From The Editor

Do APIs Still Matter?

The wireless world is abuzz with the plethora of new technologies being thrust upon us on an almost daily basis. All of these new tools and "standards" (their word, not mine!) are perfectly valid options and many hold forth the promise of revolutionizing the ways in which we work and play. Unfortunately, the unifying theme behind much of this from a developer's perspective is simply: wireless technologies = more work and more problems!. It will soon not be nearly enough to simply offer catalogs or even eCommerce via your corporate Web site. The increasingly sophisticated end-user base may expect you to offer mobile commerce or wireless location and information services. Given the competitive nature of business, you can take it for granted that if your competitor's are offering a service, you will soon be asked to provide that same service Faced with the wide variety of choices, what's an IT staff to do?!?

Given the rapid pace of change and the limited number of technology workers available to implement these changes, it seems logical that the technologies that will "win out" in the end will offer the easiest transition from today's desktop and enterprise systems to tomorrow's mobile handsets and information devices. This is why I think that not nearly enough emphasis is being placed on the existing Application Programming Interfaces (APIs) that our entire digital world is built upon. While technologies such as WAP are getting the vast majority of exposure, the probability seems very high that in two years, the majority of wireless applications will be constructed using "legacy" technologies such as Java, Visual Basic, and Linux. Why? Because that's where the bulk of developers currently are. Companies like Microsoft and Sun Microsystems essentially hold hidden assets in the form of platform APIs that developers are locked into once they undertake a development project. Billions have been spent on retraining IT staffs onto these platforms and I just don't see managers authorizing additional expenditures on additional technologies until the all-important Return-On-Investment numbers justify it. Of course, I could also argue that APIs do not matter in the age of the Internet and Wireless Web. Markup languages, in particular XML, allow developers to isolate their data and applications from underlying platform specifics thus obviating the phenomenon of "platform lock-in". While this may be true, some degree of interactivity is often required beyond the capabilities offered by HTML and WAP which is why I think that APIs do still have a role to play.

The groundwork is currently being laid as Sun continues to lay out their Java Micro Edition strategy and Microsoft continues to expand their wireless offerings via Mobile Explorer (with support for Java and Visual Basic runtimes) and PocketPC. If you haven't noticed lately, Linux also continues to gain ground in this space. Naysayers point to it as a server operating system but keep in mind that the basic Linux kernel can run in as little as 300 KB of memory. Lernout and Hauspie has demoed a very cool Linux-based device capable of speech recognition and Samsung has also demoed the Yopy PDA running Linux on the StrongARM architecture. Where does this leave companies such as Apple and Palm Computing? I think the moves they make in the next 6-12 months may solidify their future. Clearly, Palm offers tremendous products that end users absolutely love. How will they parlay that strength into developer mindshare? It will take a great deal of ingenuity and strategic partnerships on their end (such as the recent agreement between Palm, Sun, and iPlanet) but with proprietary APIs, will developers climb abord en masse? It will be very interesting to watch, to say the least.

Developers always love hot, new technologies and I count myself among this crowd. However, the guys who write the checks love that bottom line more than anything else and will soon be forced to choose between older technologies retrofitted onto mobile devices and newer technologies designed from the ground up for mobile applications. If the shortest distance between today's enterprise applications and tomorrow's mobile workforce is indeed a straight line, it would seem that the technologies of choice may already be among us.