Innocent RAD
2005-3-21
To be honest, RAD is very innocent. From the birth of the birth, it was awkward every day, "It's not to drag a control", just like the computer "is not inserting a mouse." Also in the prunity of the prequaturation, you will love it again and again, you will write the base code base. What string class, smart pointer, dynamic collection, you can write it yourself, if you use home, you Sorry to talk to people. Everyday repeated repeats, repeatedly repeatedly repeated repetitive re-repetition, but thought that they did how great, until one day, a uncle's uncle patted his shoulders: "Provincial province, we have no brick cover The building! "
The reuse is one of the cores of software development, from the function to the class, to the component, to the middleware, multiplexed scale gradually increase. The core of RAD is multiplexed. As for the control, it is actually a visual component.
The function is the earliest and the simplest multiplexing unit represents the language C. You must think that I repeat the basic concept of this Old is a waste of bandwidth, then a lot of programmers are still writing more than 1,000 single functions, is it to drag them out?
With the object-oriented concept, the class is as an object-oriented component, which represents language C , and has been widely accepted by the work of several years of MFC. With the revival of UNIX, the opposition has also been strong. The reason is very simple, and there are many people who have been able to do it. Don't say classes. Many people just piled up all the functions to be used to a class, and the result is not as good.
The component appears as the RAD concept, as a multiplex unit containing a set of functions together. Representative language delphi, not to say how many components provided by Delphi, but it provides a convenient way of implementing components, developers use a large number of well-reuse components created in this way. The componentization design of Borland also creates JavaBean, which also extends to DOTNET. If someone still doesn't understand, simply say: Rad is letting you create and use the reused components, not let the programmer "turbulence".
The middleware is a semi-finished product, which is a larger complex unit than the component, and no longer emphasizes what is achieved, and does not emphasize what kind of interface. Under this scale, general development work and design work has been separated by different full-time staff.
Compared to functions and classes, components, and middleware are more abstract, they are not crossed in some implementation form, and because of the large size of the problem, it is often necessary to accompany the relevant documentation to form a complete whole. Nothing in these ways is the best, which way to use is completely dependent on demand and problem scale.
Many radians believe that the use of components is far from the bottom, as if they can change the toilet to the bathroom, some people can find the large number, nothing more than some functions and classes, and what is related to the bottom layer. In fact, it is only VB does not support pointers, so most of the VB components only provide advanced interfaces, it is proposed all. There are also opponents who don't understand the underlying with a package. It is better to write it over again. If it is a special spirit, if it is a job, it is a nerve. It is also understood that the document and source code will not understand the bottom layer, but also waste life on this issue. So, some people use for (int * p = src, * q = dst, * e = src count; P Of course, in the territory of C does not mention the components, because the concept of a widely accepted component is based on classes. As the OO offers say, OO is just a concept. If it is a language that can do oo design is the language of the object, C is also facing the object, and the ADA is also. Similarly, components are also a concept, it is just a separately available module, because C is purely, achieving such a thing is easy in C . The reason for the C is difficult to develop components. The reasons that can be done must have interfaces that are easy to interface, while C no unified base library, and even a string of each class library. The components are the relatively high-level structure, which only receive a code frame, which can only be reused under the same framework, which greatly limits the development of components. This also gave birth to the interface independent of the language frame. Unfortunately, the boss of COM recognition is VB, C , if you want to make a look with COM, that is difficult. In general, many people are still in the old PC era. At that time, a person can write a WPS, and a person can write a CCDOS, and now these can only be Don Quixote's dream. The rise of mobile software makes many people to relive the old dream, but it is still going back to reality.
