Sometime in December, 1999 — as my days working on Jikes were coming to an end — I wrote a short memo with three predictions about how open-source would develop over the next ten years.
It’s now almost eight years later. Here are the predictions and my reasoning.
1. Red Hat will invest in creating the first true open-source Java, and so become the dominant supplier of Java in the free and open-software communities, giving it a major strategic advantage.
Those of us who were around back in the day when lots of folks cared about Java recall that in late 1999 Sun made lots of noises about standardizing Java via ECMA. I wasn’t directly involved in this, but I did know a few folks who were.
This effort fell apart in early December, when it became clear Sun was not going to give up control, a step that would be needed to make Java a real open standard.
I then felt that it was unlikely Sun would “get it” when it came to open-source, for at least a period of many years.
I predicted this would leave an opportunity for someone else to start a meaningful open-source effort, and as Red Hat then had substantial resources left from its IPO, I predicted they would do it.
2. I predicted that over the next decade open source would become more important outside the United States than within it.
Look at the per capita GDP numbers; for example, Wikipedia’s List of countries by GDP (PPP) per capita. The first four countries are Luxembourg, Ireland, Norway, and the United State. Luxembourg is a fluke since it is a small country with liberal tax laws that have drawn many financial institutions. The next three each have PPP of just over 40,000.
As I recall it, that number was around 30,000 back then, and the cost of buying one copy each of Microsoft Windows and Microsoft Office came to a few days worth, certainly less than a week’s.
It you look at the end of the list you will see that the Windows cost is not just a few days of work but can run to weeks or months.
This is just a roundabout way of restating the obvious, that what seems an acceptable cost in the United States is way beyond the means of people in the developing world.
I felt that entrepreneurs who wanted to build software companies in underdeveloped countries had only two choices. First, they could pirate Microsoft software, but that would fail for two reasons: it was unethical, and even if that didn’t stop them, then Redmond would stop them as soon as the company showed up on their radar.
That left open-source as the most promising opportunity. I also felt that the most successful companies would not be those that just replicated software written elsewhere to their own country but those that created new software uniquely built to reflect some aspects of the country’s culture. Hence there would be a need for developers who knew that culture, and the easiest way to build a team would be to get some smart programmers and teach them about open-source, so they could then shape it to their needs.
3. I predicted that, on a global basis, IBM would do better than Red Hat.
IBM had been operating as a global company for decades, and only had to learn how to do open source.
Red Hat had been in existence for only a few years. Red Hat had no experience operating as a global company, and so faced a much more difficult challenge.
I would score myself as follows:
1. Wrong. Red Hat decided to spend those IPO dollars on Cygnus. (How many people even know that Red Hat did so, or what Cygnus did before Red Hat bought it?)
2. Right, though we’ll have a better sense in a year or two.
You can make your own call.