I've been thinking today, as I search for income-producing work. I know what kinds of problems I enjoy working on, and they tend to be computational problems. I don't find excitement in typical data-management problems whose main concern is the collection, storage and display of data.
Take Amazon, for example. The whole e-commerce process is just data management. The cool part of amazon is the suggestions it makes for you. That's computations. Likewise, bioinformatics, resource optimization, cryptanalysis and web indexing/searching represent way cool problem domains to me.
It seems to me that it's in the computation that real value can be added to a system. There are a bazilliion systems that can collect an e-commerce order, all of about roughly the same feature set. The "you might also be interested in..." feature is the value-add which might help you sell more books. Tracking trucks isn't difficult, but routing them efficiently can directly affect the bottom-line.
Reading the trade press, folks seem to be questioning the value of IT investments. I have a feeling it's because they aren't creating IT projects that actually add value. Replacing a PHP order-entry system with a Java order-entry system of pretty much the same core features, maybe with a more modern architecture or language just doesn't seem to make much sense to the bottom line.
It's like replacing the carpet in a bricks'n'mortar bookstore.
So, I think I'm arguing (and certainly self-servingly) is that organizations should keep their existing data management applications, even if in PHP, and find ways to fund value-add projects. The promise of XML is that these heterogeneous systems can interface. We've taught to fear stovepipe systems, but we can just relabel them as best of breed and focus on adding to systems, instead of whole-sale replacement. You need to find the computational aspects of your business and target them. That's where you'll differentiate yourself and affect the bottom-line.
Of course, folks who know me probably find this ironic, as I'm a chronic re-implementor.