Still working on OCL. I came up with some more answers to the April 2005 exam sample question 4biii. I emailed it out to the others in our study group. I still have not gotten answers back from the others about my first submission. This second submission finishes the question. I was hoping to get some feedback before tomorrow and our study group session on formal methods. It seems that tomorrow will be another session that may not be as productive as I would like. Next week is my tern on Stuctured and OO analysis. I think that I will get to it on Tuesday. I may not be totally done but I bet I have some good stuff by next weekend.

I have another gripe about UML tools. I still think that they all suck. My current gripe is the number of connection points that you get on a class. If you have more than one association you should get more than one connection point and you don`t with most tools....Visual Paradigm, Rose, TogetherJ, Umbrello, Argo-UML, Dia to name a few. As far as I am concerned Dia is the best even though it still has this problem. Visual Paradigm is very busy, Umbrello not precise, Argo-UML slow and not precise, but Dia is just right except for the drawbacks that the others experience also.

UML tool support is severely lacking. Which brings me to the point of last post about OCL/UML/MDA. If the tools development stalls because of lack of revenue generated from sales a tools company will die. Tools are only valid if they get used. If they don`t get used then upward feature mobility is slow. Case tools etc. have been around for a long time. How is MDA going to make it if it is only used by a small number of people and no competitive advantage has been shown that generates enough interest to make this stuff mainstream. Another tree falls in the forest type reference.