More thoughts while rereading the Executable UML book. In the section, 5.5.3 Attribute checks, I am wondering about the justification for "composite value"...
... What distinguishes an executable UML model from a mere data model is that these classes have lifecycles. Objects (instances of classes) progress through...
... The classic justification for avoiding multiplicity in analysis is that the problem is better stated with full normalization and without repeating...
... Not exactly. From a modeling perspective, there is no justification in dividing an atomic value into constituent pieces if the value is atomic in the...
Responding to Balcer... ... I agree with Riemenschneider that I don't like this _for this example_. It is not a simple domain. It combines two distinct...
... example ... address are ... structure ... Suppose other domains need to deal with Address? The UI obviously needs a way to enter in address, and the label...
... The problem I cited, an authorization set, is an example of this. Specifically, it is a type understood in the problem domain where both the composite and...
... OK, so this is just plain wrong. Sorry. Three of the addresses are distinct, customer, billing, and delivery. The fourth, Shipment.deliveryAddress, is...
... easily abused ... meaningful ... of ugly ... I looked at Date's An Introduction to Database Systems chapter on "nulls". He discusses two approaches, 3...
... problem: A ... separate ... Further clarification of my response is needed to really express my heartburn with composite value attributes. The reasoning...
Responding to Riemenschneider... ... But isn't this why knowledge attributes are ADTs? All we need to know at the higher level of abstraction is the ADT...
... must be ... In so ... Two things have occured to me from your response: 1) I'm approaching this from the wrong direction. There exists a real maintenance...
Responding to Riemenschneider... ... I still see it as a matter of abstraction. The semantics of the bridges that connect subsystems must be defined when the...
... Hate to reply to myself, but I'm just flat out wrong here. (Must be a reflection of my present work environment's influence on my thinking. ... System...
... Okay. I think a lot of my problem with this has been a syntatical one; not in the sense of translating syntax via a bridge, but rather in the syntax of...
Responding to Riemenschneider... ... Just to clarify, I just used String as a <familiar> example of a data structure where one might or might not care about...
... libraries. ... IOW, in specific implementation terms. ;-) I know that you know the difference I was pointing out. One point was that this argument goes...
Responding to Riemenschneider... I'll give odds we are talking past one another again... ... I /think/ we are on the same page except for the syntax issue....
... Not really, I'm being somewhat intentionally obtuse to glean further exploration. [SNIP] ... OK. I agree with this. As I mentioned before, legal operations...
Responding to Riemenschneider... ... Say, what? ... I don't think there are any restrictions on operators. For example, suppose the ADT is ComplexNumber. I...
... OK. I can see that, but that would require a mechanism for enabling this within the toolset. Right now, I believe the path in the action language would be...
Responding to Riemenschneider... ... Sure. In MDA terms, one needs a Marking Model, which most tools usually provide as pass-through tags on the model...
... [SNIP] ... I thought all of this was implied in "bridge call". ;-) We both know that separate domains can exist within the same process/library/whatever...
Responding to Riemenschneider... ... My point here was just that neither the model developer nor the transformation engine needs to think of it as a bridge....
... 3GL IDE I guess I didn't realize that we were dealing with 3GL IDEs. ;-) ... the ... property; ... The ... supposed to ... 'composite' is at a ... ...
Decided to change the subject and get a little more direct. :-) Shouldn't 'enumeration' and 'composite' be core data types in Executable UML? Definitions from...
Responding to Lee Riemenschneider on the subject of composites: The discussion of composite values on page 67 of the book suggests the following modelling...
... I'm not sure this is true. My reading of the book suggests that core data types *are* the basis for the definition of *all* domain specific types ...
Responding to Ring... Basically I agree with both messages. However, I always have to quibble over something... ... I didn't read it that way. I read it to...
... core data ... Yes. ... This is only implied by having enumerations and composites under the heading of "Domain-Specific Data Types", but my argument is...
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.
