Responding to Willink... ... I see this (Subsystem is derived from Class) as an OO methodological issue for constructing a PIM. The Package Diagram in UML...
Responding to Willink... ... That seems like a very slippery slope. How does one deal with performance constraints unless one of the options is to introduce ...
Responding to Lahman ... To me, a non-functional requirement is one for which there is no tool support and consequently its satisfaction is subject to a...
Dear all I have posted a revised draft of our "Methods of Behaviour Modelling" paper that discusses and compares different approaches to behaviour modelling...
Responding to Willink... ... We'll have to agree to disagree here. I have to go with the conventional notion of defining functional requirements in terms of ...
H.S. ... I tend to think that a PIM cannot be executable in an elaborationist process. At least that was my claim in the paper "MDA: The Vision with the Hole."...
... There are two things that I see as lacking in the paper. 1) A discussion of how race conditions between the multiple state machines of an object is...
Lee Thanks for the comments. ... It seems to me (although I am not an expert on this) that race conditions pertain to where and how events are serialised. My...
Responding to Ashley... ... There are OTS tools that will execute a PIM without any code generation architecture, not even a default debug architecture. Also,...
Responding to Ashley... ... That's true, but for the transformation engine to do that properly the PIM specification has to be unambiguous. I believe...
H.S. ... Do you mean "sufficient for correct behaviour" or "sufficient for correct performance"? If the latter, I fully agree. If the former, I am not sure. ...
H.S. ... The discussion in "MDA: The Vision with the Hole?" is predicated on the following alignment: Elaborationist <=> Structure Centric (i.e., as described...
... conditions pertain to where and how events are serialised. My view is that where and how events are serialised is a design (PSM) issue rather than a domain...
... that a specialised (child) version of a state-machine should have some form of "behavioural conformance" (e.g., Liskov substitutability) with the...
Lee ... Did I answer your concern in my reply to H.S.? ... OK. That's what I suspected you meant! Rgds Ashley [Non-text portions of this message have been...
Lee ... Yes, you are right. Under these circumstances, behavioural conformance is not a concern. I was getting confused with the situation where superclasses...
Responding to Ashley... ... Both. However, I think the first is the important one. One can always parameterize optimization through marking. (Though such...
Responding to Ashley... ... It is true that all the translationist approaches use state machines, but that is not a necessary condition. They are used because...
Hi ... I suggest looking at the Role Object pattern. http://st-www.cs.uiuc.edu/users/hanmer/PLoP-97/Proceedings/riehle.pdf An object does not have many state...
Hi ... I see this as one of the major philosophical difficulties in UML. If the PIM is ambiguous or non-deterministic, the predictability of systems is...
Responding to Willink... ... Ah, yes -- a trip down memory lane. Back in the S-M days we used to refer to this as subclass migration or role migration....
Responding to Willink... ... I'm not convinced. That's certainly true in detail because of the complexity of the spec and across profiles because of...
H.S. ... A while back I posed the following question on the Executable UML list: * * * * Suppose you have a PIM expressed in Executable UML comprising two...
Responding to Ashley... ... 3 would be correct. However, one cannot methodologically construct the PIM to do that within the eUML or even basic OOA/D profile...
H.S. ... <snip> ... <snip> ... I am not clear on what you saying. Are you saying that, if you do not mix behaviour access and knowledge access *and* you follow...
Responding to Ashley... ... Close, but not exactly. I am saying that one can construct the PIM to be logically deterministic in the sense that A will not...
H.S. ... OK. That's what I thought. ... At least as Ed expressed it in his post, I think he was talking about this being a PIM issue. In particular, he said: ...
Having problems with message search?
Fill out this form to ensure your group is one of the first to be migrated to the new message search system.
