Let's talk about the increasingly large UML2.0, but in the words of the three boss of UML, it is not enough. It is indeed a cattle, and a word said the essence of the UML language framework. MOF as a UML language core is a basic stable elementary language, so the expansion of UML is not so worried. Those of the Endless Profile are not so terrible. That is, this MOF in the Language level, can be self-describing, so that people have the bottom - how big, not, it is defined! :)
Ok, continue to say more and more full UML2.0, the support of business modeling is enhanced - the people developed by the enterprise application system laugh; the sequential map has been changed, increase the time mechanism - real-time system People laugh; improve the components - people who engage in software architecture laugh; the expansion mechanism is specified - the tool manufacturer laughs; UML no one is used - everyone laughs. Indeed, no one will use UML to use it, go to it, it is the way of using UML now. For different areas, it can be extended based on the UML, and then the subset of UML after the expansion is used, so that the user's use will be more convenient. Some people may ask: Is this not returned to the age of all kinds of people? My opinion on this question is that users of course need to be used by the most convenient language, but because these languages have a unified basis, many studies can do it on this unified basis, and achieve better support. Although there is more language from the surface, it is substantially different from the previous 100 pairs.
UML2.0 itself has many advantages, but there is a lot of disadvantages. The first step is a semantic inaccussion (although the OMG claims to be greatly improved, I think the improvement is not enough), such an inaccurate, what do you do? But there is no way, and this thing is really difficult to improve, and you should give an understanding of OMG. In addition, UML2.0 is very pale and weak to the definition of the view. I think the reason for Omg is the most important thing to think that the definition of the view should be the problem of development methods, not the language problem. From this aspect to think about it, OMG does not make sense. There are still many other shortcomings, and it will gradually appear during modeling.
Surrounded, let's talk about it next time.
