Right now I'm still not sure whether Java will prove to be a major innovation or just a short-lived industry phenomenon, but I am certain of one thing: If Sun Microsystems doesn't start focusing its energy on linking Java to legacy systems, all the accolades won't be worth a hill of proverbial coffee beans.
A basic tenet of our industry is that customers will not abandon their existing tools and applications, along with the millions of dollars they have invested, just to be on the cutting edge of technological innovation.
Microsoft understands that fact all too well. Just about every product the company brings out will favour backward compatibility over innovation in an effort to entice users to upgrade. This is critical, because Microsoft's financial model requires that the bulk of its customers upgrade about every 18 months to compensate for the fact that Microsoft, unlike IBM, draws no significant revenue from maintenance contracts.
Granted, Java and development environments such as Java Workshop will probably be great tools when they are finally cooked, but I can almost guarantee that corporate IS sites are going to be a lot more interested in using their existing tools to generate Java code than they are in learning how to master new tools.
On top of that, IS sites are probably going to find middleware that promises to expose existing applications to Java-aware clients significantly more attractive than developing new Java applications.
Sun understands these basic issues and is beginning to work on the problem. Sun is also working on linking Java to SQL databases via ODBC, and IBM has a number of projects for linking Java to its vast array of legacy applications.
But all these efforts seem to be a much lower priority at Sun than establishing the brand equity of a new generation of Java development environments. To me, this approach is backward. To leverage our considerable existing investments, we should first focus on making Java compatible with what has gone before and then focus on a new generation of tools to optimise Java down the road.
So is Java an interesting exercise in computer science or a fundamental technological advance for IS shops?
(Write to me at email@example.com.)