My Newton story

Today the iPhone is the alpha gizmo, the one item of consumer electronics that dominates all the others.

But in 1993, the hot new gizmo was Apple's Newton, and it was a whole different thing.

Not very many people had Newtons. Apple sold fewer Newtons over the whole life of the product than it sold iPhones the evening of June 29.

Also unlike the iPhone, the first Newtons weren't even very useful. Although called "personal digital assistants" (PDAs), using a Newton was significantly more difficult than using a Day-Timer. The original MessagePad had very poor handwriting recognition, and there was no practical alternative to using it, no on-screen or slide-out keyboards. With patience, one could make notes, manage an address book and a calendar, and even send and receive faxes.

But honestly, it wasn't very good at any of these things. The return on the total investment, including the up-front costs and the time and effort of learning to use the device, was not so good.

Apple introduced several minor upgrades of the original MessagePad-- four new models in two and a half years-- addressing some of the hardware and software issues, but it wasn't until the MessagePad 2000 came out in 1997 that the Newton finally realized its full potential.

I had been watching the progress of the Newton very closely, trying to persuade IDT (Integrated Device Technology, where I was working during those early years of the Newton) to bring out a MIPS-architecture processor for this new PDA market. In 1995, I even made my own wooden prototype PDAs to show just how small a PDA could be using the technology of the day-- unaware that Palm's Jeff Hawkins had done the same thing the year before to help get the Palm Pilot project off the ground. (When the Pilot came out in 1996, I was entirely uninterested. Graffiti was a crippling defect, as far as I was concerned.)

I bought a MessagePad 2000 in April of 1997, and it was immediately useful to me. The handwriting recognition engine had been significantly improved on the MessagePad 120, but was still constrained by that model's 20MHz ARM610 processor. On the MP2000's 162MHz StrongARM SA-110 processor, the new recognizer was nearly flawless for me after just a few days of practice. Not everyone had this kind of success, but I usually saw no more than one error per paragraph of text, and it was very easy to correct those errors.

I did have to learn to print a little more neatly than I was used to, but not much. One problem continued to dog me as long as I owned the unit-- when I print a lower-case "g", I start at the top right and draw the circle counter-clockwise, and sometimes fail to close the circle before drawing the descender, especially if I was writing quickly. The Newton's recognizer often interpreted that shape as a lower-case "s".

Anyway, the MP2000 was a great fit for me. In 1996 I had joined the staff of Microprocessor Report. Attending conferences was a big part of my job, and the Newton was the perfect device for taking notes during interviews, presentations, and while visiting exhibition booths. The Newton also helped me manage my schedule. I didn't use it as my primary address book, though. I found it more convenient to use my PowerBook for that purpose, with phone numbers in my cellphone where I could actually use them.

In 1998 I had my MP2000 upgraded to the MessagePad 2100 configuration, which basically just took the RAM configuration from 1M to 4M. That gave me enough room to fiddle around with more of the third-party Newton software that was out there. There was actually a pretty good variety, mostly from very small companies that specialized in Newton software. The Newton wasn't easy to write software for, and Apple didn't support third-party developers as well as they could have, but there was some great software on the market.

I also experimented with using the Newton for Internet access. This worked pretty well with a Farallon Ethernet adapter, but wired Ethernet on a handheld device isn't a great combination. I also tried Metricom's Ricochet wireless Internet adapter. That seemed to work, but I discovered after a few weeks of testing that it was somehow corrupting the data being written to the Newton's flash memory. Either the power draw of the adapter was too great, or the Ricochet's radio transmitter was interfering with the Newton's electronics. I hadn't been making backups of the Newton as regularly as I should have, probably because it had always been almost perfectly reliable, so this Ricochet problem caused me a lot of grief.

Ultimately I decided I didn't have any strong need for Internet access on the Newton. It couldn't substitute for a laptop anyway, so I stopped worrying about it.

By 1999 or so, I stopped experimenting with the Newton entirely; I had a good software setup, the machine did everything I wanted, and it was totally reliable. I used the Newton until 2004, when I left Microprocessor Report. I had a Palm Treo by that time, and still do. It's not a complete Newton substitute by any means, but I can use it to take short notes-- or, conveniently, voice memos-- and it's a better device for calendaring and contacts because it's always with me.

In 2005 I got a Tablet PC (a Motion Computing LE1600) to fill in that note-taking gap at conferences. The Tablet PC handwriting recognizer isn't as good as the MP2000's, but it's tolerable. Tablet PCs are also huge and heavy by comparison with the Newton, but again, that's tolerable. In exchange, a Tablet PC runs a fully-featured operating system (I now have Vista on mine), mainstream applications like Microsoft Outlook and the Firefox browser, and I bought a Sierra Wireless AirCard 850 HSDPA wireless Internet card for it.

I could go on about the design elements Microsoft should have adapted from the Newton into the Tablet PC. I suppose I will, in some future column. Some of these features would be a good fit for a future evolution of the UMPC (Ultra-Mobile PC), which today is really just a Tablet PC with a too-small screen. In time, I expect the UMPC will be adapted to be a better fit for its form factor, and some of the Newton's features would help.

So far I've felt no urge to get a UMPC, which some believe bridges the gap between the Newton and Tablet PC. The forthcoming HTC Shift is very tempting, however. I've held one, but I'll have to wait to see what the final features and price are like.

Most people believe the Newton was a huge failure for Apple. I had the chance to ask John Sculley, who was Apple's CEO when the Newton project was launched, about that. He pointed out that although the Newton never paid off its development costs as a product, Apple's early involvement in developing the product category-- particularly its investment in ARM, the company that developed the original Newton microprocessor-- paid off handsomely.

It seems to me that today's technology would support the development of a fairly Newton-like device-- about 12 ounces with a 7" screen, thin and rugged, with integrated wireless Internet or Bluetooth to borrow the connection from a nearby cellphone, good handwriting recognition, and plenty of on-board storage, selling for around $400. I'd buy one, but who else would?

If you had a Newton, or have your own ideas about this product category, why don't you add a comment? Maybe we can get some hardware company interested once again.