What Happened at EO?

When GO announced PenPoint in January 1991, many developers were excited. Here finally was a system built to support an interesting new model of computing. It was relatively simple to use, modern in design, and had plenty of room for growth. Now, everything seems to have turned upside-down.

In the first week of February, EO laid off about half of its employees. According to sources inside the company, current investors have committed money to bring out the next EO machine code-named Loki. Loki is a small form factor smart phone, designed to compete with Simon for communication capabilities, and with Newton for intelligence. In the meantime, the company is attempting to restructure its debt and reposition itself, cutting back to a core staff to develop Loki.

It’s interesting to note what insiders say about the situation. According to sources, EO’s majority shareholder, AT&T, was displeased with EO’s lack of sales; the Wall Street Journal quotes EO’s sales rate at about 1000 machines per month, with total sales volume under 10,000 units. In addition, AT&T Microelectronics, the part of AT&T responsible for the Hobbit chip, has had continuing disagreements with EO over the future of the Hobbit family.

Sources say that EO rebuffed the idea of merging AT&T’s DSP design and other circuitry into the Hobbit. With no customers for this high-end, high-profit chip, AT&T could see little economic justification for the Hobbit family, certainly not at a sales volume of no more than a few hundred thousand units a year. AT&T’s calculations included the failure of EO’s OEM strategy—those unnamed Japanese companies who were “committed” to bringing out Hobbit-based PenPoint machines which never materialized.

EO and GO had another significant problem: their product development process was constipated. GO’s last major release was nearly two years old. During this time, GO released bug fixes and a modest effort called “PenPoint 2.0J”—which is little more than a Unicode version of PenPoint with a fixed text rendering engine—aimed strictly for the Asian market. In addition, GO released some mediocre applications for PenPoint such as GO Fax and GO Mail.

But the long-awaited, PenPoint 3.0, code-named Amstel, was nowhere to be seen. In fact, Amstel was so far behind schedule that in early 1993, before the merger with GO, EO had decided to stick with the slow, bulky, albeit working 1.0H version of PenPoint for Loki. Following the merger, EO redirected GO’s development effort, changing Amstel to meet EO’s needs and retargeting Loki for Amstel.

After all this, what will Amstel finally look like? Insiders say that it has changed many times, with many people getting different stories, however, it appears to be a slimmed down, faster PenPoint, to be accompanied by the long-awaited new development environment. The development tools for PenPoint 1.0 have long been reviled by many developers because of its excessive complexity and 1970’s-vintage inadequacies.

Developers actively working with PenPoint have long looked forward to an environment as friendly as the one for Newton. However, now it appears that we won’t get that either; insiders hint that PenPoint will be closed to outside developers. According to sources, EO’s new business model will be similar to the Sharp Wizard. This means that much of the software will be bundled with the device, with little additional software added by the customer during the life of the product.

Under this scheme, EO will tightly control access to the software interfaces and work with only a few selected companies to develop targeted applications. At some future time, a new development environment based on a scripting language from a third party may be announced. Of course, this environment was expected to be released early last year, and then finally by the last part of the year. However, with the February layoff the company eliminated almost all of the technical support and tools development staff. This makes it hard for EO not to close the system, de facto if not de jure.

As slow as GO’s development efforts have been, EO’s haven’t been any faster. For over a year, EO’s development staff has supposedly been focused on Loki. But where is it? Thom Hogan, a former GO and EO evangelist, spoke publicly about Loki in April 1993, when the Loki project had been underway for months and had produced a working circuit board. Certainly, EO’s staff hasn’t been working on the EO 440 or 880 all this time; those products were completed and in production for nearly a year.

A Feb. 14 story on the Kyodo News International service alleges Matsushita Electric Industrial Co. has halted production of EO machines at EO’s request. EO’s Bob Roblin has even been quoted as saying that the EO 220 (Loki) may NEVER reach the market! Indeed, one report has him saying EO hasn’t decided what hardware products, if any, the company will manufacture. What we have is a picture of a company in confusion, a state persisting since last year’s merger with GO.

Clearly EO has a problem: the product isn’t selling, developers aren’t developing, the products aren’t getting the press coverage of Newton, and AT&T isn’t willing to ante up additional money. Worse still, sources say that when word of the “temporary suspension” of future Hobbit chip development leaked out, Asian investors with whom EO was desperately trying to close negotiations backed out.

Short of cash and with no immediate prospects of raising more (similar to GO’s position when AT&T forced the merger with EO), EO chose to layoff and restructure. And that’s how we got where we are today.

The future of PenPoint is cloudy, very cloudy. When the merger of GO and EO was announced, we felt that AT&T’s intent was to foster the development of technologies and markets that would sell dial tone. The purchase of McCaw Cellular was a link in that strategy, providing AT&T with higher-quality raw material (cellular dial tone) to sell.

However, irrespective of the AT&T/EO’s “Don’t worry; be happy” attitude, neither AT&T nor any other phone company has any real interest in “creating” the market for pen computing.

The larger issue is obvious: what is the near-term future of pen computing? EO’s failure, along with the recent collapse of Slate and other dead pen companies (Momenta, Linus, et. al.) casts a shadow on the industry. Even mighty Apple is said to be unhappy with Newton sales, running at about 7,500 per month.

As industry analysts cut their PDA sales estimates, the string of failures lines up in a straight, unwavering horizontal line: the horizontal market.

The obvious is becoming unavoidable: there is no short-term horizontal market for pen computing. The available machines just don’t do enough valuable things. EO is the technical leader, but at too high a price and too short a battery life. And the Newton simply doesn’t work well enough for the consumer market.

Fortunately, several extremely good people remain at EO. But the presence of top-notch engineers can’t compensate for mismanagement and nonmanagement. The future of pen computing is cloudy and doubly so because of the disservice done to it by Apple’s relentless hyping and by AT&T/EO/GO’s mistakes.

In the meantime, people looking to make money should continue to look to vertical applications. For the next few years, the real market for pen computers will be in vertical, business-oriented, less-than-12-month-return-on-investment applications.

In time, wireless communication becomes widely available, when the battery and screen problems are solved, and when machines can compete with paper on reliability and readability; then, and only then, will there be a horizontal market.

Till then, keep it simple. Test early and often; stay close to the user (and not just your contract administrator!); stick with reliable, proven technology. And consider communications at the very beginning.