The Fonly Institute

Running Into Reality

Running Into Reality

28 Dec. 2009

It's June, 1981 and we are racing to make the first shipments of Osborne-1 computers so Adam can win a bet with Bill Godbout, one of his poker buddies. The amount is unknown, but Adam's self-image is apparently at stake, so June 30 is our deadline.

It's surprisingly difficult. Everything has to go together right – every piece has to be on hand and in working condition. Documenting the process is a full-time job, one that is apparently not being done. Assembly documentation consists of a series of “assembly trees” - lists of what goes together to create the next higher-level assembly. It's manageable when you're documenting complex assemblies like the circuit board, with hundreds of parts. After a couple of levels it gets to the actual computer people will use, and you think that it's done.

But there's more. The diskettes with the software have to be listed, and they can have different versions. Then the whole thing has to go into a box along with custom-cut foam packing pieces, so that's another assembly tree. And somewhere in there there has to be a power cord that will fit the sockets in the country of use. It gets messy fast.

Manuals in the right language, warranty information sheets, labels for the boxes and the pallets of boxes – the assembly trees pile up. It elicits the phrase from Ginsburg's famous poem “Howl”- “Trees, radios, factories - tons! They broke their backs lifting it to heaven!” (well, you have to read the whole thing).

And then you have to ensure that your people follow the instructions. Some of the first units off the line had incorrect wiring in the power panel – the small square black-anodized aluminum sheet where the power cord enters and the circuit breaker and interference filter are located. All the connections are done there with push-on wire terminals, and someone's hand has to choose the right spade lug and push the terminal onto it. Just providing a good drawing of what goes where doesn't always make it happen. Fortunately the circuit breaker will trip when you plug it in, but that doesn't get you very far, and a hot line might be connected to the metal itself where the ground wire should be.

Thus it came to pass that on June 30 the first batch of six Osborne-1's were stuffed into their boxes and loaded into someone's car trunk for the trip to the dealer. We had made it! Adam's reputation would be unsullied! It didn't matter that on July 1 the same computers came right back to be repaired – we had shipped on time!

Not long after that I sat in an executive committee meeting and heard a report from some private investigators who had been hired to go undercover in the work force and ferret out drug use. They concluded that they could find no drug use but that we had a serious quality control problem. This led to the establishment of “quality circles” - committees at the employee level to compare notes and identify quality problems and solutions. Things began to get better in manufacturing.

Osborne's manufacturing process involved a lot of “outsourcing” - not necessarily the international kind that became universal 30 years later, but within Silicon Valley. The circuit boards were assembled by a company named “Testology” in San Jose. They would sell boards to Osborne and warranty them as functional – so if there was a problem later they would take them back and fix them.

Unfortunately, Testology created their own test procedure and Osborne wasn't much involved in the process. Such an arrangement requires that both parties agree on what the test is to be to determine functionality. As the designer of all the timing circuitry I should have been heavily involved in that process. The most I can recall was having dinner with Testology's test procedure consultant. It was a good dinner, but when I began making reference to the operation of the 74LS161 and 163 synchronous 4-bit counters (building blocks I used quite a bit) the consultant went blank - he had nothing to say. It eventually came out that he had represented his qualifications as better than they really were, and that he did not understand what I had been talking about.

Thus it transpired that boards would pass the outgoing tests at Testology and be shipped to Osborne – where they wouldn't work and would be shipped back to Testology. They would pass again there, and would wind up in a loop going back and forth between the two companies. When the end came all of these boards were simply sold to the highest bidder and entered the population of spare parts.

If I had been properly qualified for my position of Vice President of Engineering I would have been alert to any sign that something was amiss – that would have been my job. But as it was I considered myself basically the chief engineer, with a staff of one and a number of expectations which I had no good plan for meeting. No one ever sat down with me and asked me to lay out my organizational structure, my priorities, and the resources I would need to make it happen. No one had told me “Lee, you can't do that with just two engineers, especially if you're one of them!

Later I learned that the VP Engineering represents the engineering department to the Board of Directors. It is an outward-facing position, and the inward-facing position is the Engineering Manager. In startup companies these two positions may be combined, but it is foolish to combine the function of Chief Engineer with these positions, as I had done. What I discovered was that in such a structure where the product is even modestly complex none of those jobs will get done!

In this situation I was trying to get the “double density” upgrade for the floppy disc system designed, and was looking for a consultant who knew the technology (give me credit here for knowing it would take too long for me to learn how to do it). I had made an appointment for one consultant to come and meet with me at a given time, which came and went. I looked for Adam's secretary Marlene, a mature Englishwoman who must have reminded Adam of one of his childhood nannies. She would have been in charge of receiving the visitor, but she wasn't there.

Marlene had had to leave her desk for some reason and, having no backup, had locked the front door! My consultant candidate had showed up on time, found the company locked up tight, and had gone off incensed. Those were the days before cell phones, so he couldn't have placed a call from the front door, and in that industrial park the nearest pay phone was nearly impossible to find.

A short time later that day I met with Adam, along with Ed Richter, an engineer referred by our General Manager Tom Davidson, who had worked with him when he turned an electronic module manufacturing company around. Ed was Dutch, a former KLM Airlines flight engineer, and quite a bit more formal than most of us. I was to report on progress on the double density system, and had to say that my consultant had gone away furious.

Adam began a tirade about how I was letting down “...all those people who had been busting their asses...” when I snapped. Sitting bolt upright and gripping the chair arms, I barked back at him “one of whom is sitting before you now!” Adam, taken aback, was shocked into silence, his head making small bobbing movements. I went on, in my most strident, drill-sergeant voice, “Do you care to conTEST that statement?”

It took several seconds for Adam to recover his voice. “Lee, I'm not going to sit through another one of your tantrums,” he said in an overly-controlled voice, as if I had been popping off several times a day. He went on to lay out what he intended to do – I would be relieved of my position of VP Engineering and Ed Richter would be made acting VP in my place pending approval of the board. I would be given the newly-created position of “R&D Fellow” with full Vice Presidential rank. Clearly this whole solution had been worked out in advance.

When Tom came into the meeting and learned of this he came to my defense - “It wasn't his fault the guy couldn't get in...he woiks continuously!” but he had either agreed to this change or hadn't been in the loop. Ed and I left the meeting with Ed muttering about how he didn't like that kind of emotional outburst, and I had been freed from some responsibilities I wasn't meeting anyway.

Some time later Ed and I went out with a real estate agent to look for a location for our R&D shop which I would run. We settled on a suite of offices a block or so away in the same industrial park. I signed the lease papers myself as an officer of the corporation on a weekend when Adam and other VPs were away. Ed, Pat and I moved our desks there and I was in startup mode again. I was much more in my element.

No Way To Run A Company

How Not To

4 Dec. 2009

It's late fall of 1980 and I'm working every available minute to get the Osborne-1 design ready. The work is going on in multiple locations – my office in Berkeley, the offices of Sorcim (“micros” spelled backwards) in Santa Clara where the firmware (what would now be called the “ROM-BIOS”) is being developed by Richard Frank and his crew and where Tom Davidson, our general manager, maintains an office, and at times in the unfinished third of a tilt-up building on Corporate Way in Hayward where I do thermal testing of the “unit”, as we call it. I put in a lot of time driving from one site to another in my 1977 Honda Accord.

Sorcim has the wire-wrapped prototype with a printed-circuit DRAM section laid out by my electronic musician friend. It's just a circuit board with spacers for legs holding it off the tabletop, a tiny 5 inch CRT monitor (diagonal, meaning the picture is 3 inches high by 4 inches wide), a 5 inch full-height floppy disc drive and a keyboard obtained as a sample from a keyboard manufacturer. When there is some question about the hardware operation I have to pack up my oscilloscope and drive the hour-long drive down there and fix it. At least once I am so tired at the end I can't make it back and wisely check in at the Vagabond Inn motel next door.

A couple of times Adam Osborne himself comes down, ostensibly for a conference, accompanied by a young woman co-worker from his book company. After strutting back and forth a few times and chatting briefly with Richard, Adam would draw himself up and say “well, you guys obviously know much more than I about these things, so we'll just be going”, and the two of them would leave. Clearly the purpose of their trip was not to confer with us but, in a sense, to confer among themselves.

During this time we try to get a metal housing made at a reasonable price. Tom takes me to Galgon Industries in Hayward , where Manfred Galgon and his crew have turned out ten metal cases for our first run of prototypes. Manfred is an immigrant with a thick German accent, and proud of his heritage. “Did you ever see a Cherman miss?” he would challenge. I briefly consider replying “Yes – out in the front office” where one of Manfred's daughters holds down the front desk – but I hold my comments.

Tom is trying to negotiate a workable price for a metal case in production quantities – the quote stands at $150 each – far more than we can afford. Even though Tom had been trained as a machinist he takes his favored approach of “beating up suppliers” in trying to argue the price down. Manfred was having none of that and steadily grows angrier. Suddenly, he erupts, pointing at me while looking at Tom, “him I will talk to!” Swinging his arm to point at Tom, he continues rapid-fire, “him I will not talk to!” and everybody gets out of the room. I would have liked the chance to talk with Manfred about better ways to get the price down, but I have to go with Tom to a nearby restaurant and listen while he laments over a few beers. Galgon remains our supplier for several sheet metal parts, an indication of Tom's respect for Manfred.

Not too long after that episode I am taken to the offices of GVO, an “industrial design” firm, for a conference about designing a plastic housing. The executive we meet with wears a European double-breasted suit, an unusual sight in Silicon Valley, clearly meant to make us feel uncomfortable about our ordinary clothes. After a while, he calls in the actual industrial designer on the project, a tall, prematurely balding young guy named Mike Levitt. He lays out a plan for two plastic housings – the first done by the quick-and-dirty “vacuum form” process in which a sheet of heat-softened plastic is pulled down over a wood and metal mold and vacuum is used to pull it around the mold. The second case, which will take months longer to prepare, would be the usual “injection-molded” type where molten plastic is forced by pressure into a machined metal mold.

Tom goes with the GVO plan, and I begin coming down to GVO to help Mike work out questions and problems. Very quickly I realize that the executive in the fancy suit will be of no help in the process, and discover that I can get in to see Mike through the back door on weekends and after hours when the executive is away. From this I construct the aphorism “always find the back door into a supplier” - that is, develop a direct channel between the technical people on both sides of a problem to get it solved in the fastest and best way.

Industrious as ever, Tom finds a plastics fabrication company in San Francisco, Ajax Plastics, that is willing to throw themselves into the task of tooling up to manufacture our beige vacuum-formed housings. Their lead technical person, Paul Bunning, is a short, wiry fellow who calls himself a “plastics artist” and is willing to put in the long hours needed to prepare for production. They built three molds and arranged them on a turntable so one would be heating, the other “pulling” the plastic sheet and the third “trimming” and removing the cooled sheet. Tom commented at one point “I had to buy a plastics company to get it done.”

Paul's artistic credentials are displayed by a wall hanging consisting of a large sheet of plastic with three ghostly impressions of the Osborne housing bulging part way out. It produces an eerie effect, as if they were invaders from another dimension who had almost made it through to our reality. Some years later, on another project, Paul takes great pride in designing a housing that would be far cheaper to build than the sort favored by industrial designers, and in charging one third of what they would charge. Unfortunately, I can now no longer locate him, and I suspect his tobacco habit combined with the fumes expected in plastics manufacturing brought him to an untimely end.

At the 6^th West Coast Computer Faire in March 1981, where we are the hit of the show with our new product, Adam and I find ourselves walking down an empty corridor behind the scenes. In an expansive mood, he claps me on the shoulder and said “the product is the company – we're going to build it up really big and sell it off really fast!” I hope he knows what he was doing, because in those areas, I certainly know nothing.

At the show no spectator could get close enough to the product to see that the display was wavering from magnetic interference between the power transformer and the CRT monitor, that random characters would appear on the screen, that the insides contained a large amount of double-sticky foam tape, our universal mechanical fix, or that the unit weighed 30 pounds due to the large iron and copper power transformer. Carrying two of them from my car four blocks to the show had nearly pulled my arms out of their sockets.

At the show a man I'd never met introduces himself and pulls out a small electronic assembly – a “switching power supply” from Astec that would replace our huge, heavy power transformer, eliminate the waver I called “the Hawaiian effect” and greatly reduce the amount of heat within the device. It was a godsend, and I immediately design it in.

Around this time Tom confides to me that he is not using purchase orders – his parts orders to suppliers are made verbally. This is unheard of, but it is part of his strategy of negotiations. If the suppliers want to, they could shut us down in a day without legal consequences. It is only their perception of mutual self-interest that sustains the whole house of cards that is the parts supply structure. If we ever falter, Tom makes clear to me, we would face disaster - the parts would keep coming only as long as everything looked good.

Upon starting production Adam demonstrates how little he knows about running a company of any significant size. He attempts to get everyone in a room around a table so he can hear reports and issue orders. You simply can't do this with a company that's larger than about 15 people – management is a full-time job and most of what's discussed is of no relevance to most of the others there. It will radically reduce the company's productivity.

By that time, with the title of Vice President for Engineering, I had set up the engineering department with a totally inadequate two engineers (Pat McGuire and myself). There was a documentation clerk who asked for much less in salary than we thought she ought to have, and a part-time draftsman. I had made a fairly common error of confusing the role of “vice president for engineering” with “chief engineer”. I tried to solve all the problems this brought about by working very long hours, often into the early morning hours. Pat puts together our floppy disc duplication system – an Osborne board with four floppy drives and enough modifications of hardware and software to make it work. Pat often works alongside me when I ran late.

One day Adam and Tom are out of town and we are having our company meeting. I am told by our 23-year-old marketing VP that there is an unacceptable level of errors on the machines we were shipping, and that production has to be stopped until this is fixed. Remembering Tom's warning about never slowing down, I dig in and resist. With the others away, I am the one with authority over production, so I hunch up my shoulders and declare “I won't do it!”

Marketing then demands that we ship all production to them before shipping to the customers. They will test the machines and let them out only when they see fit. I agree, since it would not slow down shipments of parts and endanger our supply lines. Marketing then rents a vacant third of our building and hires a raft of temporary workers to walk up and down along rows of computers, aimlessly poking discs in and out and running programs.

When Tom returns he backs me up. The meetings continue until Pat looks up and asks “has anyone ever cleaned the heads on the duplicators?” Everyone looks at each other. The fellow responsible for this sort of thing had previously been a hard disc drive technician, where the first question asked is always “did you clean the heads”? The same day we place an order with a professional disc duplicating company and our homemade duplicator setup is returned to a shelf in the engineering shop. The errors, which have been due to weak recording of the floppies, end.

The company expands into the space used for marketing's “testing” program, and almost all of the temporary workers stay and become permanent. This is no way to run a company, a lesson I am learning too late to apply.