Forgive me for being a Scott McNealy fan, but I really can't help it. Scott and the crew at Sun have done a great job over the years producing what I consider to be really good products. Scott has also provided much-needed entertainment in the form of some very quotable quotes (who was there for his presentation that started with a single cloth-covered box that could run all of the OSes that Microsoft sold? When he uncovered the box it was an overhead projector - classic McNealy). And those that know me or have heard me speak know that I am an open source fan, nay, an open source bigot. Hence my quandary. Solaris for x86 is (and has been for how long?) open source. Yet, I can't help but feel that the folks at Sun are missing the boat.

Yeah, so we can run Sx86 on desktop hardware. We can even buy that hardware from Sun. Suppose, for just a moment, that we could run Sx86 not just on desktop hardware, but on big-iron hardware. Suppose, for just a moment, that this hardware was built by Sun. Can you imagine what kinds of big-iron hardware Sun could build? Imagine x86-based blade servers, intelligent clusters of servers, massively parallel multi-processor servers - all built and optimized by Sun for Solaris and supported by Sun's worldwide support infrastructure. Can you imagine what a killer OS this would be with Sun focusing all of their OS efforts on x86?

As an interesting by-product to this, the world gets an open source OS that is truly world class. An OS that is supported by the open source community and pushed forward by a "real" computing company (that phrase is for the IT managers, not a slap at Red Hat). It would be an OS that could make a formidable run at not just the high-end server market but at the desktop, too. (Note to self: be careful when speaking on behalf of the many developers out there who develop for the open source community. There are probably many that would be offended at the thought of providing code to a company that owns an OS even if it is open sourced.)

What about SPARC hardware? Here I must speak some (Sun) heresy. I would like to help Sun take itself out of the SPARC hardware market. That's right. I'm advocating leaving SPARC behind. Take that technological muscle and put it into building big-iron x86 servers. All of Sun's "wood" behind a single "arrowhead" (sorry, Scott). Use this as an opportunity to put some focus back into the product line - Solaris for x86 from the desktop to the largest and most complex high-end servers.

For those with a big investment in SPARC servers, I am not suggesting, nor am I foolish enough to believe, that everyone needs to perform a bunch of forklift upgrades. These boxes will run for a long time. During that time, there would be ample opportunity and justification to install new big-iron x86 hardware from Sun. We did the same thing in the transition from the Motorola-based servers, remember?

This is a(nother) chance for Sun to place their bet on today's equivalent of the VHS technology, rather than the BETA that they are currently using. Everyone knew then which technology was better, and everyone knows now which technology won the market over. Sun has tried to win with proprietary platforms (SPARC and SBus come to mind), and the results are sadly predictable. Why not win with a chip-set on which Sun can develop high-end server hardware products - products that can make our open source world a little better?

I must be dreaming. Come on Sun, join me on this boat. While you're here, why not drive?