Why there is no real rapid development platform in the Java World

Xdoclet

1. Project management: You can use Maven or ant, you can use ant for smaller projects, and a lot of the content of project management can be considered using MAVEN.
2. Version management: Basically, CVS can be selected.
3. IDE: Needless to say, it's eclipse.
4. System Architecture: The architecture with spring as its core should be a better choice.
5. Performance layer: Struts or freemarker, this may be too selective, in addition to struts outside as if there is no such a broad user base.
6. Orm:hibernate should be the right choice now.
7. Auxiliary tools: Xdoclet, can be based on the needs of the project part of the content generation, reduce a lot of manual work, perhaps there should be more default extension.
8. Unit Test: JUnit or testng and so on, although many people say that JUnit has become less useful, but for the general project should still be sufficient.


The core is actually the most humble xdoclet.
Ant+eclipse+spring+hibernate+freemarker+xdoclet+css
When writing a list of Pojo attributes, add Hibernate configuration information, page configuration information, and spring configuration information, one-step generate all the underlying code; The skeleton of the system has been created for us, and a lot of tedious work that most of us don't want to do is done.

Comments:
I've thought about this a long time ago, and there are so many alternative technologies in the Java field, but I think that implementing a rapid development environment is not caused by too much optional technology, but with development thinking, how to identify a development idea (process), identify the various factors involved, and customize these factors, It is possible to achieve the goal of rapid development. To be blunt, we need a platform-independent model that describes the software (entity model, business model, interface process description, interface element description, etc.), according to which we can use the specific popular technologies (such as hibernate/spring) The functionality of the template is automatically generated. So, how to create a generic software description model, will be the most difficult part. In fact, in our development, we are already applying this kind of thought, such as we will use Velocity/xdoclet, etc. for template definition, but after all, with the specific technology, but not a higher abstract description.

Perhaps this is why MDA appears, and I believe the MDA approach will be the direction of future software development.

So, I think in software development, the so-called automatic generation, is only an auxiliary means, hide the software model, is the real cross-platform, cross-technology reusable parts.

Author
Yes, this is really a problem, from the engineering point of view, I was talking about the auxiliary code generation, and the system architecture still need to design, what method is not qualified. What I'm talking about, though, is that I've already implied a design approach: model-driven, because this is the biggest commonality of all the existing technologies, especially Java technology, which is now centered on the bean, and the bean is actually the implementation of the model. MDA, in turn, promotes a level of model drive, from the design level to the architecture level.