Welcome to 2002 J2EE. The year 2001 has been a learning experience for all of us in the Java technology universe. The lesson has been a painful one - focus on the business problem and apply technology to ensure the right ROI. About a year ago, several folks were riding out the fantasy of paper money; options would change their entire lifestyle. They were going to take a year off and get back to work when they felt like it. Their requirements for cars and houses had taken on a whole new dimension. They were asking the world to excuse them while they kissed the sky. Then the bubble burst, and we're all back to basics again; living a fairly good life, but with no drastic changes. Last year, in anticipation of this revolution, the industry introduced several products to support it. These products are now on the shelves of discount stores.
For the past few years there has been the dream of a global enterprise driven by technology. That dream is still alive, but it has been tempered a bit by reality. The reality check has impacted all technology platforms, and Java is no exception. In retrospect, the partitioning of the Java platform into three editions has made it much more feasible for developers and architects to justify using the right combination of Java APIs for the right problem. Selecting the appropriate mix of Java features out of a monolithic platform would have been a formidable task in the current economic climate. In addition, the efforts of Sun, the IDE vendors, and the application server vendors have helped architects define how Java can be applied in their particular domain. In that sense, the Pet Store Demo was much needed in the Java industry.
Let's look at some of the facets of the J2EE platform that have been affected. The core of J2EE is the EJB object model. As I mentioned in my last editorial (JDJ, Vol. 6, issue 12), the J2EE application server market will suffer as a result of companies reevaluating their problem domain's requirements. Many applications don't require the full power of a heavy and complex middle tier. The Web layer, which uses the Web application server (as opposed to the J2EE application server) and goes through homegrown data access modules to the data source, is becoming an attractive alternative. JSPs and servlets have already gained a lot of momentum. Container providers such as Tomcat and iPlanet Server are becoming viable alternatives to J2EE application servers. Not to say that the EJB market has died, only that it's being evaluated more judiciously to make sure it's the right solution for the problem.
ERP connectivity is a large piece of the puzzle and the focus of applications has shifted from global information exchange to interaction with traditional EIS sources. For that reason, the Java community is looking at JCA to provide developers with the means to get to a variety of ERP sources. At the same time, ERPs are exposing their core functionality via Java APIs and Web services. In 2002, we should see a maturing of the JCA and alternative connectivity mechanisms.
