Creative destruction at work
Netbook computers can tell you a lot about where information technology is headed in the not-too-distant future. The days of big system, multiyear efforts to design and build new government IT applications are largely behind us.
They’re small, light and not that powerful, but netbook computers can tell you a lot about where information technology is headed in the not-too-distant future. Bigger than a BlackBerry but a fraction of the size of the biggest, Cadillac-featured laptops, the netbook can run applications, such as Word or iTunes. But with ubiquitous connectivity to the cloud — what we were calling service-oriented architecture or software as a service a few years back — do their users need much besides a Web browser? Probably not.
Netbooks should tell us that the days of big system, multiyear efforts to design and build new government IT applications are largely behind us. Replacing them will be fast, nimble projects in which a confederated approach might produce the desired feature set. An example is mashup programs, which often run through the browser and offer data acquisition, manipulation and visualization capabilities similar to what would have required hundreds of thousands of lines of application-specific software code just a few years ago. This example underscores just how revolutionary the change has been in the process of software development that targets the browser. Government should take heed.
The days of having an application that handles one specific task are waning. Each semester, I ask my students if they use an e-mail client — Outlook, Thunderbird or something else — and consistently the number of affirmative responses is lower, and at this point, well under half. They use webmail. They live in the browser. Anything else seems superfluous. This is something federal IT managers should take to heart. When vendors require a client to be installed on every machine, they should have a solid answer as to why.
Building stand-alone applications is an enormous investment of time and effort. Federal agencies should instead use their mass and considerable budgetary muscle to promulgate standards and drive technical investment. If Uncle Sam wants a flavor of Extensible Markup Language optimized for its business processes, the “if you build it, they will come” saying likely will apply.
All this is not to say that big systems will not continue to exist. It’s just that the design and development activity will be increasingly focused on big systems of systems: protocols, design schemas and flexible, adaptable application program interfaces designed to move not only data but also information — or even better, knowledge. Synchronizing large databases, opening channels to collaboration and managing the security of interconnected systems will remain a part of the terrain — but the creation of lots of specific applications to be installed on each machine across an entire agency should not.
The business of government is being pushed to the browser, not just on the PC or Mac, but iPhone, Blackberry and, of course, netbook. We expect ubiquitous connectivity to our data on whatever platform is convenient to us. Dinosaurs beware. In this atmosphere of digital convergence, old-school application development geared to the PC is building for a bygone age.