I'm not the only one wondering what Sun is doing

Well, it looks like I'm not the only one who thinks that Sun is dreaming with the Java story. Have a look here

On my drive over to the Sun Tech Days presentation in Toronto, an idea kept rattling around in my head: Sun cares a lot about WORA (Write Once Run Anywhere). During the sessions that I attended, one of the things that Sun speakers stressed over and over again was the need for portability across different EJB containers. The question that I don't see addressed very often is: why does Sun care about portability? The answer to this question isn't immediately obvious. After all, Sun is a hardware company that doesn't appear to make money directly from its Java technology suite. Java, however, is the technology that brought Sun into the forefront of public / developer consciousness. So I believe the question really is one of how Sun can turn Java mind share into hardware profits. The answer to this question is that Sun wants Java applications to run on Sun hardware. In order to make that a reality, Java must be portable, otherwise there would be too much friction involved in migrating an application from one hardware / container platform to Sun's. By adopting this strategy, they have effectively tied the future of Sun to the idea that folks could develop Java code on their existing hardware, but that eventually, someday, those folks would reach into their checkbook and buy some Sun hardware. While this is not necessarily a bad strategy, I think this is a horrible strategy for Sun. Java portability is a double-edged sword that lets folks switch to Sun hardware, but it also makes it possible for folks to switch away from Sun hardware. The only way that this strategy can work effectively in the long term is if Sun can outperform the competition in the price / performance. I really don't see how this is possible, as Sun builds and sells proprietary hardware. Their hardware commands a considerable price premium over competitor's products today. In fact, no Sun hardware can be found in the top 10 results for TPC-C raw performance or price/performance. This premium is likely to continue to exist in the future because they lack the economies of scale that are available to the competition. Unless Sun can turn this around, it is likely that businesses will exploit Java's portability to switch away from Sun hardware.

I've said the same thing time and again - by making Java portable, Sun forgot what their core business is. They make virtually all of their money from Sparc servers. Thus, making it easily possible for people to migrate from eexpensive Sun hardware to cheap intel (Linux or Windows) hardware, Sun shoots itself in the foot. Business case studies will be written on this eventually - as an example of how not to invalidate your own market.