Macworld Magazine, December 1990

The Iconoclast by Steven Levy


The Soul of a New Macintosh

The Twisted Tale of the Mac LC, The Computer That Refused to Die



In early 1988 the top executives of Apple Computer had many priorities for improving the Macintosh product line. Making a truly affordable color Mac did not top that list. But fortunately for them—and for us—two unheralded engineers realized that there would be considerable clamor for such an item. They resolved to design a color Macintosh that could be produced at a much more modest cost than the current rage, the IIcx. The result of their labors, as well as the good efforts of literally dozens of Apple workers, is the Macintosh LC, described in detail elsewhere in these pages.


But it was not easy. It was certainly not easy. The road to the Macintosh LC was to be pocked with detours and twists, treacherous obstacles and tempting exits. A stark recitation of the various code names for the project gives an indication of the shimmying course: Strange, Spin, and, finally the eponymous Elsie. It took resolve, a steady vision, and even some judicious insubordination to bring the vehicle home, ultimately parking it in the space marked Victory Circle.


Victory, however, was the furthest thing from the minds of Paul Baker and H. L. Cheung in the early days of 1989. Management had just axed their pet project, a color computer called Spin. They had finished the prototype around Christmas and had eagerly demonstrated it to any Apple executive they could collar. They had even gone to Apple's production facilities in Singapore to look into the potential cost of manufacturing the machine. But though Apple's brain trust generally agreed the computer was a fine one, Spin was deemed not enough of a change from previous models to justify a go-ahead. Baker and Cheung had a dead project.


Baker and Cheung

It was a rough blow for two veteran Apple hands. Baker, now 35, had worked for Apple since 1979, and had spent five years as a key player on the Lisa team before leaving the company to work for the Macintosh's original designer, Jef Raskin, at a company called Information Appliance. When that experience reached its end, Baker returned to Apple, hoping to finally associate himself with a hit product. H. L. Cheung, two years younger, had been the ninth employee Apple hired in Singapore. After immigrating to Cupertino in 1987, he headed the Apple II-family design team. Then he wangled a transfer to the Macintosh side, where the action was.


Since returning to Apple Baker had been working on a project codenamed XO, pulling some of the Mac SE's features into a chip set that could be retrofitted into the Mac Plus. Cheung, the manager of Baker's team, was working on two stripped-down modular Macs called Green Jade and White Jade. Those projects fizzled, but the vision behind them made sense: quality Macs for lower cost.


The idea resonated with each engineer's personal orientations: Baker had two children, and was passionate about Apple maintaining its place in the education market. Cheung was still drawn to the mystique of the Apple II, a computer first conceived as a color machine for The People. So why not a Macintosh iteration? To quote the system description Baker circulated in 1988, the prospective computer would be "a small color Macintosh system based on the Mac II design." He listed the features that would distinguish this machine from its predecessors:


  1. Small size (single floppy base configuration is only 90mm tall)
  1. Single processor expansion slot (no NuBus interface)
  1. Built-in video subsystem (no Mac II video board needed)
  1. Expanded system accommodates larger power supply
  1. A low-cost RGB monitor based on Apple IIGS for reduced cost


Baker and Cheung worked on the computer in off-hours during the early months of 1988. Eventually, of course, the project needed approval from above, so that the custom chips could be designed and manufactured. It was at one of these reviews that the machine–then called Spin–spun off course. The superiors decided to scrap Baker's idea to use a low-cost monitor; they dictated instead that the computer use the same video scheme being implemented in the high-end Macintosh IIci, then being readied for market. This would mean that Spin would require the powerful–and costly–68030 microprocessor.


At that time, Baker temporarily left the project to work on yet another high-end Mac, the IIx. He returned, in October 1988, in time to ready prototypes of Spin. The computer looked like an Apple IIGS with the power, and the same slick high resolution color, of a full-size Mac II. Twenty prototypes were produced. Baker and Cheung distributed them through the company "to anyone who would take it," says Paul Baker. Yet the project was canceled. The low-cost color Mac, which by that time was not so terribly low cost, seemed a fatality.



Spin Control

If that was the darkest hour for the project, dawn was quick in arriving. The fortunes of the mother company were flagging, and the myriad critics of Apple Computer were quick to identify the cause—its computers cost t oo much. Almost coincidentally with the squashing of Spin, John Sculley was promising a crowd of hostile stock analysts that Apple would one day dazzle its customers with a budget alternative to its current offerings. What he didn't say was that Apple had just snuffed its most viable low-cost Macintosh development effort. Fortunately, the project could be revived—but now that everyone would judge the computer by its price tag, the cost would have to be brought down considerably. The Spin team was summoned to the office of then-Apple VP Jean-Louis Gassée to discuss what adjustments would be made to make Spin a truly low-cost computer.


Speaking first at the meeting was Bill Goins, a former Dynamic engineering manager who had joined Apple some months before to manage the Spin project. Armed with some props—cheap color MS-DOS clones like the Tandy—he outlined various design alternatives. Since you couldn’t have all the features of previous Macs at a fraction of the price, he argued, some beloved features of the Mac II would have to go. Jean-Louis Gassée had his own point to make. Leaving his seat and literally getting down on his knees, he implored the engineers: Please! Make it color! So the low-cost Mac would be color. Befitting the project’s new high priority within Apple, the team was beefed up considerably. There was soon a squad of engineers and designers who would lay out the circuit board, program the new ROMs, design the case, and perform the dozens of other jobs required in the creation of a new machine. But, as Bill Goins later recalled, “First we had to have the The Plan.” A controlling vision that would dictate what to put in and what to leave out.


In early 1989, The Plan was to cut costs to the bone. As a result, the prospective computer (now code-named Elsie) began to shape up as a kind of low-rent color Mac, with the penny-ante Apple IIc as its role model. It’s almost shocking to contemplate now, but the prototype they produced that spring resembled a small one-piece box like the Apple IIc with a built-in keyboard. It had a single 800k floppy drive that could be used without a hard drive, since much of the Mac system software would be embedded in the ROMs. The power supply was lodged on the wall plug. It included a single megabyte of RAM. The machine would be able to output only 4-bit color, a crude alternative to the Macintosh 8-bit standard. Powering the computer was a simple 68000 chip, the same one found in the Mac Plus.


This version could indeed be produced cheaply. The problem was that it wasn’t a very good computer. The processor was overmatched—it took aeons to draw the color. Obviously, a more powerful processor was required, as well as more storage. It was a tough decision, because the more powerful the processor, the more expensive the computer would become. Ultimately a solution evolved: there would be more than one low-cost Mac. Around April, a team was spun off to work on the original Spin design, to create what would become the Macintosh IIsi. And a few months later another team would break off to produce the black-and-white Macintosh Classic, using the 68000 processor. (Some of Paul Baker’s XO work found its way into this computer.) The LC itself would navigate a middle course, using the Motorola 68020 chip. All the computers would use a hard drive.


“The Plan was to do it without a hard drive. But it really isn’t a Mac without a hard drive,” admits Baker.


Since low-cost manufacturing and components were essential, Baker and the other engineers practiced a frugal, almost miserly form of design. At one point, there was a furious controversy over whether a plug should terminate in a male or female connector, the distaff alternative being chosen to save a few cents. Perfectly good components used in previous Macs required a redesign so they could be produced at a slightly lower cost. Making a new keyboard from scratch saved $2.50 per machine. And do you wonder why there’s no way to hook an external floppy disk drive to the LC? Hey, a connector would have cost $1.50 per unit.


While such vigilance could slice a few dollars off the production cost, the real savings came from redesigning the system from the ground up. “We learned something very important about cost reduction,” says Cheung. “You can’t do it by eliminating features.”


In fact, the team was eager to introduce entirely new features into the Macintosh family. The strategy in these cases was to build the feature, get it to work, and only then seek approval. “The key to getting something accepted is to make it a fait accompli,” explains Paul Baker. A prime example is the voice input capability now included in all the new Macs. When engineer Eric Harslem suggested it, everyone agreed it was a neat idea. A custom sound chip was designed, and the LC team implemented a scheme whereby the user can simply pick up the mike (included with each computer) and verbally annotate documents. “When [the executives] heard it, they were convinced,” says Baker.


Lowering the Boom

As the project stretched on, the Elsie team began feeling the pressure. After all, consensus had it that the fate of Apple depended on the success of these low-cost Macs. People who wandered into project manager Bill Goins’ office would be startled, even unhinged, to see a chart indicating that these new computers were destined to replace the Mac SE, the Mac Plus, and the Mac IIcx— the products that, basically, kept Apple in business. “I felt the strain,” admits Cheung. “The whole company was on our back.”


The designers had to fight off the deadline pressure and focus on the computer itself. “Even though we knew it was important to get it out,” says Baker, “we had to do the right thing.”


Every detail took on importance, and that certainly included the physical appearance of the new computer. Once the built-in keyboard was scratched, the designers considered all sorts of weird approaches. For a while, recalls Bill Goins, people were charmed with a “boom box” shape—a long rectangle that you could, if you wished, transport on your shoulder. But since a boom box cannot support a computer monitor, that idea died a wistful death. Eventually the industrial designers came up with a distinctive shape—a flat box, something like a squashed MacBottom disk drive, with a slightly sweeping curve in front. Originally, Jean-Louis Gassée objected to the curve. But the Elsie team, resilient to rebuffs, decided to take an internal poll on the matter. The vote was 22-to-1 in favor, when John Sculley chanced by.


“What’s the score so far?” he asked, and was informed of the tally.


“Make that 23 to one,” he said, and that settled the matter.


The designers had to make one more big adjustment on the LC—its method of producing color. All along, the idea was to use 4-bit color, which would demand less of the processor and the monitor. But the first few developers who saw the prototype complained. The Macintosh standard, they said, was the high-resolution 8-bit color of the Mac II. Why make wimpy Macs? The objection made sense. “If we’ released it with 4-bit color, we’d have a new (and inferior) software standard—we’d be living with it forever,” says Wayne Dyer, the Macintosh product marketing manager.


The design team members came to realize that despite the expense it would incur, 8-bit color was the right thing to do. Typically, they did not inform their superiors that The Plan had changed, but began to quietly implement it in their spare time. So by the time they revealed to the executives that the LC absolutely, positively had to have 8-bit color, the design team was able to show how this could be pulled off—indeed, they had done it.


Off the Books

That sort of anecdote is repeated over and over again when the designers recount the tale of the Mac LC. It was a computer essentially hatched from the bottom up. When the putative wise men in charge of the company would strike down part of the plan, or even the entire project, Paul Baker, H. L. Cheung, and the others kept going, even it it meant the things were carried off the books for a while. Persistence paid off—we now have the Mac LC. “This can be another Apple IIe, or Plus, or SE,” says H. L. Cheung, not doing much of a job of hiding his pride.


Speaking of the Apple IIe, one of the neatest things about the LC is its ability to emulate that computer with the addition of a board that sells for under $250. Just the sight of an old copy of Raster Blaster running on a Mac can make an old Apple II owner get all misty. Rather more significantly, this ability will enable the Mac to smoothly replace the hundreds of thousands of Apple II machines in schools—and probably rescue the company from oblivion in the crucial kindergarten to 12th-grade market.


But the Apple II emulation board almost didn’t make it out of the gate. In mid-1989, Jean-Louis Gassée killed it outright. Just said no. What did the engineers do? “The design team never missed a lick,” says Bill Goins. “They just kept working on it.” And when engineer Rob Moore finally produced a prototype that worked, the executive staff of course was convinced that the skunkworks emulation board operation was justified.


But that’s the story of the Macintosh LC itself: the low-cost color computer that Apple produced in spite of itself.