When asked if I would write an editorial, I jumped at the opportunity to express a growing concern I have for the direction of corporate Web development.

In today's market, upper management often selects an application server for the enterprise before knowing all their business needs. This one-size-fits-all attitude can set up many projects for failure. I'll use a race car-team analogy to make my point.

A successful race team has been racing a stock car for many years. Because of improvements in technology, the team knew they'd need a new race car. Upper management (who had never been to a race) purchased a dragster. The dragster was popular, and many teams were purchasing them. They read in a magazine that it was the fastest car you could buy. The race team now has a challenge: they have six months until the biggest race of the season, and they don't know what to do with a dragster. They find themselves sitting in it, looking at the oval track and wondering why they even wanted a new car in the first place. And now that they have it, can they learn to drive it and make it go in circles when it wasn't designed for such a race?

I've worked with several companies who've chosen WebSphere or WebLogic as their enterprise application server. These companies have a business that's driven by PowerBuilder and a staff who knows PowerBuilder. Now they must learn Java and n-tier development at the same time. In my opinion, these teams are set up for failure. They'll spend 18-24 months and millions of dollars learning a new technology, and when all is said and done they'll have an application that performs poorly due to a team that didn't understand the tool and n-tier development.

This problem could be fixed by using an application server that's right for the problem. If your project team consists of PowerBuilder developers, move your PB to EAServer. This allows the team to focus on one objective: n-tier development. After the team has successfully developed n-tier (distributed) applications, they can develop new components in Java or convert existing PB code to Java. The team may choose to write new components in Java, but they'll most likely never cost-justify converting PB to Java. Another reason that most corporations won't convert PB to Java is that once you're running in EAServer, why rewrite to run in a different server? EAServer has all the functionality that upper management is looking for.

Another area of concern is when upper management believes the hype that they should be choosing just one application server, and put all their eggs in one basket. It's unrealistic to think an enterprise will support only one application server since they do support a number of different databases. Every company I've ever been in has several databases. Why? Because the software they purchased runs on different databases. It's not an issue - all developers know SQL and can move between databases. This will be the same for application servers. Many companies we deal with today already have more than one application server.

The last point is that application servers are more likely to crash than database servers, simply because they're in their infancy. Even though upper management doesn't want to hear of it, the new technology does help the enterprise in a multitude of ways, but it is what it says it is: new. For this reason, large critical projects won't and shouldn't share application servers. Since the cost of the application server licenses is insignificant compared to the cost of the project development, a project team should make the choice of having the application run on its own application server. Choose the deployment application server to meet the needs of the project, not the other way around.

Don't let upper management set up your pit crew for failure. Point out simple facts: your team knows PowerBuilder, EAServer is an open solution that supports CORBA and EJBs, and you wouldn't be interested in running your perfect, bug-free application alongside some "always-down" application that continually crashes the server.

With these arguments, maybe you won't get stuck with a dragster that you don't know how to drive, that costs a ton, and performs poorly on an oval track.