Sunday, September 16, 2012

The Soul of a New Machine

I vividly recall reading "Soul of a New Machine" by Tracy Kidder back in 1982. I loved the book because the hero of the story was an electronics engineer. There are plenty of books and movies and TV series about doctors and lawyers and police officers and even news reporters. But here was a book where the hero was an engineer.

In preparing his book, Tracy Kidder got a preview of this world in the late 1970s when he observed the engineers of Data General design and build a new 32-bit minicomputer in just one year. His book starts with a description of Edison de Castro who is one of the founders of Data General. As a young engineer in his twenties, he led the team that developed the PDP-8 for Digital Equipment Corporation. Now leading his own business, he needed to respond to the latest development from his former company.

For those that don't know, the PDP-8 and DEC (Digital Equipment Corporation) started the minicomputer boom and the PDP-8 was both Bill Gate's first computer and also the genesis of the internet. It was the PDP series that led to the development of UNIX and the C language, and I cut my early computer teeth on a PDP-11 which was a more advanced model.

This was the age of DEC, Data General, and Wang — all Massachusetts companies — as the leading computers that I would call personal computers, at least compared to IBM mainframes. This was before Radio Shack and Apple and the IBM PC and the "micro" computers that became the personal computers we think of today. As micros increased in size and capability, they often adopted technology first developed on the minis.

In the story, Data General has moved their main development lab from Boston to Research Triangle Park in North Carolina. A few northeastern patriots refused to go south, and they remained behind to develop a new computer before the team in RTP could do it. It was a sort of "skunk works" development only partially endorsed by management. The company was expecting its team in NC to produce a new product, and commissioned the small Boston team as a "plan B."

At one point Edison sneaks into an office to look at the new PDP-11. He opened the drawer. (Computers in those days looked like file cabinets and had several drawer like sections that opened up to give access to the components.) He looked at the computer chips inside which conveniently have their model numbers printed on top. He did a quick inventory, figured out the total cost, and knew he could design a comparable computer at a competitive price.

At that time, minicomputers were 16 bit, and the PDP-11 was the first 32 bit mini. (Mainframes were 64 bit at that time — a key difference between mainframes and "minis." The minicomputers were growing up.) As we know today with the transition from 32 bit to 64 bit to 128 bit in Windows and MacOS, the wider the data path the more efficient … and faster … and — well — just better the design. Those were halcyon times. Ever since then I've kept a close eye on the component marketplace. I've read about new microprocessors and support chips with eager anticipation and I've written before about specific chips and my prognostications.

So, it was interesting to read the news in the Electronics Engineering Times, a newspaper for nerds, that Apple is now the leading chip purchaser, having taken the mantel from HP. That can be misleading because IBM is one of the largest chip producers and consumes its own products without public notice. So it may be that IBM is still the number one chip consumer, but, then again, the volumes that Apple is selling could even outdistance my old employer.

The news is that Apple Inc. became the largest buyer of chips in 2010, overtaking Hewlett-Packard Co. by spending $17.5 billion to grab nearly 6 percent of global production.  Apple's chip purchasing was up 79.6 percent from $9.7 billion in 2009. In that year Apple had been in third place behind HP and Samsung. It was sixth in 2008.

Apple spent more than 61 percent of its 2010 semiconductor budget on chips for wirelessly connected equipment such as iPhones and iPads. In contrast HP spent 82 percent of its 2010 chip-buying budget on chips for notebooks, desktop and server computers. This is clearly working to Apple's benefit. Smartphone shipments increased 62 percent in 2010 and tablet computer shipments exploded by 900 percent, driven by the debut of the iPad. Meanwhile global PC shipments, not counting tablets, grew by 14.2 percent in 2010.

As a result Apple is expected to keep on increasing its semiconductor spending during the coming years at an above-average pace. In 2011 Apple’s semiconductor spending is expected to exceed that of Hewlett-Packard by $7.5 billion, up from $2.4 billion in 2010.

Apple's surge to leadership in semiconductor spending in 2010 was driven by the overwhelming success of its wireless products, namely the iPhone and the iPad. Is it any wonder that Google (Roid), Microsoft (Win7) and HP (WebOS) want a piece of the action! These products consume enormous quantities of NAND flash memory, which is also found in the Apple iPod. Because of this, Apple in 2010 was the world’s number one purchaser of NAND flash.

Chuck and I recently traveled through four states stopping at almost every Apple store to buy an iPad.  They were always out of stock. We would call, they said they had some, and by the time we got to the store, they were gone. In one case, the last one was sold to the person just ahead of us in line.

Don't you wish you had that problem at your business? For those looking for a good investment, I suggest Apple Computer. The P/E is really low for such a hot stock and the future seems so bright that you will have to wear shades. The secret to success in the stock market is growth, and Apple continues to show growth that is amazing and no sign yet of slowing down, despite strong competition by other world class companies. They just can't keep up the the undisputed market and thought leader.

By the way, the Boston team of Data General did succeed and produced the Nova, their most successful product. And, today? Well, DEC was purchased by Compaq who, in turn was purchased by HP. Wang went broke. And Data General? It is just gone.

And how should we view Data General's legacy?

Maybe the most fiercely innovative companies simply can't help but flame out. It would have been nice if his company had endured, but what Edison de Castro and his small team did lives and breathes in the fact that computers are even more broadly integrated into our society.

Data General didn't last nearly as long as some of the other U.S. high tech companies, but it did leave a lasting influence. Shouldn't that be enough? Edison, now 64 years old (the age of yours truly) lives in the Boulder area. I always hope to run across him in a coffee shop and tell him how the book about him and his team influenced me and my career. Who knows, maybe we can become friends on FB.

Originally written on June 11, 2011.

No comments:

Post a Comment