One difference between UML1.4 and UML2 for state diagrams is the loss of the SimpleState and CompositeState element types from UML2.
In UML2 there is only a State metaclass for both simple state and composite state. That state has a boolean isComposite that returns true if the state has one or more Regions. For UML1.4 we have an abstract FigState with subclasses for simple state and composite state (plus others but I'll worry about them one I have these first two working). I'm busy working on on FigVertex which will replace FigState and its subclasses for UML2 (the old UML1.4 ones will remain untouched so that I don't break anything for UML1.4 in the UML2 work). FigVertex will alter its behaviour depending on the value of isComposite so that it can act like either of our current FigSimpleState and FigCompositeState classes. Regards Bob On 7 May 2011 14:08, Bob Tarling <[email protected]> wrote: > I had previously thought that both the UML2 activity and state > diagrams had changed so much that we need a complete reimplementation. > > I took a look this moring at trying to move my new state diagram > forward and realised things for that diagram are actually fairly > similar to UML1.4. We do not have to have a complete new diagram for > state diagrams. > > I have the original state diagram code in core ArgoUML now working > against UML2. Please give it a try. > > There are some thing I had to do by checking UML version within the > diagram code. Mainly this is for the following. > > 1. No internal transitions in UML2. > > I couldn't find anything for internal transitions in UML2. So I have > StateBodyNotation skipping any code related to this. See commit rev > 19369 > > 2. No top state > > In UML1.4 a Statemachine has a top State. In UML2 this does not seem to exist. > > I'm not sure if this is information I can infer from somewhere else > but for the moment I have worked around any calls to setTop/getTop. > These are all in the package org.argouml.uml.diagram.state see commit > rev 19370 > > I'd appreciate a review by anyone more familiar with this diagram type. > > This leaps us ahead a lot quicker with another UML2 diagram (which I > think will also fulfil our AndroMDA requirements is completed) > > Regards > > Bob > ------------------------------------------------------ http://argouml.tigris.org/ds/viewMessage.do?dsForumId=450&dsMessageId=2732955 To unsubscribe from this discussion, e-mail: [[email protected]]. To be allowed to post to the list contact the mailing list moderator, email: [[email protected]]
