If business is war, then in this age of the digital economy the application development department must be the equivalent of the munitions factory. The trouble is - as is often the case in war - most of the weapons on hand at the beginning of the conflict are tired relics harking back to a bygone era.
Take a look at most of the entrepreneurial efforts that make up the digital economy, and you'll find that although the vast majority of them are startup companies using the Internet, most of them are still just modernising some sort of traditional business model, whether it's hawking books or selling steel.
And if you delve a little deeper into these efforts, you'll find that the companies in question tend to rely on applications written not for mainframes, but for either Microsoft's Windows, Linux, or Sun Microsystems' platforms.
That in itself is not a bad thing. After all, the primary goal for these companies has been to get to market as quickly as possible. For many of them, this meant that getting the application up and running was a lot more important than making sure it scaled to meet future business needs.
Today, many of the companies are more successful then they ever dreamed possible, but the cost of sustaining that success is starting to show as they struggle to find scarce developers to help them maintain and expand their operations.
Of course, the folks who have worked in application development for traditional businesses have always known about scalability issues. In fact, many of them tend to haughtily dismiss the efforts of the startup companies as naive attempts to rewrite the fundamental laws of physics and application development.
After all, the reason these companies continue to base their businesses on legacy applications is because they are convinced that mainframes are the only platform capable of efficiently running large-scale applications. And from their perspective, it's only a matter of time and effort before they Web-enable these existing systems to crush upstart rivals relying on immature platforms.
Alas, as is often the case in the early stages of any major conflict, both sides are laboring under some serious delusions. For starters, most of the legacy systems traditional businesses rely on are so old that any attempt to Web-enable them becomes an extremely complex undertaking. Just cleaning up all the data alone can take years.
It's no wonder that many traditional businesses have elected to spin off new units chartered with reinventing the company.
This avoids the ordeal of hauling the rest of the company into the new century.
The simple fact is if companies have to wait for their existing applications to become truly Web-enabled, the market will simply pass them by.
The only real asset these startup companies have is experience, especially when you consider that most of them seem to think that scalability is something you can bolt on to your application after the fact. In truth, many of those startup companies are now waking up to the fact that they need to rewrite their applications while the business simultaneously grows at 100 per cent every three months, now widely considered to be the actual length of a single Web year.
Just as many traditional companies are trying to bring 20-year-old legacy applications to the Web, many of the startup companies are trying to upgrade two- or three-year-old legacy applications to the next generation of Internet infrastructure.
Who will win? Well, history tells us that every civilisation that gained a technological edge over its rivals carried the field.
In the new economy, that means that if you don't know what an object-oriented application development framework is by now, you are about to be outflanked by your enemies.
So just remember, it's rarely the biggest or the fastest that wins the day; it's almost always the smartest.
Michael Vizard is editor in chief
of ARN's sister publication InfoWorld. Reach him at email@example.com