UML in practice, the current situation and some of the recommendations [2]

The fourth problem is the small number of programmers rely heavily on automatic code generation framework, which is in fact the key issue is not UML. In order to use certain tools they generated is not consistent with the code reading habits of Chinese people, willing to give up everything else. UML for them is just a code generation tool only, so they make the most of little value UML, more formal.

The fifth problem is that project resources. Most of the projects the domestic real resources available are very limited, if the foreign company to other developers to see whether the development cycle or human capital are basically ridiculous. UML should actually run through, not only in the initial design stage, it should run through the entire project team, including design, management, development, testing, everyone should use it.

I think in general the company if not sure do not have too much blind faith in UML / RUP, in fact some of the items do not need-based lightweight RUP or UML. UML is now intend to study with colleagues, it is also possible to first lay down the future, there will be more outstanding technology to replace it, now take the time to learn and practice as much about OO design, something more basic. In fact, just familiarize yourself with the tools and the exchange paradigm, the future for you, just a few days to a month or something, do not care much about. For those who now have decided UML actual project to be successful, my suggestion is as follows:

First, the project cycle at least plan to extend the usual 50% to 100%. You must follow all explained that the only way to make a meaningful UML itself, can effectively improve the quality of project design.

Second, the project began to focus for a period of UML development training. Consistently throughout the project, to be regular all-round exchange of object-oriented design (of course is a combination of UML), the attempt to improve design quality through the exchange.

Third, customers and the company leadership withstand the pressure to put as many design problems to solve in the initial design stage, rather than be forced to rush to complete formalism of UML design, and then cast aside the design and began coding.

Fourth, properly handle the part of the "elite" programmers resistance. To divert the flow, can not be diverted should be considered even subdue him completely out of the project group,. UML magnitude of team development does not encourage individual heroism, no less than cruel to simply not go forward.

Fifth, not too high expectations on the project. Eager acceptance of both the face of higher credit to the government clients, or companies only know how to act recklessly not understand technology leadership, must keep a low profile, it is not necessarily an immediate impact as an experiment ... ...

Well, to write so many, Good luck.







Recommended links:



My Favorite Covert Surveillance



windows media player Runtime error fixing tip



OGM TO AVI



ASF to MPG



DivX to iPod



How to "control" Yahuo sequels?



Photoshop Production - wire and spark



Collective wisdom light green hope



Five 8-word mantra! Speculative memory are Major contributors spit voice



LISTS Debugging



Evaluate Java And JavaScript



J2ME Game NORMAL-FRAME



The Basics Of Intrusion Detection Rules



Rotating Pear Tree Production Xiangjie