Paul in his home office. Behind him is his Apple II system. The system was hand delivered to Paul by Steve Jobs for use in the development projects talked about below. The Disk II disk drives with the are serial numbered 3 and 4.

Two weeks later, I had left my safe, secure position with IBM to work at the three (now four) person company. Within days I had my very own Apple II work station. This computer was hand delivered to me by Steve Jobs. I learned that the Basic was go into Apple's next generation of computer, code named Apple Annie. Woz was very excited about this new machine. It was to have plug-in program ROM cartridges and lots of custom LSI.

One problem that we had to solve was that of getting 6502 object code files into the Apple II from our development system The development system consisted of a national COPS microprocessor with a 6502 compiler created by Bob Shepardson. The input to the compiler was a deck of punched cards. The output was paper tape. How do you read paper tape into an Apple II? Woz had the answer. He built a card for the Apple II that would drive a paper tape reader. One day while Woz I were setting up the tape reader, I noticed that Woz seemed depressed. I asked what was wrong. He replied that he had developed a floppy disk drive for the Apple II. He was really proud of that, but Apple's management had given him an impossible schedule for the delivery of both the hardware and the disk operating system. I said I could do the disk operating system (DOS). Woz was delighted. After a quick consultation with Bob Shepardson and Steve Jobs, Woz and I started to work.

When Woz showed me the designs of the disk controller hardware and software driver. I was truly amazed. At that time, all disk drive controllers were big cards with dozens of large and small scale integrated circuits. The design Woz created required only seven small scale integrated circuits. What was even more amazing was that Woz's design had significantly better performance ( data density, reliability, cost) than existing controllers. When Woz started this design, he did not look at how other people had done it. He thought about how it should be done. Using this process he created something remarkable. This became my real world example of what was to later be called "thinking outside the box." In my later life as an engineering manager, I have told the Woz Controller story to many engineers as way getting them thinking on a different path. Unfortunately, most engineers will never attain the level of Steve Wozniak's creativity.

It took Bob Shepardson and Steve Jobs a couple of weeks to negotiate the details (cost, schedule, re-scheduling the Basic, etc.) and sign a contract. If you looked at this contract, you should be amazed. The cost to Apple was very small compared to the value Apple received. The schedule was very short considering the work to be done. The product specifications, deliverables, acceptance criteria, penalty/bonus clauses and legal mumbo jumbo that I have become used to in the years since were all missing. Those were the good old days!

Now that I was no longer doing the Basic and because we had other work coming in, Bob needed to hire another programmer. I knew the perfect candidate. Kathleen O'Brien, my life partner, was a very good programmer - and we worked well together. A few weeks later, Kathleen was Shepardson Microsystems employee number 5.

During the time I was working on the DOS, big changes where happening at Apple. They moved from their small office space behind the Good Earth Restaurant in Cupertino to their new World Wide Headquarters on Bandly drive. (We used to joke about the World Wide part. Apple was a tiny start up that had just begun delivering its first production products. Shepardson Microsystems did a lot of projects for similar small start up companies that were going nowhere. Why should Apple be any different? At one point Steve Jobs offered to buy Shepardson Microsystems to form the nucleus for Apple's software development organization. Bob refused the deal. Steve was only willing pay for the Shepardson Microsystems with Apple stock. Bob might have been more interested, but Steve would not increase the offer beyond 10% of Apple's stock.)

One of the big changes came when Apple hired Jeff Raskin to manage Apple's technical writing group. The task of writing the user manuals for the DOS and the new Basic fell on Jeff's shoulders. The task was particularly difficult since no form of specification existed for either product. Jeff had a nearly finished version of the DOS to work with. For the Basic, all he had was the syntax checker part of the code. Jeff's solution was to write the manual as the specification. This was all well and good, but Jeff had big ideas. As he was writing the manual, the specification for both products grew well beyond the scope of the original agreements. For DOS, this lead to several follow on, last minute contracts to cover Jeff's additions. Fortunately, the scope of the DOS changes where limited due to shipping deadlines. This was not the case with the Basic.

Our little Basic grew and grew and grew. It would no longer fit in 48k of RAM. We were going to have to develop code segmentation and overlay methods. Jeff acknowledged the size of this monster by naming it NOTZO BASIC. We called it NutSo Basic.

The final chapter in our association with Apple came soon after a meeting with Apple in October, 1978. I have scanned Randy Wigginton's minutes of this meeting. The meeting.gif file is an small image of the introduction to this document. The meeting.txt file is an OCRed text file of the document. As you can see, the first part of meeting covers fixing some defects in the now shipping Apple DOS 3.0. The second part of the meeting covers discussion about the now bloated NOTZO Basic. I fixed the DOS defects by giving Apple a marked up listing. The Basic problem was fixed a short time later, when Apple canceled the Apple Annie project and the Basic contract. We at Shepardson did not mind. Atari wanted us to write a Basic for their new Atari 800 computer. That is another story.....

Every time I recompiled the code, I incremented a revision counter. The counter started at Rev 0.1. Whenever I got to (n).9, I would roll the counter over to (n+1).0 The first listing I gave Apple was Rev 2.8. They (I forget who) decided they could not call it DOS 2.8, so they changed it to DOS 3.0. Apple did the beta testing with this version (2.8 renamed 3.0). When Apple shipped the DOS for revenue, they incremented it to 3.1 to indicate that the code had changed from the beta version. As a final note, when I transferred the source code to Apple in October, 1978 the Rev number was up to 6.3.