The Future of Software: Apps to Die For

Service-oriented architectures may lead to the next transformative application, such as a hands-off tax return.

When it comes to what the next killer application might be, there's good news and bad.

First, the bad news: The next killer app is hard to find. There are lots of candidates but no clear winner. Often mentioned are existing tools such as search, speech recognition and autonomic security that are being refined constantly, but for which the potential benefits already are pretty clear. "The punch line is that the next killer apps are going to be new ways of doing old things," says Dan'l Lewin, VP of .Net business development at Microsoft.

Now the good news: The emerging software infrastructure known as service-oriented architecture is seen by researchers and IT executives alike as the foundation for a new generation of software components that, years from now, will top CIOs' lists of technologies they can't imagine living without. And that hints at an additional layer of benefits for IT execs: Those future components often will be developed from within IT departments, eliminating the licensing and quality issues that CIOs claim plague today's commercially available software products.

That means that companies out in front developing service-oriented architectures also may have a jump on building the next potentially transformative application component. Some will build components that solve universal business problems, but many more will simply and elegantly address processes specific to particular industries. For instance, a component that lets hotel guests check in from a taxi using their cell phones may be killer for the hotel industry, but it probably won't be of use to a tire manufacturer.

Conversely, an app that uses XML Web services to aggregate shipping options from DHL International, FedEx, and United Parcel Service might prove helpful for businesses across industries. The movement of parcels is a universal business need but not one that would be considered a core part of most businesses -- much like payroll, for which most companies rely on third-party service providers. "The next killer app may very well come from a large global enterprise that provides a set of services that are pretty horizontal in nature but haven't been deemed to be a mission-critical competitive advantage," Lewin says.

The shipping scenario is an example of a concept sometimes referred to as mega-aggregation, which has been an area of focus for researchers at MIT's Sloan School of Management. Professors and graduate students there theorize that the ability of Web services to build applications that can pull data from a variety of sources becomes much more compelling if there's some kind of agreement about the terms being used, a concept they've dubbed semantic, or ontology-based, integration. In other words, if a Web-services component has specific instructions on the descriptive parameters it must conform to when grabbing data, it increases the likelihood of delivering useful information.

If a person were to ask a financial-services mega-aggregator about his or her net worth, that mega-aggregator might return information that conforms to several descriptions. One site might supply real-time data, another might only update information every two days, and a third might include speculative data based on expected stock fluctuations. "The semantics could be so fouled up that it causes problems in your life," says Michael Siegel, principal research scientist at MIT's Sloan School. But if the mega-aggregator approached the task with instructions to search for actual net worth at a point in time, it could specify that to its data sources, returning an accurate snapshot.

Intuit Inc. is working toward a sort of killer app that borrows from these concepts. Within the next five to 10 years, the company plans to take nearly all the manual labor out of income-tax filing, says Scott Cook, founder and chairman. Future incarnations of the company's popular TurboTax software will use XML Web services to reach into a taxpayer's various financial accounts, pull the data needed to populate portions of a tax return, and combine it with W-2 data stored by the taxpayer's employer. It will then trigger an E-mail from the employer with the employee's completed tax form, requiring that the taxpayer merely click to approve the form before it's sent to the Internal Revenue Service. Intuit already is on the road to making this happen, having established connections into hundreds of financial institutions. "It's a combination of thoroughly connected and intimately personal with guard rails," Cook says.

Even an innovation that might appear specific to a particular business could hold implications for other industries. "It's as much about building capability as it is solving a particular problem," says Bill Godfrey, CIO of Dow Jones Inc. Godfrey's IT crew is building a nimble publishing system based on the Java 2 Enterprise Edition platform. While the system will solve publishing workflow issues, Godfrey says, it also will give Dow Jones the ability to publish variations on a theme. So the same piece of content described in different ways by XML tags could easily be routed to multiple publishing formats, from newsletters and wire services to desktop information products and yet-to-be-thought-of content-distribution methods.

Godfrey characterizes that capability as rethinking how Dow Jones reaches its subscribers. Similar advances could be in the offing for other content-driven industries such as entertainment or professional services. "The companies that can leverage the changes in the supply chain can gain an advantage," he says. "They're future-proofing their businesses."

One way Dow Jones is looking to future-proof its business is to work toward taking syndication one step further by delivering its information products in the context of its customers' work, putting the products directly into the applications they work in most frequently. A bond trader would access Dow Jones financial news from within a bond-trading system, rather than having to toggle to a separate desktop application. Engineering such change starts with resources, and Godfrey is looking to reduce the portion of Dow Jones' IT staff that's focused on legacy systems from 50 percent to 30 percent. That would translate to 20 percent more of his staff's time spent on strategic projects rather than caring for legacy technologies.

Part of what makes service-oriented architecture so promising is that it essentially dumbs down application development by enabling every application and data source to speak the same language, such as XML. That means less time spent on custom coding and more time spent solving business problems. It also means the next killer components could come from the unlikeliest of sources. "The barrier to entry for whomever it is that develops the next killer app will be greatly reduced," says Doug Heintzman, director of technical strategy for IBM's software group.

IBM is betting heavily on the future of Web services by undertaking a massive transformation in which its product code base will be deconstructed so that it can deliver functional components rather than packaged middleware apps. About 80% of IBM's code base will be separated into components by the end of 2005, Heintzman says. He expects that eventually, IBM will have lots of company as Web services continue their migration from the middleware layer to the application layer. "We're going to see the same market forces that are driving us to do that showing up in other places," he says, especially as core business systems such as enterprise resource planning, customer-relationship management and supply-chain management increasingly are broken down into assortments of components.

In the past, predicting a killer app was difficult, but building one was even harder. As the ability to build apps on the fly grows, trying to predict which ones will be transformative won't be easy. But whatever components emerge as the killer apps of the future, their arrival will be signaled by the same indicator: They'll alter the way business is conducted.