1987

Paravant

In early ’87 my contract with NCR through Elias was finishing up and Elias didn’t have any definite work lined up, so I started looking around on my own and managed to get a short contract with a company called Level 5 who had what was then called an “Expert” system. In a few years these would become known as Artificial Intelligence or AI systems. The contract was to design and code an interface between their Expert system and the Common Data Dictionary system which ran on a VAX computer.  The contract was only for four to six weeks so while working on it I started contacting people to see about additional work.  

Soon after I had left the failing Harris Word Processing division, my good friend, fishing, and diving buddy Craig also left the Harris and started a new company, so I called him, and it was good timing for his company needed someone to do testing on a system they were developing.   I wasn’t too keen on doing testing, but Craig explained it wasn’t just testing, but also involved developing software that would automate the testing. Well, that did sound somewhat more interesting so Craig asked me to come in to talk to his VP of Engineering Rick, so I went in for an interview the next day.

Their company was a small startup company named Paravant and had less than twenty employes. While Craig was the President and recommended me for the job, he said Rick was the V.P. of Engineering and did all the hiring of technical staff, so I needed to talk to him.  Rick was a nice enough guy, we sat and talked as he reviewed my resume and asked me some questions.  He explained that their company did both hardware and software development and they had two contracts, a small business contract with the U.S. Army and the other with Miles Laboratories.  The contract with the Army was to develop a ruggedized hand-held PC about the size of a Webster’s desk dictionary that ran MS-DOS which was a novel concept at the time with the smallest “portable” computers being the size of a small suitcase, it would a few more years before “laptops” became common.   It sounded like a pretty cool project but that it was not the project they needed the testing. 

 Rick explained the contract he wanted me for was with Miles Laboratories, which in addition to making Alka-Seltzer, made drugs and medical products including blood glucose meters.  The contract Paravant had with Miles was to develop a device about the size of a paperback book that a blood glucose meter would connect which would then transmit the meter’s blood glucose data over a phone line to a doctor.  Rick explained the device was at the end of development and they were about to enter a testing phase of the project and they wanted me to head the testing effort.  I didn’t think much of doing testing work, but hey, testing work was better than no work at all so I accepted the contract.  It was the first job in over ten years that I would not be writing code. As it turned out the testing was a lot more complicated than I initially thought. 

Miles Laboratories was a leading manufacturer of blood glucose meters and glucose test strips which was a huge business considering, several times a day, people with diabetes prick their finger, put a drop of blood on a test strip, and put the strip into a meter to determine their blood glucose level. With of 30 million people in U.S. having diabetes, its big business.  Microcomputer chips were showing up everywhere and Miles Laboratories had developed a glucose meter containing one which would record the readings of the test strips over several weeks which the patient then could take to the doctor who could review the data. Paravant’s device had the benefit that the patient didn’t have to make a visit to the doctor.  The device had a microprocessor, some memory and a built in modem, all things I was familiar with.  Paravant had a hardware engineer and a software engineer that  I would be working with to do the testing.

The device was considered a Type B medical device and thus required significant testing for FDA approval.  Fortunately, Type B medical devices are the least stringent of the FDA classifications.  Even so hundreds of tests needed to be performed, tests for which my job was to define and document in detail.  Paravant had just moved into a new facility which had both office and warehouse space and was to share an office with the software developer for the device and start work the following week.

After being provided the product description, copies of the hardware specification, and printouts of the software I begin designing tests, lots of tests, hundreds of tests, page after page of tests, for each test describing what it was to test, what the inputs were and the expected outputs.  I had written a lot of code for work and had worked a lot of hardware as a hobby so when I thought I had every possible test  I would review the tests with the hardware and software engineers, and they would then suggest more tests.  What happens when the meter is unplugged, what happens when the phone is unplugged, what happens if the device memory fills up. In the following weeks the number of tests grew to a stack of paper over two inches thick.   

 

Graduation