Pioneering the Laptop: Engineering the GRiD Compass

Alan J Weissberger, CHM Volunteer
aweissberger@sbcglobal.net

This report covers the Grid Compass panel discussion, which was held March 15, 2006 at the Computer History Museum in Mt View, CA. It was part of the Odysseys in Technology lecture series which was sponsored by Sun Microsystems Laboratories.

Introduction

GRiD Systems Corporation was truly a pioneer in mobile computing, and many of the technologies present in notebooks and tablet PCs today would not exist were it not for the GriD Compass.

Introduced in 1982, the GRiD Compass 1101 was likely the first commercial computer created in a laptop format and one of the first truly portable personal computers. With its rugged magnesium clamshell case (the screen folded flat over the keyboard), switching power supply, electro-luminescent matrix display, non-volatile bubble memory, and built-in modem, the hardware design incorporated many features that we take for granted today. Software innovations included: a multitasking operating system (with a hierarchical file system), a graphical user interface, remote file system, and integrated productivity suite (including word processor, spreadsheet, graphics and a data base manager). In addition to the GriD Compass Computer, there were two types of networking services offered by the company - GriD Central and GriD Compass Central. Remote services were offered va GRriD Central. The company also offered a personal version of GRiD Central and GriD Compass Central.

John Ellenby founded GRiD Systems Corporation in 1979 and was later joined by co-founders Glenn Edens (VP of Development) and David Paulsen (Manager of Hardware Engineering). These latter two co-founders participated in the panel session chronicled below. Carol Hankins (Software Development Manager- responsible for GRiDOS and applications development), and Craig Mathias (Manager of the Network Products Group-responsible for communications and networking products including GRiD Central) were also on the panel. New York Times technology correspondent John Markoff moderated the panel session and led the Q &A/discussion that followed.

Opening Remarks by Jeff Hawkins (founder of Palm Computing and Handspring)

Having worked at Grid Systems for 10 years, Hawkins was quite pleased to introduce the panelists and moderator. Before doing so, he provided a testimonial for the CHM and then told a few of his personal "GRiD stories."

Hawkins is a big fan and supporter of the CHM, which he referred to as "a museum in the making." With the world's largest collection of computer artifacts, it is sure to be a world-class museum. The CHM will be a popular destination for people from all over the world. It provides a foundation for the historical significance of computing. This evening's "Pioneering the Laptop" event is part of the archival process, which documents the history of computing. The event is part of the Odysseys in Technology series, sponsored by Sun Microsystems Labs (Hawkins thanked Sun Microsystems for their sponsorship of this event).

Jeff Hawkins then told us his three personal "GRiD stories:"

1. Product positioning: What you first say about a product sticks for a long time. At $8,150 the Grid Compass was very expensive at the time it was commercially introduced in 1982. Grid Systems was thus remembered as the company that made expensive laptops - even if they made much less expensive laptops later on. Were the magnesium case laptops as rugged as people then believed? Jeff wasn't sure, but confirmed that the rugged image, coupled with the high price tag, enticed the U.S. government (Dept of Defense and NSA) to be one of GRiD's best customers.

2. Technology adoption: GRiD was designed for business executives in the early 1980's, but its keyboard was perceived a threat to those same executives. They didn't have any terminal or other keyboard device in their office, and so were uncomfortable with a PC. They were afraid of typing, thought they might appear inept, and even felt it was a demotion to type (in those years, secretaries took dictation and typed memos for business executives). This contradiction slowed GRiD adoption - at least in that targeted market segment.

3. Solid Invention: GRiD was a remarkable product in many ways. Grid filed a design patent on the laptop's flat panel display that folded and hinged over the keyboard. Initially, the patent wasn't assumed to be very valuable, as GRiD Systems Corp. was the only company making portable PC products at that time. Years later, GRiD licensed the patent to other laptop manufacturers and collected significant royalties. One of the licensees challenged the patent in court and won. That rendered the patent useless and cut off the licensing fees. [Glenn Eden later corrected Jeff, saying that the patent was re-instated and GriD Systems collected royalties untill its dying day]

Hawkins introduced the panelists - all of whom he cited as "computer pioneers." Then he introduced session moderator John Markoff of the NY Times.

John Markoff's Opening Remarks

John stated that we would explore the design and engineering of the first laptop computer - the GRiD Compass 1100. He thinks of the Grid as the "missing link" between what was done with the large computer systems of the 1960s and 1970s and the PC industry, which emerged in the 1980s and 1990s. The GRiD was about more than portability. It was as interesting inside as it was outside. There were many first- time- ever attributes for a PC. These included: a multi-tasking OS, a graphical user-interface, advanced link management, and network file system.

Markoff then related his single "GRiD story" about a hard core GriD user. Admiral John Poindexter was a Grid user and hacker when he was National Security Advisor in the Reagan administration. He would use his Grid (with integral modem) to dial into the federal governments computers in Washington DC, where he reportedly "hacked" into an IBM mainframe PROFS system, while visiting the Reagan Ranch.

Glenn Eden's Slide Presentation on GRiD Systems Corp and the GriD Portfolio

Former Xerox executive John Ellenby founded GriD Systems Corp in January 1979. Glenn Edens and David Paulsen later joined as co-founders (in Sept 1979 and November 1979, respectively). They moved from a garage to their first office building in August of 1980 at which time they started hiring people. The company was very secretive (not even telling job applicants what they might be working on) and remained in stealth mode till March 30, 1981 when they issued their first press release. The original GriD concept was inspired by Alan Kay's vision of a Dynabook (more about this later in the session). Ellenby wanted to develop a portable email terminal, while Glen envisioned a portable personal computer. Indeed, Eden said his primary contribution was to push the founding team to do a whole computer, complete with local processing and storage (rather then a dumb terminal with only email capability).

The GriD Compass 1100 computer was actually the key element of the "GriD Navigator System." The other two elements were GriD Central and GriD Compass Central (see below). The GriD Navigator System was described in a brochure as being a "tool to let the manager and business professional gain command and control of information critical to sound business management." It was said to "help you gather, analyze, and present management information for more timely, thorough business solutions with less effort." The GriD Compass was described as "the first personal, portable computer designed to connect you to all the resources within your organization, outside information services and to GriD's advanced management software tools."

The original GriD Compass 1101 included many state of the art components and modules: Intel 8086 microprocessor, 8087 arithmetic co-processor, 256K bytes DRAM, 384K bytes bubble memory (non-volatile storage), electro-luminescent active matrix display, 57 key keyboard, integral 300/1200 bps modem with audio, telephone handset and speaker, RS-232C/RS-422 serial port, IEEE 488 GPIB peripheral interface, GriD link LAN, and a clock/calendar with lithium battery back-up. It weighed a little more than 8 ¬? pounds and consumed 60 Watts of power. The GriD Compass was manufactured in Mt View, CA in 1980 and 1981 (in those years, off-shore manufacturing of computers was not nearly as prevalent as today). They had a full, integrated suite of software which all fit in 256K of memory. GriD could also handle "compound documents" at that time

GriD Central was described in a brochure as "a 24 hour service to provide advanced applications software and remote file storage over the phone (line) to any GriD Compass user." It was based on a large fault tolerant, multi-minicomputer configuration, located at the GriD headquarters building in Mt View, CA. The GriD Central was actually based on a cluster of IBM Series/1 computers with mirrored IBM 3330 compatible storage (from CDC).

GriD Compass Central, based on the Intel 80186 microprocessor, was intended to link up to 32 GriD Compass laptop PCs together as a personal version of GRiD central. It was described in a brochure as "a compact table top computer". Said to be a lot like Apple Talk, the GriD Compass Central's capabilities included remote file storage and printing. Compass Central was said to "let you share data and peripherals and exchange messages directly with other Compass users."

Glenn noted that the GriD Compass was used extensively in space exploration. It was the first portable computer to be taken into space via the NASA space shuttle. NASA modified the GriD by replacing the modem module with a Shuttle Bus interface board, designed by Rockwell International. That enabled the GriD to communicate with on board instruments. The GriD Compass flew on all of the NASA shuttle missions. It also flew on Air Force One (the U.S. President's private plane).

The GriD management team was "obsessed with briefcases." They wanted this portable computer to be carried in a briefcase by mainstream business professionals. In an April 1982 Business Week article, the GriD Compass was touted as "a Porsche for top executives." In December 1982, Fortune magazine named the GriD Compass one of the Products of the Year under the category of "Smallest Computers."

The end of the GriD era came in March 1988 when it was announced that Tandy was buying GriD Systems Corp. The deal was approved on June 28, 1988 at a special meeting of shareholders. Glenn said that MSFT later re-introduced the product Tandy acquired as the "tablet PC."

Panel Discussion led by John Markoff

Each of the three other panelists stated their role at GriD. Carol Hankins headed up the software development team, after working at Zilog (1 year) and Xerox Parc (for 17 years). Craig Mathias was originally hired by Carol to work on an Operating System for the GriD computer (GRiDOS). After two months into that effort, Craig became manager of the "Load Center" project, which later became the Grid Central network. Dave Paulsen left Apple Computer to co-found GriD. He was enticed by "the challenge of putting all the interesting technology together in one box and making it work."

Markoff asked the panelists if they all had the Dynabook (Alan Kay's concept) in mind when they designed the GriD Compass? Mostly they had John Ellenby's briefcase in mind- whatever they produced needed to fit into it. The Dynabook inspired the concept of a portable computer as a stand- alone device.

However, the marketing and sales team got the pricing wrong. Instead of forward pricing the GriD Compass to create elasticity of demand and a larger market, they used bottom up pricing based on the typical 3 X the bill of materials. The $8,150 per unit price was very expensive (especially in 1982) and put the unit out of reach of many business professionals. Evidently the company did not have the "deep pockets" to lose money initially and make it up on volumes, which would stimulate the future market for the GriD Compass.

GriD Compass did drive the creation of flat panel displays. It got five Japanese companies excited and interested in that technology (they later perfected and mass marketed flat panel displays). This was an important legacy for GRiD.

How did the GriD team stumble on the (ultra-secretive) national security market? Craig quipped, "We can't talk about that." Glenn chimed in that "we had a customer located at Freedom Circle, which we later discovered was the CIA." The spooks liked the rugged capability of the GriD Compass. At $8,150 the military thought it was a bargain. The Compass "took off at the highest levels of government and filtered down from there." For example, the White House Communications agency showed strong interest in the product.

What about some other attributes of the product? The first GriD Compass ran on AC power (first battery powered machine came years later). Radiated electromagnetic energy was not an issue for the GriD Compass, so it did not interfere with co-located electronic equipment.

One unique feature of the Compass was its use of bubble memory modules for primary, non-volatile storage. The NSA and other government agencies were worried that someone could steal the machine and then have access to the stored classified information. To prevent this from happening, Carol's team wrote special software that enabled the security conscience user to erase the bubble memory in the event of emergency. She referred to this capability as "scrubbing bubbles."

Markoff asked about the roots of the clam-shell design. Where did the idea come from?
A PC that laid flat was difficult to use, while one that could bend forward was a much better design. Bill Moggridge (of the IDEO design firm) generated many hand drawings before the team settled on that design. A magnesium case was chosen as a much better alternative then plastic material. Some spooks were concerned about how to destroy the laptop in the event of a threat to steal it, but there was no self destruct capability added for them.

What was the corporate culture like? Craig said it was unique in many ways. He has never seen such a competent team elsewhere in his entire career. Everyone came from different (engineering) disciplines. John and Glenn had a lot of experience with UNIX. There were software people hired from HP, Zilog, and IBM (Series 1 programmers). Carol assigned her newly hired software engineers to work on something they had never done before. This resulted in build up of a lot of software expertise - the team members had acquired experience in other member's areas. Work became a very software engineering oriented, collaborative effort. Carol stated, "our mission was to become profitable before we ran out of money."

There were modules built by other companies that were integrated into the GriD Compass. To get the "IBM Selectric" feel, the keyboard was outsourced to a company named Rafi. Racal Vadic built the integral Bell 212A compatible modem for the Grid. It was a great challenge to fit that 1200 bit/sec modem into the unit. Full telephone support (complete with auto dial and pre-stored phone numbers) was available. There was an additional charge for the optional telephone handset.

A fair amount of Xerox experience was put to good use in design of the User Interface (UI). Carol led the group that did the UI design. She named some of the group members and said the UI was very consistent. It had something the marketing team called "leveraged learning." Once you learned the set of commands required for one application, you could use those for other applications as well. Hence, you leveraged the command set that you learned.

As well known VC Ben Rosen invested in both GriD Systems Corp and Compaq Computer, Markoff asked, "Why didn't GriD become Compaq?" The panelists provided several reasons. They stated that the entry price of the machine was too high, that it was too early to market (as a portable), and that it did not provide IBM PC compatibility till 1985 (so could not run programs written for the DOS/MS DOS operating system). Another reason for not merging the two companies was that GriD Systems was focused on direct sales, whereas Compaq zealously pursued the retail market. Glenn opined, "we should have raised more money, but we were trying to get to profitability quickly." Carol closed by saying that "we didn't move to plasma displays quickly enough." The GriD electro-luminescent active matrix display used considerably more power then the early Compaq PCs.

Software was a big advantage for the GriD. They were the first supplier of an integrated software package. It included a multi-tasking OS, text editor, spread sheet, business graphics, Basic interpreter, and Navigator user interface. It was easier to transfer data between applications then with other PCs. The team never thought of unbundling the software from the hardware, as they thought that would negate their integrated hardware/ software advantage. By 1985, the PC market was moving strongly towards MS DOS and away from proprietary OS's. The latter proved to be a big disadvantage for GriD.

The end game for GriD came in 1988 when Tandy acquired the company. The GriD people believed that there were at least two advantages of being acquired: 1] Tandy's retail sales force would generate higher sales volumes, and 2] lower cost hardware would result from the components that Tandy could buy at deeply discounted prices. Tandy thought they would be getting a direct sales force (from GriD) to complement their own retail sales network. So each company wanted what the other had and they didn't. The panelists were not around after the acquisition so couldn't comment on what happened to the GriD after that. Markoff said, "It fell off my radar screen."

Answers from the Question and Answer Session (all audience comments excluded)

The gross profit margin on the Compass 1101 was "pretty good." But the net profit was negative for some time (the company incurred operating losses). The original production run was 3k units per month for first few months. There was one known departure from the "3 x bill of materials" pricing. When designing Grid Link and computing its cost, the company charged $8 for a terminating resistor that cost them only 35 cents.

What about being one of the last computer vendors to have a proprietary OS? In order to grow in the portable PC marketplace, IBM PC software (i.e. OS) compatibility was essential. That was the big push behind the Grid Case- a product introduced in 1985. The waning interest in the Grid OS and ultimately its demise was caused by its lack of compatibility with MS-DOS. Grid Central was going to migrate to UNIX, but it did not happen because it was perceived to be too difficult to port.

GRid Systems Corp had eight different target markets. The oil industry was just one of those vertical markets and not the sole focus of the sales force.

GriD Central had a concept of tokens to purchase software. You could buy software on-line. That is, you were able to download new software if you had purchased a token.

A member of the audience displayed an HP 75 (a portable computer that predated the Grid). It included a BASIC interpreter that could handle complex variables. The panelists suggested that the user would have been a lot better off if he had used a Grid Compass instead of the HP 75. [Craig Mathias reports that GriD Systems Corp didn't use HP 75s, but did have an HP-85 that was used to test GPIB (IEEE-488) peripherals and to experiment with graphics programming constructs.]

The corporate culture was one of innovation. Building the 1st laptop (Compass 1101) was a real gamble. No one knew if it was going to sell. There were other products developed by GRiD that were later picked up by major companies. What MSFT now calls the Tablet PC was pioneered by Grid as the "Convertible." The Grid Pad introduction was neck and neck with Apple's Newton (an early PDA).

Mathias remarked, "Clearly, we were not good at marketing." [After the session, Craig clarified that remark: "Well, it's not that we weren't good at marketing. But we pioneering an awful lot of innovations at once, and the right marketing approach wasn't obvious."].

Grid was used to launch every satellite off the space shuttle from 1984-1987. The launch sequence software for the shuttle was actually programmed into the on board Grid Compass. The GriD could communicate with all the electronics gear on the shuttle via a Rockwell built communications card that replaced the integral modem. Later, NASA programmers developed quality of life applications for the GRiD. These included calendar scheduling for the astronauts. NASA called the GRid - a Portable On board Computer. In the Jan 1986 Challenger shuttle disaster, the Grid's were recovered and they still worked.

Grid was on Air Force One, but was it also the "President's football," which could be used to activate nuclear weapons on the President's command? The panelists could "not confirm or deny" if any U.S. government agency had nuclear missile launch codes programmed into the Grid Compass.

There was no competition with the Data General 1 minicomputer - Grid Compass was more primitive and not as expensive. Gavillon was a more serious competitor, but it was not in the same league as the Grid Compass, according to the panelists.

1st round of GriD Systems Corp funding was seed funding from individuals. $2.5M was raised in the 1st venture funding in Aug 1980. The three original VC investors were: Institutional Venture Partners. Citicorp, and Mayfield fund. $3.5M was raised on the 2nd VC round. From GriD's start of operations in August 1980, the Venture firms' requirement was for the company to make 10 working machines within one year. Indeed, the team completed this difficult goal one year later - in Aug 1981. Impressively, GriD Systems was able to bring the first product to market with only $6M of VC money- a paltry sum by today's standards!

The Grid experience influenced Jeff Hawkins when he started Palm. He related back to the Grid experience of having a lot of software operate within a very small machine. The powerful software that the Grid team developed, with the small amount of hardware resources required to run it, was really astounding.

Raj Shaw (an early Grid employee) asked the members of the audience that worked at Grid to stand up. Incredibly, it seemed that at least ¼ of the audience did so. A wide round of applause followed.

John Markoff thanked the panel members for their great stories and asked the audience to give them a hand. More applause followed.

CHM CEO and Executive Director John Toole closed the session by thanking John Markoff for his encouragement and stimulating questions to the panel. Toole stated that it was really exciting to get a team together (the Grid team here tonight) that had made computer history. He thanked the panelists and offered them a small gift as a token of the museum's appreciation.

Additional Resources

Pioneering the Laptop: Engineering the GRiD Compass (CHM event announcement and video of entire panel session.)