Laptops And Mobile Users: Everything Old Is New Again

Although PDAs, communicators, and more exotic devices get all the press, it's long been clear that the prosaic laptop is still the workhorse of the mobile worker. This fact was reiterated last week at Giga's Mobile and Wireless Enterprise conference. When we surveyed attendees on what platforms they were using in their mobile applications, the laptop was the runaway leader. Rifle through the bags of traveling sales staff or peer into the front seats of police cruisers (if you're not careful, the latter opportunity could quickly follow on from attempting the first!) and more likely than not, it's a laptop that you'll find supporting their work.

This finding raises a number of fascinating questions: What are people using all those laptops for? Why aren't they replacing them with cheaper, lighter PDAs (and conversely, when will they do that?). And how are desktop applications being reconfigured for the mobile user on the laptop? Nor is the era of the laptop coming to an end. Even as PDAs grow in power and popularity, substituting for laptops in many scenarios, the laptop is being reinvented as the tablet PC (which in many ways represents the power of a laptop with the interaction style of a PDA) and the modular computer (which promises to deliver the power of a laptop in the form factor of a PDA).

At first glance the traditional laptop has a number of disadvantages compared to a PDA. It's large, and often heavy. It has a comparatively short battery life. The need to boot it up makes casual use or spontaneous data entry impractical, as well as eating into valuable battery reserves. (Although improvements in hibernation and standby have made strides in this area, we're still some way from the instant-on experience of PDAs). Keyboard input requires a lap to rest the laptop on, or some other surface, and no manufacturer has yet come up with a widely acceptable substitute for the mouse, although various kinds of pads, sticks, and balls all have their adherents.

Despite this list, we see no slowdown in the number of companies that are deploying laptops. In many cases, a full-size display is the only way to get the job done effectively: the task simply demands that much screen real estate to display the information that the user needs. In others, intensive data entry or direct manipulation of graphical elements make a keyboard and mouse the best option. It's also relatively easy to add (and later, swap) a wireless card to a laptop. By contrast, wirelessly enabling a PDA often involves either a clumsy external sled on a conventional PDA or, in the case of an integrated wireless PDA or communicator, a hard commitment at purchase time to a particular network technology and carrier.

In addition to those advantages, there's also the lure of minimal effort to remake a desktop application for mobile use, a requirement that is becoming surprisingly common. In some cases, the need may be to support telecommuters or mobile workers. In other cases, the development of a uniquely mobile application for a specific business purpose becomes the catalyst to mobilize other traditionally office-bound applications as mobile workers seek to reduce the need to return to base or even to completely untether themselves from office visits.

So if it must be a laptop, this raises the challenge of making a desktop application mobile. That last question, which at one time seemed to have a simple answer, has taken on a renewed complexity as well as urgency. There was a time when most applications were being developed as fat clients in a client/server model. Revamping could be as simple as adding a mobile database to a sufficiently powerful laptop. Synchronization requirements, while not always trivial, are at least a known problem. With more modern generations of applications having been built using good principles of layered architecture and then delivered with technologies such as Web servers and application servers, revamping those applications to the laptop is considerably more challenging. The applications may require a local Web server on the laptop, serving up an HTML interface from servlets or Active Server Pages, or even a local application server running EJBs.

If it's acceptable for the application to only be available when the network is, a relatively thin client, perhaps as little as a browser, might be enough. On the other hand, putting some functionality locally in a Web server might improve responsiveness as well as reduce network demands, perhaps even allowing some functions to be carried out when the network is unavailable.

At the other extreme, guaranteed application availability requires complete independence from the network, which could imply a laptop powerful enough to run the whole application locally--user interface, application server, database, and all. Although demanding, this is often still easier than either redesigning the application to repartition it for a laptop (which creates two code lines to maintain) or rewriting it for a PDA.

Even this fully-loaded laptop approach may not be enough in some cases. If an application depends on back-end logic in an ERP system or other interfaces to processes that can't be carried around on a laptop, it will still need a network connection to access those processes. In that case, the best that can be hoped for is to decouple the local application as much as possible, for example, by using asynchronous messages between the laptop and the back end.

Of course, we should be careful not to let the pendulum swing too far toward the laptop. All of the reasons that enterprises began adopting PDAs still apply, such as light weight, low cost, long battery life, and convenience. The underlying lesson in all of this is that good usability, good design, and good quality are all underpinned by the same philosophy: fitness for purpose. And there are a lot of purposes for which the laptop is still fittest.

Carl Zetie is VP of research at Giga Information Group.