Eulogy on the Newton, and why Windows CE will be BIG

Before I get into current topics, a few paragraphs about an old and now departed friend, the Apple Newton MessagePad. The Newton will always have a special place in my heart, and I’ll likely continue using my MP 2100 for some time to come. There isn’t anything else out there that can match its handwriting recognition capabilities, at least not for me.

Hundreds of handheld and mobile computers have made it to our Pen Computing office through the years, but none enchanted me as much as the original Newton MessagePad. I remember the day I picked mine up at the local Apple store, right off the truck, and how I stayed up all night admiring the MP. It was so unlike any other computer I had ever seen, so elegant and—in many ways—futuristic. I read every book on Newton I could find (see reviews in early issues of Pen Computing Magazine) and contacted their authors. I tried to understand the concepts behind the MP’s handwriting recognition, and was rewarded with very acceptable recognition right from the start. True, it would get much better over the years, but it worked even then.

Delving into Newton’s brilliant innards and capabilities was like exploring a magic realm. Some of the novel concepts worked, some did not, but it was clearly a revolutionary device.

Though the original MessagePad’s communication capabilities were widely criticized, for late 1993 they were, in fact, quite advanced. You have to remember that times were different. This was before the Web, and before spam and junk e-mail began clogging up your in-box. Between NewtonMail and that little old 2400-baud "cigarette box" modem with its perennial rattle, I often had quicker and more reliable e-mail access than I have now. And Newton could even download system upgrades over NewtonMail.

Though it certainly had its flaws, there has never been another piece of equipment that made me feel like the original MessagePad did. It was like being there at the birth of something entirely new, like seeing the future, and being part of it.

The real future, which is now already the past, didn’t turn out the way it was supposed to. Apple—long on brilliance and short on resources—simply couldn’t do it. Newton’s career was also a textbook example of "regression to the mean," the the theory which says that outbursts of excellence (or lagging) inevitably regress to the average. You can only maintain excellence for that long. The conception of Newton was a period of excellence, but it eventually burned itself out.

Apple almost completely missed out on desktop connectivity and synchronization. Had a connected MessagePad simply shown up as a folder on the Mac or Windows desktop, things might have developed very differently. But the Newton Connection Kit never developed into the simple, reliable tool it needed to be.

I remain convinced that the Personal Digital Assistant concept—I still prefer the term PDA over all the handheld/palmtop/palm-size, etc permutations—and the flashes of innovation and technological brilliance we’ve seen in Newton virtually guarantee a next round. But I will never forget the original Newton MessagePad. It was special. And Apple was stupid for killing it.

But on to the present and future. It’s becoming increasingly clear that Windows CE may be slated for a career far more important than initially expected. My crystal ball says that CE will eventually become the standard operating system for a very large part of the overall PC/NC market, while Windows 95/98/NT will be reassigned to server and backoffice duty. Let’s take a look at the current situation:

Windows has grown into an almost unmanageably large and complex operating system. The Microsoft people working on the many parts of Windows are all very competent, but the overall thing has simply gotten out of hand. There are far too many settings, control panels, resources, and so on and so on, that can, and do, go wrong. Windows 95 is definitely not very stable, and I’ve spent far too many hours of my life tending to its problems. Windows 98 has some of the roughest edges worked out, but it’s much bigger yet and if it crashes on Bill Gates, it’s almost certainly going to crash on others.

And Windows 95 still basically makes us live with limitations of the original 1981 IBM PC and the original version of DOS. Using anything other than COM1 or COM2 is asking for trouble if you have a lot of applications requesting COM ports. IRQs remain at an extreme premium which makes adding devices an exercise in frustration. On my plain vanilla Pentium II/333 Gateway 2000 all IRQs are taken, and two are double-booked.

Then there’s Window’s reliance on Intel processors which are expensive, inefficient, and could heat a log cabin in Alaska. In a laptop, they burn your lap. The heat sink in my Gateway alone is bigger than most handhelds. But since Windows requires Intel processors (discounting the Alpha), there really isn’t any competition and the latest Intel processors continue to cost a small fortune.

So what we have here is an operating system—based on legacy underpinnings—that has grown far too big and cumbersome for the average user, running on a single and somewhat outdated processor architecture. If the trend continues, future versions of Windows will be bigger and clumsier yet, and the way things go, future Intel X86 chips may well need to have water-cooling.

All of this is marginally acceptable in a stable desktop environment where power is not a problem, where the configuration remains the same, and where the system is started once a day, or left running overnight.

In a mobile environment, it is not acceptable. Power is a big issue. The computer is frequently turned on an off. Peripherals are plugged in and removed during the course of a day.

And, increasingly, the Wintel overhead and complexity is too much even for desktop systems. Windows 95/98 is the end of a long, arduous road. Let it die before it gets much worse. Then let’s put Windows NT, hopefully clear of all MS-DOS legacy, onto the server and be done with it.

The future belongs to a new approach, and that approach is Windows CE.

Contrary to common belief, Windows CE is not just a shrunken version of Windows 95. It was actually built from the ground up and has an entirely new and optimized kernel. True, Windows CE on a handheld PC looks just like Windows 95, but that’s just a shell and Microsoft did it to provide that fuzzy familiar feeling. One look at a palm-size PC shows that, beyond the Start menu and a couple of control panels, the user interface is entirely different and optimized for the platform, as it should be.

In order to make Windows CE simple to program for, Microsoft identified hundreds of Windows APIs, all already familiar to Windows programmers, and optimized them for CE. Microsoft also "componentized" CE into some 120 modules, so that OEMs can include only what their device needs. And Windows CE toolkits are just extensions of already familiar Windows programming tools.

Unlike Windows 95 and Windows NT, Windows CE can run on just about any 32-bit processor architecture. This means the end of Intel’s monopoly. There will be a healthy competition between makers of Win CE processors. Those chips—already fast, energy-efficient, and inexpensive—will become faster, more efficient, and even less expensive.
Since there is no longer a need to live within the x86 or the DOS legacy constraints, devices will also no longer be crippled by inadequate communications and resource support. Those limitations simply no longer exist.

So don’t be fooled by Windows CE’s meek beginnings. Windows CE is not about the tiny keyboards and dim 480 x 240 monochrome screens of the first generation HPCs. It’s about providing the right tool for the job, no more and no less. It’s about doing away with unneeded complexity and breaking with a past that has become increasingly burdensome.

Take a look at the latest Windows CE devices in the CE section of this issue. The first generation of palm-size PCs is quite impressive and far more polished than the first HPCs were. And the new class of "super handhelds" is sensational. I’ve been carrying my NEC MobilePro 750C with me wherever I go. I love its almost full-size keyboard, its big, bright color screen, and the softmodem that allows me to browse the Web for hours. My old Pentium lap burner has been relegated to desktop duty.

Windows CE will keep evolving in leaps and bounds, and in ways that will surprise you. It will quickly become the operating system of choice for mobile computers, and for millions of other PCs. Finally Microsoft is on the right track. It’s good to be able to feel optimistic about Windows.

-Conrad H. Blickenstorfer is editor-in-chief of Pen Computing Magazine and general editor of Digital Camera Magazine. He is also a mobile technology contributor to the Fortune Magazine Technology Buyer's Guide.