When I read this article, it took me a while to figure out exactly what the difference between view-based and object-based models are. By the time I got through the article, I think I finally get it.

I think they may have something here. For many people, being able to visualize objects and their interactions helps them understand how to program much better. The question is whether having an "intermediary" layer to view a specific aspect of all classes is better than having one class representation, and being able to focus on the various aspects of that class.

Personally, while I have no data or backing other than gut feeling, I feel that both would be useful. I think that the primary interface should be descrbied like they do, so that each object looks like a singular graphical object. However, I think a mode allowing you to view a certain aspect of multiple objects (such as inheritance, containment, stack traces) is also useful in different situations. There are many ways to look at a problem, or a design, and I think different views are more useful, depending on what you are trying to accomplish.

And that's all I have to say about that.

^

|

|

hey, that's my line.

--Amanda

IMHO, I believe that the more "layers" that exist just serve to confuse the student, especially at the introductory level where abstractions are definitely (or were for me) the hardest part. Perhaps if that functionality was there, but not forced upon the student, I would find it the most useful.