PowerBuilder is dead. I first heard that statement when I joined Sybase back in 1996, nearly a decade ago.

I had been hired to specialize in two of their hottest new Web development products, namely GainMomentum and web.SQL.... You remember those, right? Hmmmmm.

Here we are nearly a decade later and I sit in awe watching just one of our PowerBuilder/EAF development teams demonstrate a rich data-aware Web application they built in less time then most "modern" development projects spend putting together automated build environments and arguing the advantages of design patterns. I take a pause and think to myself, "PowerBuilder is dead?"

I have to admit I am a bit of an odd duck. I think I might hold the distinction of being just about the only person to have spoken at both JavaOne and Microsoft TechEd. I've seen the birth, and watched the death, of a dizzying array of development products and trends. Each was the sure-fire replacement of the legacy 4GL RAD environment. In the end, however, they have all missed the mark.

The 4GL RAD revolution wasn't fueled by the invention of a new approach to modeling or new language constructs. What fueled the revolution was simply productivity, and an understanding that what drives businesses to build business applications is to make data manipulation and display simple and fast. If something is going to be a viable alternative to the 4GL product genre such as PowerBuilder, it's going to have to truly deliver in a few simple but powerful ways.

First, it needs to make data binding dead simple. Let me say that again. It needs to make data binding dead simple. It should let developers discover data and bind that data to forms and tables just as easily as the DataWindow does. Contrary to popular belief, that does not mean providing developers with some "seven-layer burrito" of data abstraction layers and frameworks. It means "point and click."

I've heard the big boy analysts say that most Web development projects spend 80% of their time on data binding. Given the lack of sophistication of most Web technologies today, it frankly does not surprise me. PowerBuilder developers spend about 5% of their time doing data binding. Until a tool can claim the same level of productivity, it will never fully replace the 4GLs.

Next, it has to do the dummy work for me. I hate writing a hundred forms and data objects and mock objects and controller servlets and action classes and insert next big object du jour... These tools are supposed to make my life easier, not harder. Right? Any tool that wants to compete is going to have to be smart about building these objects for me...and in such a way that I can then edit and tweak where needed...just like what PowerBuilder allows me to do.

Finally these tools have to realize all of this data, all of these business rules, and all of these objects are irrelevant without a way to build serious application user interfaces. If you ever want to see a Web developer's eyes jump out of his head, tell him you need a treeview and a tab pane. Oh, and you want to update multiple queries in multiple panes and update them all at the same time and in the same transaction! Bring a towel. It's guaranteed someone is going to have to clean up when someone's head explodes. Applications have windows, frames, tabs, and trees, and they all depend on data views and they all depend on transactions. That's real life. Oh, and for good measure, I don't want to litter my whole application with this junk too. I want it componentized rather than copied into every page and I want it extensible.

What all of these self-proclaimed PowerBuilder "replacements" have forgotten is that productivity is king. They and, to a large degree, the developer community have become more enamored with elegance and design than function and delivery. Businesses chose PowerBuilder to get stuff done and make their businesses more successful. That was the only metric that mattered then and that is the only metric that matters now. That's the metric by which any competing tool must be judged and, unfortunately for most available technologies today, that is where they fail.

Today, PowerBuilder is still able to deliver on these key requirements of true rapid application development in both client/server as well as Web development with the Web DataWindow. And with future emphasis on .NET SmartClients and mobile development with PocketBuilder, PowerBuilder is anything but dead.