Issue 7908: Design principles (uml2-rtf) Source: (, ) Nature: Clarification Severity: Critical Summary: have a general problem with the UML 2.0 specification. A graphical modelling language is essential for succesful software development. However the more I read about UML 2.0 the more I had the impression that UML 2.0 has not been developed with actual real-world software development in mind. Just to give one highlight of UML 2.0 is the merge relation between packages: The relation leads to bad designs and incomprehensible software systems, e.g. like like badly designed inheritance hierarchies etc. Especially consider the following case: a trifle change in the diagram (change the merge relationship into e.g. an access relationship) causes a tremendous amount of changes on the code and the configuration level. The only way to handle this is to forbid the merge relationship and to hope that nobody is blind enough to actually use it. Reading the manual, I stumbled over numerous similar issues. I'm sorry to say but I'm very disappointed with UML 2.0 as it is Resolution: Revised Text: Actions taken: November 10, 2004: received issue August 23, 2006: closed issue Discussion: One can argue for and against the thesis in this issue. However, it is, in essence, a critique that does not really indicate how the problem is to be fixed other than to redo all of UML 2.0 from the beginning. Since that falls outside the scope of an RTF and also of reality, this issue will be closed with no changes made. Disposition: Closed, no change End of Annotations:===== m: webmaster@omg.org Date: 10 Nov 2004 05:30:43 -0500 To: Subject: Issue/Bug Report -------------------------------------------------------------------------------- Name: Markus Flückiger Company: none mailFrom: markusflueckiger@yahoo.com Notification: Yes Specification: UML 2.0 Specification Section: Design Principles FormalNumber: ???? Version: 2.0 RevisionDate: 03/08/02 Page: 640 Nature: Clarification Severity: Critical HTTP User Agent: Mozilla/4.0 (compatible; MSIE 5.0; Windows 2000) Opera 6.01 [de] Description Hi I have a general problem with the UML 2.0 specification. A graphical modelling language is essential for succesful software development. However the more I read about UML 2.0 the more I had the impression that UML 2.0 has not been developed with actual real-world software development in mind. Just to give one highlight of UML 2.0 is the merge relation between packages: The relation leads to bad designs and incomprehensible software systems, e.g. like like badly designed inheritance hierarchies etc. Especially consider the following case: a trifle change in the diagram (change the merge relationship into e.g. an access relationship) causes a tremendous amount of changes on the code and the configuration level. The only way to handle this is to forbid the merge relationship and to hope that nobody is blind enough to actually use it. Reading the manual, I stumbled over numerous similar issues. I'm sorry to say but I'm very disappointed with UML 2.0 as it is