The Stealth Turnip

I was working on two separate postings, "Rejection" and "Virtual Pheromones", but put both on hold until later this weekend. Tonight I wanted to write in celebration of an anniversay, if celebration is the correct word.

Twelve years ago today, Iraq invaded Kuwait. Following this act of aggression, the Pentagon started the operation known as Desert Storm, and we in the US tied yellow ribbons around our trees to honor the service people serving in the Middle East. We also bombed the hell out of Iraq, eventually forcing it to return behind its borders.

I, along with most of the people in this country, was very interested in the events in the Middle East, playing a radio at work to keep informed, reading the papers, watching the news at night. I was doubly interested because I had worked on Saudi Arabia's air defense system, Peace Shield, when I was employed at Boeing.

Peace Shield was a cooperative venture between the Saudi government and the US Airforce. Boeing's role was managing the development of the software for the system. It was the second largest computer system ever developed, with software contributed by multiple contractors, all using their own development techniques, naming conventions, and coding practices.

My job was to find a way to integrate the meta-data from all of these different companies, creating automated tools to pull the data from the code, using it to pupulate a common data dictionary.

(The dictionary was a requirement mandated by the Air Force, which was, I believe, a bit concerned about taking acceptance of several million lines of code with only minimal understanding of how data flowed through the application.)

The code was FORTRAN and the data dictionary was based in an Oracle database, all housed on a VAX/VMS system. I tweaked and reviewed code and tweaked some more, taking FORTRAN coding sheets home with me, working seven days a week, at least 12 hours a day. I even had the use of one of the earliest portable computers, a Compaq.

Portable is relative. The Compaq weighed about 20 pounds and looked like a huge toolbox, with amber and black screen. The operating system was DOS, and we had installed Lotus 123 on it to aid me in my effort. Unfortunately, the hard drive eventually crashed -- they did this a lot -- and we had to get it replaced.

Replacing the hard drive on the compuer wasn't a problem, Boeing had spare hard drives. The real problem was that the Lotus 123 discs were locked down to the initial installation and we couldn't use them on the new hard drive. Back in those days, software companies would alter the installation discs to prevent the discs being used more than once. We eventually had to go IBM to get new discs in order to install the software. .

Finally, I was finished with the application, everything was staged and ready to go. I started the program and sat back and waited. And waited. And waited. My application was running happily away, consuming the content of program file after program file, stuffing intermediate data into the database. A lot of data was getting stuffed into the database.

Gradually, and then more significantly, the system started to slow. My application was running at decreased speed, and I could hear engineers calling out to each other, trying to figure out what was going on. Finally, I received a frantic phone call from the DBA that whatever it was I was doing, I had to stop. Now.

My application had taken down the entire database system.

The program worked, but it worked too well. Just trying to process the intermediate data was more than the system could handle. Trying to use files and other storage systems and more efficient algorithms didn't work - there was too much data, too deeply embedded into too much code.

I had to admit defeat.

I took my findings and a carefully prepared report of the problem into management and reviewed it with them. These were good people who listened and reviewed and concurred with me -- it was too late to try and populate the data dictionary from the existing code base. The base was too big, and the coding practices of each of the contractors differed too drastically.

It was my first programming failure.

However, not all was doom and gloom. Peace Shield was a fun project, with an interesting and eclectic group of engineers. To reduce the stress of the project, the HR group would pull together fun little events that served to build team unity, and to break up the tension. One such event was a paper airplane contest, a once in a lifetime event when you consider the minds we had on the project. We ended up with creations that ranged from the simple but highly accurate 'traditional' paper planes, to a huge origami swan, and variations in between.

And then there was my entry: The Stealth Turnip.

Now, you'd have to see The Turnip to appreciate its beauty. Several strips of paper were pulled up around a central core, creating a turnip like shape, with venting to allow even flow of air. Streamers at the end were folded into a helicopter like shape, and the other end was gently flattened to give it a point. The inventors of the F16 couldn't have been more proud. By my calculations, the Turnip should act as a powerless helicoptor, floating gracefully down to earth, immune to the puffs and breezes of air in a building that size.

The day of the contest dawned and we gathered in the central atrium, contestants on the third floor, spectators on the ground. One creation after another was let lose to float, or crash, to the ground. The swan was particularly graceless in its head over butt fall to the floor.

When my time came I stepped up, looked out over the assembled watchers, pulled back the hand holding The Turnip and let her go....

...whereby she promptly, quickly, and with amazing speed dropped like a rock straight down, earning the dubious distinction of being the plane to land furthest from the target.

I eventually gave The Stealth Turnip to my manager who was rather fond of it, as a remembrance of our working together. This was after Boeing closed down the operation having failed to meet the objectives the government provided.

Boeing lost 1.2 billion dollars on Peace Shield, and I met a project I couldn't complete. I also designed a plane that couldn't fly, and Lotus created software that couldn't be reinstalled, on a computer that had a habit of throwing hard drives.

And today's anniversay of the invasion of Kuwait 12 years ago reminded me of all of this. Mistakes. A lot of mistakes. And now I hear that we want to go back. Go bomb Iraq some more.

Mr. President, if you want to give your Dad a present for Christmas, may I suggest a tie instead?

Posted by Bb at August 02, 2002 08:55 PM