News

"I sing the segment wireless" or "Come in, remote apps!"

• By Jack Vaughan
• February 28, 2004

For too long, wireless computing has been "The Next Big Thing." Yes, Wi-Fi computing has been successful, but truly unique and working new remote computing applications are few.

There are good reasons why so-called mobile or wireless application development has been a slowly evolving discipline. Standards have never taken hold across the client pool as they did with the World Wide Web, which has meant too many choices. Meanwhile, people brave enough to make choices in this area have had to grapple with the usual blind alleys and false starts that any new technology invites.

Individual cell phone manufacturers, tablet PC software/hard makers or PDA manufacturers have had no compelling reason to give up their own takes on how to create mobile apps. Java and .NET are standards, sure. But in the mobile device world, unlike in the computer space, they are just some among a very wide set of standards.

You can choose assembler, C+, .NET, Symbian, Java, JavaScript and other tools to create such apps. Depending on your organization's experience and mind set, one or the other will seem the better path. For your app, you can strive for "always-on" operation or, based on a growing body of learning, go for a more asynchronous "don't-count-on-an-'always-on'" approach.

Few software companies have dedicated more of their resources to wireless computing than Sybase and its iAnywhere group. The company forged even deeper roots in that which is wireless last year when it purchased AvantGo, now part of iAnywhere, a company that was early to the Web wireless application show.

In the March issue, Application Development Trends talked with wireless application development leader Dr. John Halamka of Harvard Medical School about a remote application that could feed hospital case log notes, lecture notes, exam calendars and the like to the PDAs of medical students (see http://www.adtmag.com/article.asp?id=9014). Halamka used Sybase iAnywhere software on the project. We also spoke with Ojas Rege, formerly with AvantGo and now director of Mobile Solutions for Sybase's iAnywhere Solutions subsidiary. Few have a better view on the trends in this area. So what does Rege see?

Let's start with what he has not seen. "We haven't seen that one size fits all in mobile computing," said Rege. There has been a proliferation of development paradigms, he noted, and that has not been all bad since people don't want to be stuck with a single solution.

But a lot has had to be unlearned along the way.

"In the early days, the wide-area wireless [segment] was especially where expectations weren't met," said Rege. "There are some remote apps where wireless is essential, but in most cases being able to synchronize [at some point] is just fine. We say 'Understand your application. Don't get lost in technical details.' In 1999 and 2000, people were so focused on pure wireless that the applications didn't work or had performance issues."

Field service applications, little discussed when the wireless Web first raised its head, keep arising. They got there through trial and error. "People have field service applications and they may even have 95% [wireless area coverage], but the fact is that they have to work all the time. Where we have seen people do this effectively is where it can be dispatch driven," said Rege. The advice for pursuing Phase 1 of a wireless application: "Get it to work first," he noted.

Early -- circa 1997 -- attempts in this area saw developers trying to implement still-new Web apps in the yet-newer mobile environment, explained Rege. "Categories we see lot of use in over the last two years are on the CRM side, specifically for sales force automation," he added. Then there are field service inspection and field data collection apps. Meanwhile, he said, there has been an uptake in various areas of decision support or productivity enhancements. That involves Personal Information Managers (PIMs) and e-mail-oriented corporate portals.

Among the questions that those on the bleeding edge have had to try and answer is: If I am doing data collection, do I need a full database on the device? Am I going to be doing a lot of searches and retrieval? The answers to these questions will influence your technology decisions, said Rege.

The area continues to be unsettled. Through the brief history of mobile computing, application development planners have had to grapple with the fact that the platforms are very consumer oriented -- "gadgety," if you will. The short shelf life and fast turnover of consumer gadgets haunts this application space. Moreover, decreeing a specific type of device for a large pool of users is not often a favorable path.