Current Cover

Message Board

Mobile Strategies: Software Development

Don't panic. You may already know more than you thought about mobile software development.

From Pen Computing #7 December 1995

In the last issue we looked at a service company that used paper based forms to track its service crews, work orders, and customer accounts, and how it might use mobile computing technologies to improve its operation. The technology options we explored included pen computers and both wireline and wireless data communications. Each scenario allowed the service personnel to stay in contact with headquarters and the company to improve its crew tracking and dispatch as well as improve its billing cycle to customers. All of this applies to pen applications in government as well, so we decided to include the second installment of my "Mobile Strategies" manual in the government section. So let's get started.

In the last issue, we examined how a company or an agency might apply these technologies, but we did not discuss one important detail of the strategy. Just exactly how does one go mobile without any expertise in the mobile and wireless area? Doesn't that require lots of training and new development software? Don't panic! You may already have more knowledge than you think.

Today, there are indeed a number of pen-based environments in which to build your mobile application. However, if we look at the overall pen computer arena (excluding PDAs such as Newton and Magic Cap devices), most mobile applications are running under DOS or Microsoft Windows for Pen Computing, which is simply Windows 3.1 with Microsoft's pen extensions, so you don't necessarily have to start from scratch on an unfamiliar programming environment. Same with hardware. With the proliferation of more powerful microprocessor, 486s (and soon Pentiums) have found their way into the pen computing world. You no longer have to deal with odd machines with processors named after J.R.R Tolkien characters that don't run any of your software. And the newest breed of small disk drives and flash memory technology provide ample storage in mobile computers, so you don't have to spend a lot of time squeezing bytes out of programs.

All of this means that you can now build and deploy mobile pen applications with the same tools that you now use for client/server desktop computers. If you have an IS staff versed in Visual Basic you're off to a great head start.

Start small
Your first challenge is to decide on an application. My advice is to keep it small. A small-scale pilot aimed at a key area where mobile technology yields a good return on investment will win the hearts of your finance people. If you look at the worst case, a small failure is better that a big one. It is also important to find a strong sponsor among the users so that you have support. Such people are invaluable and can lend an almost religious fervor to keep the effort going.

Get the users involved from the very start. No only can they help you design a good application, but they can ultimately help you to sell it to others. Win them over and you have a powerful sales force.

Get your software developers in the field before they start. Have them spend a week looking over the process used now. Have them follow a real user and actually help, or do the job themselves. The knowledge they will gain will be important in developing a useful product.

Design considerations
Now that you have selected a candidate for a mobile application, you need to decide how to design the user interface. There are two design approaches for developing pen applications.

• The first is taking an existing application and making it pen-aware. This gets you off the mark quickly, but there is a big downside. Let's say you have this great building code inspection Windows application. It allows your inspectors to type violations into the computer while inspecting a facility. You decide to simply put this application on a pen computer and hand it to your inspectors. For a while, they will struggle trying to enter data by hand. Then they will complain that the handwriting recognition is not up to the job and demand their old laptops back.
• The second approach is a pen-centric design where applications and the processes they automate are completely redesigned with pen computers, pen operation, pen data entry, and mobile users in mind. This way, you get all the benefits of the pen as a fast and efficient pointing and editing device.

Obviously, the latter approach takes a bit more skill and experience, but it's not as difficult as you may think. As stated above, for most pen applications, you can now use the desktop development environments you're already familiar with. The real key to success is to start with an open mind.

The best applications I have seen take a fresh look at a problem rather than copying paper forms or desktop programs. They present the user with list boxes, scrolling lists, radio buttons, and check boxes to maximize the advantages of the pen and minimize the need for handwriting. Whenever there is a need for a user to enter text, they provide a small on-screen popup keyboard right where it is needed. You wouldn't want to write a novel on such a keyboard, but it's sufficient for brief comments. Applications may also allow the user to write on screen and save the input as digital ink. This method can also be used to capture signatures, say on your customer's approval of invoiced work.
When lists and number entry don't suffice, there are innovative alternatives such as number pads, wheels, or columns. A number pad could be your basic telephone touch pad, but don't forget the "-" or "/" symbols if they are needed. Number wheels allow the user to select each digit by spinning a small number wheel with arrows for forward and backward spinning. Number columns are simply a column of numbers from 0 to 9 which drop below the entry box for each digit. The user just taps the desired number in a column for each digit and the number appears in the entry field.

If you remain suspicious of pen-only computers, companies like Husky offers systems that include a keyboard. Norand addresses the numerical data entry issue by offering pen computers with numeric keypads; their PEN*KEY 6100 and 6300 models are so equipped and weigh in at only 2 and 3.5 pounds, respectively. Many other pen systems have optional keyboards. Although you increase size and weight of the package whenever you use a keyboard, it is certainly an option when lots of input is required.

Why not wireless?
Lets assume you decided to program your new pen application in Visual Basic. You've settled on a design philosophy and laid out a number of pen optimized screens. You have also selected a pen computer which meets your needs in weight, battery life, and usability by your users. But now you discover that a wireline approach to mobility involves docking stations or special cabling in elaborate and expensive setups at key locations as well as additional computers to handle communications with the new portables. You suddenly realize that each of your small mobile computers is going to need floor space for 20 pounds of support hardware, as well as careful handling of it all. This is when you ask: "What about going wireless?"

This is actually a lot easier than it used to be. PC Card-sized modems provide wireless access to the nationwide ARDIS and RAM radio networks. But what about the software and utilities needed by Visual Basic programmers to make use of wireless technology without reinventing the wheel? Welcome to the world of VBXs (Visual Basic Extensions). One company which may provide exactly what you need is Dynamic Mobile Data System Inc. in Staten Island, New York (see sidebar).

Before you design your wireless mobile application, there are a few more things you should investigate. Wireless services such as ARDIS, RAM, or CDPD will charge you for sending and receiving data. Review their offerings and compare pricing. Charges are typically based on the number of packets sent or received and although pricing is starting to drop, costs can add up. So design your application to make the best use of an expensive medium. Transmitting data only instead of entire screens, and limiting the transfer to changed data only are good ways of economizing on wireless transmissions. As for coverage, RAM and ARDIS have been around the longest and cover the largest areas. While RAM has a high speed throughput in most cities. ARDIS is just now deploying new comparable speed channels. CDPD is still very new and not available in many areas but users who have tried it are amazed by its speed.

Soon your mobile pilot will be underway and you will start getting calls from all of your envious business associates, asking how you did it. You don't have to let them know that it actually was quite easy.
- Dom Giangrasso

Dom Giangrasso is a Manager of Special Projects for Consolidated Edison of New York. and has been working with mobile computers since the days when the first 30-pound Compaq portable was considered state-of-the-art. He is a contributing editor for Pen Computing Magzine and also serves on the World Market Strategies PDA/Handheld Forum Advisory Board.