : > This is issue # 3999 Marcus Andersson, maran@d1o67.telia.com > Attributes obsolete in UML 1.3 [...] > On page 3-40 in the UML 1.3 specification it says: "Note that an > attribute is semantically equivalent to a composition association; > however, the intent and usage is normally different." I think Marcus has just addressed a quite essential point. My question would be: Shall we still agree for composions and attributes to be presented as semantically strictly equivalent ? My feeling is that attributes represent something strongly internal to objects (thence the common principle, not to use any public attribute). IMHO, the strong "internality" of private attributes and their restricted visibility scope make them something different from plain composition. Ideally no alien object should be able to interact with an other object's attribute without that letter object being able to directly control that interaction (I would suggest reflexive mechanisms to allow more transparency). Composition allows the "weak entity" (using old E/R vocabulary) more freedom. It could maybe be compared with the difference between a State Federation (the USA) and a Federal State (the Federal Republic of Germany). (American and German Citizens' Opinions are welcome :o) I agree the most popular OO languages don't genuinely support that attribute vison, but should UML be constrained by today's hipes ? (I think that last issue has been discussed earlier.) Regards Francois -- Francois Taiani http://www.laas.fr/~ftaiani Ph.D. Student http://www.cnrs.org LAAS-CNRS mailto:francois.taiani@laas.fr X-Sender: jwarmer@pop3.NL.net X-Mailer: QUALCOMM Windows Eudora Version 4.3.2 Date: Wed, 01 Nov 2000 11:21:31 +0100 To: Francois Taiani From: Jos Warmer Subject: Re: issue 3999 -- UML RTF issue Cc: uml-rtf@emerald.omg.org, maran@d1o67.telia.com In-Reply-To: <39F7F6AE.E0C2F267@laas.fr> References: <4.2.0.58.20001025150349.00cd2ca0@emerald.omg.org> Mime-Version: 1.0 Content-Transfer-Encoding: 8BIT Content-Type: text/plain; charset="iso-8859-1"; format=flowed X-UIDL: $V;!!%/0e9>e#e9A~kd9 Hmm, interesting. Can anyone make explicit which modelelements (at M1 level), are needed to replace one Attribute modelelement? Since attributes are features and associations or associationends are not, the above transformation seems far from obviouis to me. Jos At 11:17 AM 10/26/00 +0200, Francois Taiani wrote: Hi everyone, Juergen Boldt a icrit : > This is issue # 3999 Marcus Andersson, maran@d1o67.telia.com > Attributes obsolete in UML 1.3 [...] > On page 3-40 in the UML 1.3 specification it says: "Note that an > attribute is semantically equivalent to a composition association; > however, the intent and usage is normally different." I think Marcus has just addressed a quite essential point. My question would be: Shall we still agree for composions and attributes to be presented as semantically strictly equivalent ? My feeling is that attributes represent something strongly internal to objects (thence the common principle, not to use any public attribute). IMHO, the strong "internality" of private attributes and their restricted visibility scope make them something different from plain composition. Ideally no alien object should be able to interact with an other object's attribute without that letter object being able to directly control that interaction (I would suggest reflexive mechanisms to allow more transparency). Composition allows the "weak entity" (using old E/R vocabulary) more freedom. It could maybe be compared with the difference between a State Federation (the USA) and a Federal State (the Federal Republic of Germany). (American and German Citizens' Opinions are welcome :o) I agree the most popular OO languages don't genuinely support that attribute vison, but should UML be constrained by today's hipes ? (I think that last issue has been discussed earlier.) Regards Francois -- Francois Taiani http://www.laas.fr/~ftaiani Ph.D. Student http://www.cnrs.org LAAS-CNRS mailto:francois.taiani@laas.fr _____________________________________________________ Klasse Objecten tel : +31 (0)35 6037646 Chalonhof 153 fax : +31 (0)35 6037647 3762 CT Soest email : J.Warmer@klasse.nl The Netherlands internet: http://www.klasse.nl