What's in JAVAEE?

Is Java EE 5 the simplest version in the history of Java EE, as Sun said, which will greatly reduce the development difficulty? After the advent of Java EE 5, how should enterprises choose? Will enterprises that use J2EE 1.4 or earlier upgrade to Java EE 5?

Java is 1 1 year old this year. In this 1 1 year, it has been applied to enterprises, desktops, Web, mobile and other fields. Among them, J2EE and its related products are most widely used in enterprises. However, from the first J2EE version of 1999 to J2EE 1.4, people have been complaining constantly, not because it is not powerful enough, on the contrary, it is because it is too powerful to use.

Over the years, Sun has been trying to change all this, but he has never been able to do so. After the introduction of J2EE 1.4 in 2002, its complexity reached its peak. Especially EJB2.0, it is very difficult to develop and debug. Perhaps Sun was determined to change all this, or perhaps it was forced by the market and the pressure of developers, and finally released the specification of J2EE5 (now renamed Java EE 5) in May this year, and declared that Java EE 5 would be the simplest version in the history of Java EE, which greatly reduced the development difficulty. But the industry is skeptical about this. Has Java EE 5 really become simple? This seems a bit sudden! Can Java EE embark on a simple and bright road in the future?

Java EE 5 was born

Java EE 5 was finally born in a long-awaited way. At this year's JavaOne conference, Sun specially sent a propaganda group to give a big blow to Java EE 5, saying that the biggest change of Java EE 5 is to simplify development. It is also mentioned that EJB3.0 can improve the efficiency of development and testing through the latest comments and other features provided by Java. Sun's advocacy has not completely dispersed, and the boos in the industry have come and gone. Richard Monson-Haefel, a senior analyst, said that although Java EE 5 has made many improvements, it still cannot overcome its size and complexity. This senior analyst is currently one of the designers of EJB3.0 specification. From the mouth of the designer of Java EE 5, I think it must carry a lot of weight.

Maybe Sun doesn't want to hear such comments. However, Monson-Haefel is still talking about various problems of Java EE 5.

Monson-Haefel said: "The functions in Java EE are highly integrated, and we will not be able to use some of them alone. Moreover, many developers of Servlet, Java databases and JavaServer pages generally only use some related features of Java EE. However, the Java EE specification requires that all features [APIs] must be used. In order to make the program run normally, these developers have to build a complex project to meet these requirements. This is because Java EE still maintains the programming mode in the late 1990s, which means that Java EE is API-centric. Monson-Haefel even alarmingly said, "The application of Java EE has come to an end! " .

Not everyone is as pessimistic as Monson Highfill. A JBoss spokesman said: "We don't totally disagree with Richard Monson-Haefel, a senior analyst at Burton Group, about the complexity of the platform, but no one agrees with his conclusion that the application of Java EE has come to an end."

IBM has always been a loyal supporter of Java and stood up for Java EE. Jim Knudson, as the Java EE architect of IBM, has publicly stated that "although the Java EE platform is somewhat complicated to use, it cannot be the reason for us to give up this platform".

As soon as Java EE 5 came out, it attracted so much discussion. Of course, these discussions are good and bad for Java EE 5. So how to choose the main user group of Java EE: enterprises?

Will enterprises pay for Java EE 5?

After the advent of Java EE 5, how should enterprises choose? Will enterprises that use J2EE 1.4 or earlier upgrade to Java EE 5?

Maybe Monson-Haefel's words will scare many enterprises. But Jim knudsen is still optimistic about this. Although he agrees that the Java EE platform is very complicated, he said that some complexity comes from the needs of users, and we should let users achieve the results they want. He said that IBM users who expect to realize SOA intend to build SOA on the Java EE 5 platform.

Ram Venkataraman, director of product management department of JBoss, also supports the idea of increasing the complexity of Java EE platform, because it is needed to handle various services, not only Ajax-style Web services that simply call relational databases, but also financial services that handle large-capacity transactions.

Maybe many business leaders will try Java EE 5 after listening to these praises. This is especially important for those enterprises that have built or will build SOA architecture. Because Java-based SOA is generally built on J2EE server, it forces enterprises to make a choice whether to upgrade to the new version of Java EE.

If the enterprise chooses Java EE 5, it may increase the subsequent cost. This is because Java EE 5 is based on Sun's latest JDK, so Java EE 5 has added many new APIs. Although these APIs seem to make Java EE 5 easier, developers must learn a new set of APIs to achieve programming model persistence. And these enterprises must increase the training cost for this.

In fact, Sun's original intention is also good, but it is not appropriate to simplify Java EE. Perhaps Sun still doesn't understand that blindly adding functions can't effectively simplify the use of Java EE, and sometimes it may be counterproductive. If Sun can add a WYSIWYG drag-and-drop system for development, it will really simplify development. Adding a lot of APIs like this, or even changing the thinking mode of programming, can't make the Java EE platform simpler, just change the medicine.

Since Sun handed over the design of Java EE platform, he has also been responsible for the increasing complexity of Java EE. As a member of JCP, Monson-Haefel and all members of JCP should be responsible for what Java EE 5 has become today. In fact, JCP members have been discussing this issue with Sun's Java EE 5 development team for a long time, but Sun is not active. In this way, both Sun and Sun should be responsible for this.

The director of JCP, Onno Kluyt, has a different view: "The complexity seen by Monson-Haefel is not the responsibility of JCP. JCP is a standard-setting organization, not an organization responsible for optimization, and should not be responsible for the complexity of Java EE 5. " At the same time, Kluyt also mentioned that the best way to make Java EE 5 easier is to have a good IDE. Simplicity is the concern of every JSR and its central task. However, Java EE developers don't know how to do it. After all, whether Java EE is simple or not, its developers have no say, and those who use Java EE have the most say. "

Java EE: where is the way?

Maybe the chatter above makes our heads bigger, but we can leave it alone. Finally, we have to let Sun make the final conclusion. Finally, Karen Padir, Sun's vice president in charge of enterprise Java platform, made a moderate exposition on Java EE 5: "The platform is of course complicated because it is the specification of application server. Moreover, the application server is very complex, because complexity exists mainly for developers to work on the application server, but application developers do not necessarily involve complexity, because they only use the tools necessary to complete the work. Developers can ignore the complexity they don't need. "

Just like Microsoft's Office, although it is complicated, not everyone will involve all complicated functions. So Java EE is not as complicated as we thought. On the contrary, with the support of JDK5, the performance of some applications, such as Web services, will be greatly improved. Therefore, if enterprises want to build new applications (especially SOA), they can still consider Java EE 5. For those systems that have been used for many years, be careful when upgrading. But overall, Java EE 5 is still very good.