Code Trials

At my last company, Garden.com, we had a number of rounds of venture capital before going public in 1999. Our primary Venture Capitalists were Austin Ventures, led by John Thornton. Over the years I've got to know John pretty well and appreciate his knowledge and sound business sense. He has always been generous with his time and advice to me.

However, I remember feeling disappointed when we first started in 1995 because he was not always as encouraging as I expected about our business. This can often be the case because a VC's primary goal is to make investments and to get a good return. As an entrepreneur you are typically starting a company out of passion and a burning desire to create something. Of course you'd like to make money, but frequently this is not your primary goal.

As a result, when I started Digital Cheetah in 2001 the last thing I wanted to do was get any external financing. If you can afford to go alone, that is often the best strategy because you can create the company precisely the way you want. However, if you need the cash then remember VC's are for funding first and just be grateful that they are willing to invest in you at all.

One of the things I learned whilst working on my Ph.D. was a sense of rigour. My advisor: Professor John Tucker fought long and hard with me to curb my sloppy ways and instill a deep seated sense of precision and rigour in my thesis and code. This was not an easy task but I believe it was the single most important lesson I learned during my Ph.D.

Nothing frustrates me more than when I see sloppy code or practices. If you have a set of coding standards for testing, documentation, naming conventions, and even layout - use them. If you don't, then maybe you should get some.

One of the reasons I am drawn to Python for my next major project is that the language has a very clean semantics and when you write using Python it just flows very naturally. The precise indentation requirements make even the layout of code always look consistent. Also, with facilities like doctest it becomes possible to combine the code, with the documentation and testing in a very clean and precise bundle. One can certainly not accuse Perl or even "C" of being rigourous.

No one would ever say John Tucker was a great coder, but he certainly helped me write better code and design stronger systems that have lasted for years. I think a big reason for this is the sense of rigour he instilled in my over twenty years ago. Thanks John!

I recently read the thought provoking book: Outliers by Malcom Gladwell. I was particularly pleased to see Bill Joy and The Beatles featured, amongst many others. The book attempts to explain why some people succeed far more than others. It was fascinating to read about the early lives of Bill, John, Paul, George, and Ringo and how they grew into such huge successes.

The basic premise of the book is that when and where we grow-up, and how many hours you work, has a much bigger impact on our chance of success than genetics and intellect. It traces the lives of highly successful people applying this premise. He is so convincing that I definitely changed my view of how I achieved my own success.

I used to believe I was just born brilliant, but thanks to Mr. Gladwell's book now I'm not so sure ...

Over Twenty years ago I shared offices with one of the best coders I've ever worked with: Chuck Chan. We were both working on our Ph.D.'s
and we spent many hours discussing programming, girls and beer. I got on very well with Chuck and was quite certain we would both be writing code for the rest of our lives.

When I was at IBM I worked with another great coder - Larry Raper. Larry and I worked closely together on one of IBM's great lost technologies the System Object Model (SOM), and it was through the work on SOM that I received the majority of my patents. Larry and I had many coding discussions together and I have fond memories of those times. Larry was always going to be a coder, I felt sure of it.

Chuck started working for Microsoft in 1989 and today he is a General Manager. Larry is now retired but he spent the last five years of his career as a manager. Both of these great coders took the manager path that I see all too often in our industry. There is only one reason I know that great coders move into "management" - money.

One of the main reasons I started Digital Cheetah was because I didn't want to end up as a manager. Nothing against managers - a good manager is almost as hard to find as a good coder. What is sad is that our industry almost always takes the best and most experienced coders and makes them managers. I don't understand why this is because rarely do the best coders make the best managers.

I think companies like Google understand the importance of experience in their coders and only wish that more companies would compensate them fairly. An experienced coder is easily worth ten or more juniors, sometimes a lot more.

Over the last twenty years I've had thirteen patents published. Most of these were applied for when I was at IBM in the early '90s. IBM had good financial incentives to publish inventions and patents. It also turns out that IBM Austin produces more patents for IBM than anywhere else in the corporation, so its not surprising I wrote so many.

Since starting Digital Cheetah in 2001 I have developed some very creative solutions that are certainly worthy of patent submission. However, I have not applied for one patent. There are two reasons for this:

• The cost of processing a Patent.
• The restriction they can impose to freely use the ideas.

In the early days of Digital Cheetah the cost was the biggest reason for not pursuing a patent. However, these days I am less inclined to apply for a patent because although they offer some protection against their use, they also impede my own ability to use the same ideas in the future. For most small businesses I believe "patent protection" is an illusion and all that you really do is limit your own use of the ideas.

Most of the time the only possible reason for applying for a patent is if you plan on selling the company. Then having a patent portfolio can be attractive.