We don't spend a lot of time in IT thinking about a world without applications, but it's time we started.
At first blush, the idea seems reasonable enough. It's easy, for instance, to draw a picture on a whiteboard of a network cloud filled with services - some of them building on external services and providing for particular local needs, some of them simply purchased as they're used - and a set of clients interacting with them. Look more closely at that picture, though, and you'll probably find that some application thinking has crept in. Did you truly draw an application-less picture, or did you end up implying that the services are applications?
This is why the document-centred interface that was a part of IBM OS/2 was almost never used as the designers expected and why Apple's OpenDoc layer for Mac OS never took off. The intention was users would completely ignore the underlying applications - indeed, they would be applet bits, combining to get the job at hand done, perhaps never configured the same way twice. But we users kept thinking we were double-clicking a textual document to get our word processors to open. We stayed in the application mindset.
Organisations are striving to deploy service-oriented architectures. But if we don't figure out how to break the application mindset, we can't extract the value in designing and deploying services on networks.
One of my clients is wrestling with this issue and discovering that users aren't getting the message. Here's an example of what they proposed to the business people:
"Suppose that, when customers phone the call centre, instead of running them through a script, we listen to them first. They tell us about the problem. We open up what looks a lot like a blank page. It's connected to a service that offers suggestions that might help the call centre operator. As the problem is described, this service suggests what to ask, or where to steer the conversation. Eventually, we have most of the problem described.
"We've yet to ask any of the standard identification questions, but that's one of the suggestions. So now we say, 'Can I just get some information from you?' We start to capture things in the order the customer uses, and more suggestions appear; if the user gives us his telephone number, for instance, a service looks it up and suggests his address, so we just ask if this is where he lives.
"Meanwhile, other services have analysed the issue and perhaps offered a solution, estimated a service call time or suggested some other action that might be appropriate. Still other services might be offering our latest deals that might be applicable to the customer's needs."
With your own experience as a customer, you might think this is a pretty enlightened way to deal with somebody. As the business client, you want to ask things like, "Don't we need to verify who they are first?" and "How do we get their account number?" That's linear, process-driven application thinking in action.
Learning how to think without applications and how to bring the business along will be difficult. The payback in the marketplace will be immense. Each service will be able to grow organically: low-cost updates, and no more massive changeovers ever again, plus a very natural way to approach customers and suppliers. Mastery of application-less thinking is IT's next challenge.