Maximum PC

http://www.eweek.com/article2/0,1759,1629416,00.asp

So far, I have to re-read this since it goes over my head. In essence, it talks about how IBM has viewed the importance of UML to the tool space, and why Microsoft has a long battle ahead of them. I think it's interesting to note that Grady Booch - co-creator of UML - is working for IBM. In fact, I think IBM has done a pretty good job with the whole Application Life Cycle program. Borland jumped the boat as well, with the acquisition of Together, but Microsoft has yet to include one. Anyhow it's an interesting read for developers since it talks about what IBM has in store for us in terms of developing products that aids us in developing tools for the tool space.

I enjoy reading Grady Booch's articles and comments. He consistently provides his insights in a clear and frank manner. He is also interviewed in Software Development this month discussing MDA.

Regarding MS and their unwillingness to use UML, which this article really didn't discuss, I must say that I am not suprirsed - this is MS after all! I wish the examples of the shortcomings of UML would have been clearer, unfortunatly, from what I've seen the examples trying to explain the shortcomings are harder to understand than the system the examples describe. Not a very good start to convincing people that a second model architecture is needed. How their products turn out - who knows.

The problem with the Rational tool suite is that it is quite overwhelming. They must have more than fifty tools and the website doesn't do a very good job of showing you where to start and how the tools that might be appropriate for your business integrate together. Use x -> y -> z or x -> k -> j type of thing. This is the foothold that I think MS wants to climb, but why then need to dismiss UML, and create MSML to do it, is not clear to me.

I'm fairly comfortable with UML, so much so that I'm actually learning it right now (along with 4 other topics, hence why I'm so busy). UML does have its shortcomings, but I just don't see why MS needs to shoot off in a tangent. That's the problem with Microsoft, instead of participating in solving problems, they want to make more problems by coming out with their own, proprietary solution. I'm not saying that MSML won't work, it's just that UML is really a good tool.



I haven't messed around with Rational enough to know how overwhelming it is. If IBM can work on what you stated, Rational can be a very powerful tool for IBM. As per Microsoft, they will always do whatever it takes to lock you into place. Hence why MSML, UML is a good tool, and if you have a proper understand of OOP, UML is actually not that difficult to learn. Anyhow, I'm sure that AOP will further extend to UML such that AOP and OOP will be properly implemented in UML sooner or later. I don't think MS will embrace AOP for a long time, and by that time, they'd rather come out with another tool to piss on everyone elses parade.

Good question about AOP support in UML - I'm not sure how that works.

FYI, I just made up MSML. They probably call it something else, but I don't know what it is.