Hi Marco - I agree with your inclination to do a domain analysis on the "entire" system. Very few legacy systems have a well-understood and effective overall...
... There are three approaches to this type of problem. a) You can assume that the interactions between the components are fixed, and merely model a specific...
Responding to Marco... ... Does this mean that there multiple versions of hardware that have basically the same functionality? For example, the hardware...
A question has come up about 'delayed events to self' in the context of developing a model compiler. Are these possible in executable UML (the book does not...
... They are completely possible. In fact, the microwave oven example makes use of delayed signals to self. Their behavior is exactly as you describe below. ...
One scenario we had considered was two delayed events to the same instance, expiring at the same time, one self directed, the other not. We thought the...
... Not at all. A domain is a "subject matter" (page 14), and a subsystem is used to partition a domain for manageability. The confusion comes into play with...
What is the line between a delayed event and an alarm which should be modeled explicitly as a different object? Is it whether delivery of the event can be...
... My questions are: Is the existing HW/SW- subsystem a domain, or does it span multiple domains? or is this your question? My default (not chosen) mode of...
Responding to A_k_vincent... ... Since Balcer addressed the other questions, I will poke at this one. I see two qualifications. The first is the more general...
One quasi-legitimate use for self-directed events is to break up long processing chains, but return control to the state machine to see if there is something...
Responding to Riemenschneider... ... But a subsystem is also a unique subject matter with well-defined boundaries or else one could not allocate requirements...
Responding to Bestler... ... OK. But if one assumes the rule that self-directed events always have priority over external events, then there isn't anything to...
... issue ... I really shouldn't respond to this, again, because this was done to death on the SMUG mailing list. ;-) What are your definitions are of "problem...
I would have thought it was obvious that a "domain" is a scope that *excludes* everything that you don't want to talk about while a "sub-system" is something...
Responding to Riemenschneider... ... The problem space is where the requirements are defined, implicitly or explicitly. It is the non-software environment...
... "sub-system" ... enthusiast ... discussion. ... I'm confused as to how this reply applies to the context of Executable UML. You seem to be focusing on the...
Responding to Bestler... ... Quite true. Like everything else in OO development, we abstract with a very flexible view of logical indivisibility. So the...
... It has nothing at all to do with implementation, but rather with the fact that boundaries are very fluid in many business environments. What is a...
... An additional thought that is even more relevant to executableuml: It is not reasonable to require a modeler to freeze the system boundary *before* they...
... build/buy ... Here's another definition question ... I define the system as the application domains, all of the service domains, and all of the realized...
... I'm reading this as saying that the conditional access module was defined as a subsystem in the set-top box application domain, and a new requirement came...
There have been several occasions in developing the architecture where we have said something like "we would never do/have never done X and can't think of a...
Responding to Bestler... ... The last sentence I agree with because the tools should not be invasive in the solution design; that's what using UML and abstract...
** Editor's note: The following post contains a reference to a specific tool...while we don't like tool discussions in this group, it's only used to...
Responding to Riemenschneider... ... Yes or no, depending on what the difference is between 'domain' and 'subsystem'. If a subsystem always has a "firewall"...
... In this case, you are saying that domain and subsystem are synonymous, so there is no need for the two terms. ... defined ... Then why not just make all...
Responding to Riemenschneider... ... Yes. ... I do. I don't use the term 'domain' at all expressly to avoid the confusion inherent in the overloading with...
... And this is exactly what is implied by "subsystem" in the "Object Lifecycles" book by Shlaer and Mellor, "Executable UML" book by Mellor and Balcer,...
This conference appears to be closely related to this group's purpose. -- MJB ... C A L L F O R P A P E R S ============================= The 5th OOPSLA...
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.
