Technologizer

Newton, Reconsidered

A hands-on assessment of Apple's pioneering, ill-fated Personal Digital Assistant, twenty years after its original unveiling.

  • Share
  • Read Later
Marie Domingo

In the grand scheme of things, 1992 is such recent history that it barely qualifies as history. When it comes to portable gadgets, however, it’s an era that’s nearly unrecognizable to us 21st-century humans.

Sure, there were pocketable gizmos back then: The Game Boy, for instance, had been around since 1989, and the Sony Watchman was hot stuff. There were even miniature computers, such as HP’s 95LX. But in 1992, nobody had an MP3 player. Or a GPS handheld. Or a smartphone. (Less than five percent of people in North America had a mobile phone, period.)

And in 1992, nobody had a PDA. That’s Personal Digital Assistant, in case you’ve forgotten, and even though nobody had one, lots of people were talking about them. Apple CEO John Sculley had coined the term in the keynote speech he made at the Consumer Electronics Show in Las Vegas on January 7. He announced that Apple would release PDAs–pocket-sized information devices, easier to use than a PC and selling for under $1000–in 1993.

Twenty years ago this week, on May 29, 1992, Sculley spoke again at another CES, in Chicago. This time, he didn’t just talk about PDAs. He brought one with him. It was a Newton, a prototype of the device which Apple planned to start selling in early 1993. Actually, Apple had multiple Newtons on hand that, which was good: The first one it unveiled on stage had dead batteries and didn’t work.

Using a second unit, Steve Capps, one of Newton’s creators, showed how you could use it to order a pizza by moving topping icons onto a pie and then sending out a fax. In 1992, that was show-stopping stuff.

The hype surrounding Sculley’s CES announcement of the Newton was immense, a precursor of the hoopla that would later accompany Steve Jobs‘ keynotes for iPods, iPhones, iPads and other post-Newton gadgets. But it didn’t change anything right away. Sculley, in fact, was demoing vaporware: The Newton was nowhere near ready. Apple held another unveiling fourteen months later at Macworld Expo in Boston, when the product which it officially called the MessagePad finally went on sale on August 2, 1993. (It was one of what was supposed to be a line of Apple products based on “Newton Intelligence,” most of which never came to be; Apple also licensed Newton technology to other companies.)

The earliest MessagePad reviews tended to accentuate the positive, but public sentiment quickly turned against Apple’s PDA, so much so that the Newton, like Microsoft Bob, remains convenient shorthand for “technology flop.” People remember that the handwriting recognition didn’t work–or, more specifically, that Garry Trudeau’s Doonesbury featured a week-long sequence in August of 1993 in which the handwriting recognition on Mike Doonesbury’s Newton-like PDA didn’t work. They recall that the Newton didn’t sell in huge numbers. They know that Steve Jobs axed it when he returned to Apple in 1997.

But simply dismissing Newton as a failure is unfair. Microsoft gave up on Bob after about a year; Apple, by contrast, stuck with Newton for six years. It released seven distinct models and worked with companies such as Sharp and Motorola, which released their own Newton-based gizmos. If someone other than Jobs had been in charge of Newton in 1998, the year it went away, it’s at least conceivable that there might be Newtons of some sort even now.

What Newton wasn’t was a hit. Apple sold 50,000 MessagePads in the device’s first three months on the market; the company trumpeted this figure as evidence that the gadget was selling briskly, but it was more likely a major disappointment. (In July 1992, MacWEEK‘s Jon Swartz reported that the company expected to sell a million Newtons in the first year.)

A lavish 1993 coffee-table book on the Newton project was titled Defying Gravity, but Newton’s problem was that it never quite took flight. In the six years between its premature debut and untimely death, it wobbled and sputtered like a leaky balloon, neither soaring nor crashing.

Of course, most of the people who have only a foggy understanding of why the first Newton failed to live up to the early irrational exuberance have a good excuse: They never actually used one. Including me.

I was a working technology journalist in 1992, but I paid only scant attention to the Newton’s launch. I didn’t try to get my hands on one when they went on sale in 1993, and didn’t bother to attend the unveiling at Macworld Expo Boston even though I worked nearby. In 1995, I bought my first PDA, Psion’s wondrous 3c; I don’t remember a Newton even being on my list of possibilities.

Several years later, I reviewed some of the final descendants of the original model, such as the MessagePad 2000, and met with the Apple executives who ran the Newton business. But when it dawned on me that 2012 marked the 20th anniversary of the Newton, I still felt like I wasn’t in a position to express informed opinions about it.

So I bought one.

Thanks to the modern miracle known as eBay, it’s not hard to acquire a Newton. I lucked upon a remarkable specimen: a first-generation model, the MessagePad H1000, running version 1.0 of Newton OS. It wasn’t just in mint condition, in the original boxes with all the original accouterments and documentation, plus a shrinkwrapped introductory videotape. There was no sign that it had ever been booted up. By buying such a virginal example, I would get the same Newton experience that the earliest adopters got when they plunked down their $699 in 1993.

eBay turned out to be an embarrassment of Newton riches. As long as I was shopping around, I bought a vintage external Newton fax/modem, capable of operating at a blistering 2400-bps. And a memory card called the Newton Enhancement Pack. And the Newton Connection Kit, which included hardware and software for hooking the MessagePad up to a Windows PC.

A few weeks later, my sister-in-law–who didn’t know of my Newton experiment–told me that she’d uncovered her own MessagePad H1000 and asked if I wanted to check it out. It turned out to be in nearly as pristine condition. I gratefully borrowed it as a backup.

This article isn’t a history of the Newton-here’s a good one–but rather my notes after six weeks as a Newton user, at long last. I had fun and wound up with a new appreciation for this groundbreaking gadget. But I also got a better sense of why it wasn’t the epoch-shifting breakthrough that Apple promised and pundits predicted.

Looming Large

Even before I used my new H1000 for the first time, I realized how unfamiliar I was with it. For one thing, I didn’t know it was so big. As the Wall Street Journal‘s Walt Mossberg pointed out in his 1993 review, it’s roughly the size of a VHS videotape. Technically, you can squeeze it in a coat or pants pocket–as long as you don’t mind everyone knowing that you have something roughly the size of a VHS tape stuffed in your pocket. A shirt pocket, however, is out of the question.

The MessagePad is an example of the gadget category which many people (including Steve Jobs) have disparaged as “tweeners.” Neither truly pocketable nor capable of replacing a full-blown PC, tweeners have never gone away. Current examples include Amazon’s Kindle Fire and Samsung’s Galaxy Tab 7.7.

That might help explain why so many people appeared to be oblivious to my MessagePad when I used it in meetings, on airplanes and at a fancy banquet: At first glance, a reasonable person might mistake it for a Kindle or a Nook, especially when its screen is shut off. At one point, I silently sat it on the table during a lunch I had with an employee of a legendary Silicon Valley company. (No, not Apple.) She didn’t exclaim “Hey, a Newton!” Instead, she gestured at it knowingly when we discussed the current tablet boom, as if it were an example.

Once I’d put four AAA batteries and a watch-battery backup into the MessagePad for the first time, powering it up felt like bringing it out of cryogenic suspension. Newtons, it turns out, begin their lives believing that it’s 5am on January 1, 1993. And the only way to set the year to 2012 is to flip the calendar forward, one month at a time. I tapped the MessagePad’s screen 230 times to set the date, watching the months flutter by like pages falling off a calendar to indicate the passage of time in some old movie.

As I did, I was already struck by a fact about the PDA’s screen: It’s terrible. Terrible.

This surprised me. Back in 1993, reviewers had plenty of beefs with the Newton, but display quality wasn’t one of them. Critics apparently thought the screen–monochrome, 240-by-336 pixels, no backlight–was okay. High-quality color LCDs already existed, but they were the stuff of $4000 laptops; nobody would have expected one on a $699 gizmo.

The H1000’s display is greenish gray on grayish green; there’s a contrast setting, but it only lets you choose between gradations that amount to bad, worse and illegible. In optimal light, with contrast set just right, the MessagePad’s screen is readable but unappealing. Anywhere else, it can be a challenge to make out. At a cocktail party in a murky bar– the sort of place where I generally peek at my iPhone a dozen times an hour–it was hopeless.

The screen doesn’t even look too impressive in the introductory video which came on a VHS tape with new MessagePads.

[youtube=http://www.youtube.com/watch?v=64QuJdJmCbA]

20/20 hindsight may make the MessagePad’s screen look worse than it seemed in 1993; its battery life, however, benefits from a couple of decades of diminished expectations. Back in the 1990s, people squawked that the MessagePad H1000 drained its four AAA batteries too quickly. I found, however, that I could go for a couple of weeks on a set. In an age of smartphones that conk out after less than one day, that was more than enough to keep me happy.

Now, about that handwriting recognition. (It was, incidentally, developed by a team of Russian computer scientists who later went on to create Evernote, the gem of a note-taking app for the iPhone and other devices.) The H1000’s core apps include a note taker, a calendar and an address book, and the principal means of entering information into all of them is to write it with the PDA’s stylus. The original Newton TV commercials certainly played up the feature.

[youtube=http://www.youtube.com/watch?v=V7ie78LuL0k]

And in this excerpt from a 1993 episode of public TV’s Computer Chronicles, Apple product manager Tony Espinoza shows off the recognition working correctly–albeit slowly–and demonstrates most of the PDA’s other major features.

[youtube=http://www.youtube.com/watch?v=aJkI6xOPspM]

As you write words on the screen using the stylus, in either cursive or block letters, Newton OS checks them against its dictionary, which contains 10,000 words, plus any you add yourself. If the software manages to understand your scrawls, and the word is in the dictionary, it’ll convert it into editable text. But if it fails to decipher the characters you wrote–or it does decipher them, but the word isn’t in the dictionary–you’re toast.

(There’s an option you can turn on which, in theory, lets the MessagePad recognize words that aren’t in the dictionary, but in my experience it didn’t work. Ever.)

It’s difficult to say exactly what percentage of the time the PDA has managed to correctly turn my handwriting into text, in part because it theoretically gets better at the job as it goes along. When I write painstakingly and use words that are in the Newton dictionary, the handwriting recognition works, sometimes. Manually adding words it doesn’t know–like my own last name–helps.

Overall, though, the MessagePad is iffy for words even if they’re in the dictionary. It only seems to know proper names if they’re extremely commonplace ones, such as “John” or “Perkins.” And it’s downright abysmal for brand names. Especially ones that didn’t exist in 1993–strangely enough, “Google” and “Facebook” aren’t in its dictionary.

All of this is crippling, and leaves me thinking that one piece of conventional wisdom about the Newton–that it was doomed by crummy handwriting recognition–is correct. How could Apple have expected a PDA with an unusable input system to sell? How could anybody?

Sculley and company weren’t the only ones who failed to flag the recognition as a product-destroying flaw. The Wall Street Journal’s Mossberg cautioned that it was “probably still a bit too imperfect for most people” but still praised it as by far the best he’d ever seen. Investment fund manager Roger McNamee, who would eventually help run Palm, told the New York Times’ John Markoff that the MessagePad had “the best handwriting recognition that I have used.” InfoWorld‘s Stewart Alsop declared that “the handwriting recognition beats everything I’ve seen hands down (and I’ve seen them all).”

Mossberg, McNamee and Alsop were right: The Newton did have the finest handwriting recognition anyone had seen up until that point. But that didn’t really matter. It needed to be good enough to be useful, and in its original form, it wasn’t.

By the time Apple started selling the MessagePad H1000, it may have understood that the recognition was a disaster in the making. My friend Phil Baker–who’s worked on interesting gadgets as old as Polaroid’s SX-70 and as new as the Barnes & Noble Nook–led the team that built the second-generation Newton, the MessagePad 110. (His industrial designer was a new Apple recruit: the not-yet-legendary Jonathan Ive.)

Phil and his team were already at work on the 110 before Apple had unveiled the first MessagePad, and were startled by the favorable reviews in publications such as PC Magazine. At the original model’s launch, he told me, “I got worried when the marketing manager was entering words in the dictionary by hand an hour before the demo.”

Graffiti to the Rescue

Once I got tired of entering words into my Newton’s dictionary, I turned to a magnificent piece of software called Graffiti. It felt like a homecoming: The streamlined form of handwriting recognition, which has you enter simplified characters one-by-one, was the primary means of input on Palm’s PalmPilot PDAs for years, including several models which I owned. But before Palm put Graffiti on PalmPilots, it offered it as an app for the Newton and other early pen-based handhelds. (It developed the software after its own Newton-esque PDA, the Zoomer, crashed and burned within months of its 1993 launch.)

Graffiti may have been a tiny company’s attempt to rectify the Newton’s most glaring weakness, but it had a Cupertino seal of approval. An Apple division called StarCore distributed the software as part of the Newton Enhancement Pack, in packaging that called it “fast, accurate and frustration-free.” Even Apple, it seems, acknowledged that the standard Newton recognition was sluggish, inaccurate and full of frustration.

It had been eight years or so since I retired my last Graffiti-based PalmPilot, but the Graffiti characters were still lodged in the back of my brain. I was able to immediately begin writing on the MessagePad at a rapid clip, with nearly flawless accuracy. If every MessagePad had come with Graffiti preinstalled, I concluded, the early history of the Newton could have been radically different–and Garry Trudeau wouldn’t have had nearly as much fun.

Having mastered text entry, I moved on to the MessagePad’s built-in applications. They aren’t bad, and in many ways they’re roughly equivalent to the notepads, calendars and address books on modern smartphones. The gestures remain engaging–especially the zig-zag you draw to delete items, which self-immolate in a puff of smoke.

NewtonOS also has a feature called Assistant, a sort of proto-Siri which may be the single most wildly ambitious thing in the software. Instead of tapping your way around the apps, you can enter plain-English commands such as “Call Bryan at home” and “Remember order paper clips.” It’s hobbled by the balky handwriting recognition, but with Graffiti, it works.

Little reminders that Newton software is twenty years old do lurk everywhere. There’s no unified home screen with icons for all your applications in one place; in fact, there’s no icon for the Notepad at all. And bopping between programs is an odd experience–applications and dialog boxes pile up on top of each other like a stack of Post-It Notes. You’ve got to close them one at a time to work your way back to the ones on the bottom. Clearly, the Newton’s creators were still getting a sense of how a mobile operating system should behave.

They were also swatting bugs, or failing to do so. My MessagePad keeps slipping into a strange limboland–I can draw on the screen, but the PDA doesn’t respond to any input until I reboot it. It sometimes claims it’s out of memory when it isn’t. Neither of the Getting Started memory cards that came with either of the Newtons in my possession work in either unit; this may be due to a defect that was discovered after the first MessagePads shipped. And ActionNames, a calendar program on the Newton Enhancement Pack card, spawns error messages every time I try to use it. (Its manual acknowledges the possibility of problems and cheerfully suggests upgrading to a MessagePad 110.)

I’m still not sure if all of these glitches bedeviled the first owners of MessagePads, but if they did, they couldn’t have helped the device’s reputation.

Figuring out how to get additional software onto the MessagePad was no cakewalk. The Newton Connection Kit I’d bought reminded me how much computers have changed since the early-to-mid 1990s. The cable used a DB9 serial connector, an archaic standard long ago nudged aside by USB. As for the software, it came on a stack of 3.5″ floppy disks–which didn’t really matter, since it was a Windows 3.1 program which likely wouldn’t run on a 21st-century computer anyhow.

In theory, it should have been possible to hook it up to my 2012 MacBook Air using special software and an adapter that let the MessagePad’s serial jack connect to one of the MacBook’s USB ports. After failing repeatedly to get it to work for more than a half-second at a time, I gave up. And then I remembered that I had the Zenith laptop I used in the mid-1990s boxed up in my garage.

The Zenith had a serial port and a floppy drive, and it ran Windows 3.1. It worked fine with the Connection Kit. I ended up using my MacBook to download Newton apps — including a faux Tetris and an e-book of Alice in Wonderland — and then transferring them to the Zenith. Then I’d copy them to the MessagePad.

I had faint hopes of using my MessagePad and fax/modem for e-mail. It had come with a post card you could snail mail to Apple requesting a starter kit for the NewtonMail service; Apple apparently used the postcard as a means of buying some time, since NewtonMail wasn’t ready until months after the first MessagePads shipped. But NewtonMail is the only e-mail program that’s compatible with NewtonOS 1.0, and I assume that Apple discontinued it at some point in the past two decades. Even if it didn’t, it ran on Sprintnet, a dial-up network that’s no longer with us.

Unable to use my fax/modem to get online, I jotted a note on the MessagePad, then faxed it to myself. That worked perfectly.

A Newton Post-Mortem

Even by Newton standards, my original model, running 1.0 software, is a relic. Eventually, Newtons would enter the Internet age; clever people have even used later models to tweet and connect to wi-fi networks. And as Phil Baker explained to me, Apple was busy refining the MessagePad hardware and software even before the first MessagePad was sold.

Like Bullwinkle repeatedly declaring that he’s about to pull a rabbit out of his hat–“this time for sure!”–the company kept releasing new Newtons which looked like they might be the model that would make the MessagePad a bona-fide success. There was the 110, the 120, the 130, the 2000 and the 2100, plus a fascinating proto-netbook called the eMate 300.

Newton hope tended to spring eternal. In March of 1994, the New York Times titled John Markoff’s story on the MessagePad 110 “Apple’s Newton Reborn: Will It Still the Critics?” A year and a half later, a Markoff piece on NewtonOS 2.0 was called “Apple’s Newton Poised for a Rebirth.” But neither the 110 nor OS 2.0 nor any of Apple’s other improvements did the trick.

Along the way, the Newton handwriting recognition got much, much better. That didn’t help enough–which may be a sign that consumers simply aren’t as interested in taking handwritten notes on an electronic device as Apple expected. After all, the only major post-Newton attempt to popularize handwritten input, Microsoft’s Tablet PC, also fizzled.

Even back in the 1990s, I remember becoming convinced that Apple was energetically pushing the Newton in the wrong direction. Sheer technological potency wasn’t the problem: Compared to the first model, the final Newton PDA, 1997’s MessagePad 2100, had  fifteen times the clockspeed, 90 percent more pixels and more than twelve times as much RAM. It also had vastly better software.

But the 2100 moved the Newton even deeper into tweener territory. It was taller, wider and thicker than the H1000, resulting in an even less pocketable gadget. It started at $1000; the H1000 had been $699. And it still didn’t come standard with the hardware and software you needed to exchange data with a PC.

Might Apple have served the Newton better by moving it in a different solution? We don’t have to treat that question as an imponderable. We just need to look at the case study provided by Jeff Hawkins’ Palm Computing.

After Palm’s big-time partners, Radio Shack and Casio, bailed on the Zoomer II, the company retrenched. With almost no resources, it got to work on a PDA that was far smaller, simpler and cheaper than the Zoomer. The new gadget used Graffiti instead of conventional handwriting recognition, and it came with a docking station and software that made syncing calendar and contacts with a PC a breeze.

Palm called its creation the…Taxi. But only briefly. When that moniker turned out to have trademark problems, the company changed the name to Pilot, shortly before the first units went on sale in 1996. Starting at just $299, the Palm Pilot became the phenomenon that the Newton was supposed to be but never was. It eventually evolved into the Treo, which was one of the two most significant pre-iPhone smartphones, along with the BlackBerry.

There was certainly some Newton in the Pilot. And there’s an awful lot of Pilot in the iPhone 4S, the iPad and every Android device–starting wtth the home screen’s grid of icons and the way apps run in full-screen mode. Had Apple followed Palm’s path–smaller, simpler, cheaper–it might have made all the difference.

Or maybe not. Newton fans took Steve Jobs’ 1998 decision to discontinue Apple’s PDA badly. They accused him of bearing a grudge against the gadget based on its origin as the brainchild of John Sculley, the man who had fired him in 1985. Perhaps so. But various other Apple executives had flirted with shuttering or selling Newton for years before the company’s co-founder returned. (Sculley was sacked by the Apple board just over two months after the first Newtons went on sale, in part because he was spending too much time being a visionary and not enough time selling Macintoshes.)

Another point to consider: Jobs didn’t seem to have any problem with plenty of other products and technologies which originated during the Sculley era, including PowerBooks, FireWire and QuickTime.

“The Newton achieved cult status, but the market was not big enough to see it become a success,” says ZDNet’s mobile-gadget guru James Kendrick, who owned an original MessagePad back in the day. “It was too expensive and big and clunky to become a mainstream hit. Jobs killed it out of necessity as the situation Apple was in when he returned was too bad to continue the expensive Newton line.”

Still, Kendrick sees hints of Sculley’s failed experiment in Jobs’ blockbuster gadgets. “I believe that Siri, introduced on the iPhone 4S, has a direct lineage with the Newton,” he told me. “The natural language interpretation was begun way back when on the Newton, as Apple tried to make the Newton work in a natural way.”

When Jobs decided to shut down the Newton division, color screens were still unaffordable, touch input was crude and wireless data didn’t get much more exciting than two-way paging. When he launched the first iPhone nine years later, technology allowed Apple to build the sort of devices it wanted to create in the 1990s, but couldn’t. He may have killed Newton, but he didn’t kill the dream behind it so much as press a giant pause button–and after finally spending quality time with a MessagePad, I’m more convinced than ever that he made the right call.

[MORE APPLE NOSTALGIA: Earlier this year, the Apple II turned 35.]

  1. Previous
  2. 1
  3. 2
  4. 3