- This topic has 1 reply, 2 voices, and was last updated 18 years, 8 months ago by
.
-
Posts
-
I am contemplating the use of the UML modeling feature of MyEclipse for an upcoming project. However, I’m having trouble gaining comfort with some aspects of the tool and would appreciate some feedback on my observations. I recognize that some of my discomfort comes from the fact that I developed my preferred modeling style with Rational Rose in the past, so I’m trying to sort through which of my issues are best resolved by changing my mindset and which might be candidates for requesting changes in the tool. For this thread, I will be working through some of my typical use case modeling scenarios to set up my comments and questions.
The first thing I like to do is to structure the model with some top-level packages. When the model repository is created initially, I am in the package-centric tool, which seems like what I would want so I lay out packages for the various views I expect to support. The packages are layed on the initial class diagram that I renamed “Model Structure” and includes a “Use Case View” package with subordinate “Use-Case Model” and “Business Use-Case Model” packages.
Comment 1: I can’t right-click on a package and create a subordinate package. Instead, I have to create the packages on the diagram and draw dependencies from the child packages to the parent package and then go to the Properties page for each of the child packages and change the Namespace property. I would think in package-centric view that I should be able to create child packages by right-clicking on a package in the Outline view.
Comment 2: I don’t like the way associations (in this case dependencies) are routed. I would prefer that they be routed between the centers of each object. It seems that associations are drawn from a finite set of anchor points for each type of object and I don’t like the placement of these anchor points; for instance, they will overlap part of the name of an actor.
Next, I wish to create a global view use case diagram for the Use-Case Model. To do this, I found that I needed to click on the Use-Case Model package in the Model Structure diagram and then click on the Use Case Diagram icon in the toolbar. I give this diagram the name “Global View of Actors and Use Cases”. This diagram may disappear at some point down the line as the model grows and it becomes overly crowded, but it’s a good starting point for discovery of actors and use cases.
Comment 3: Again, I tend to expect that I should be able to right-click on the package in the outline view and create a subordinate diagram. Or at least I should be able to have the package selected in the outline and then click on the toolbar icon, but the toolbar icons are not active when an object is selected in the Outline.
Now I’ve figured out that when a package is selected in the outline, I can click on a package icon in the properties for the package and it will create a child package. This is nice. I use this method to create packages for Actors and Use Cases.
Next I create the first set of actors and use cases on the global view diagram and draw associations. I select each of the actors and use cases and change the Namespace on their properties to put them in the desired packages.
Comment 4: Again, I wish I could right-click on a package or select a package and add an actor or use case without having to add them to a diagram. This would help with allocating them to packages right of the bat, but I guess this is okay.
Comment 5: The use case names are placed inside the use case icons. I tend to prefer having fixed-size ovals with the use case names below the icons, but that does not appear to be a style option for MyEclipse UML diagrams.
Comment 6: There does not appear to be a way to wrap use case names inside the icons. When I try to resize to force a wrap, the icon won’t shrink horizontally. Oh well. Long use case names are frowned upon, but they do occur from time to time.
Next, I like to create local view diagrams for each use case for documentation purposes. I select a use case in the global view diagram and click the Use Case Diagram icon in the toolbar and it creates a new diagram subordinate to the use case just as I like. I name the new diagram “Local View” and add any actors and use cases involved with the use case to the new diagram.
Comment 7: When I add the involved actors and use cases, the associations, includes, or extensions are not included. I know that it is not always desirable to include relationships on diagrams, but I prefer the option to suppress the relationships that automatically appear. I’ve never really bothered ot name relationships in use case diagrams, but this makes it difficult to get the relationships on the diagram. When I select the properties for the main use case in the local view, I see compartments showing the end points but I can’t drag these onto the diagram.
That’s all I have for now. Again, I would appreciate any feedback others might have for my comments.
