Now that Sun's chosen the GPL license for Java, there are several opportunities on the table for other folks. Now, these chances were open if you wanted to get into a commercial relationship with Sun in the past, but the dynamics of the relationship can now be different. The obvious first one is being one of the first non-Sun commiters. Here's another one that some chatter in #redmonk this morning made me think of:
"The Performance People"
Chip companies like Intel and AMD can more easily send their teams of people to optimize Java for their chips. Now, they've been doing this already under the old terms, but there's a broader chance than just optimizing: these companies could dedicate people to simply working on the overall performance for Java and Java-based code.
Aside from being a good citizen, this could be part of a wider branding story along the lines of "[Intel|AMD] is a performance company. From chips to software, we make things run faster and better." Of course, these "performance people" would work on more than the JDK and JRE, they'd volunteer to work on all sorts of open source projects.
My first choice would be Jakarta Commons; not because commons runs slow or anything, but because it's everywhere so any "performance fixes" on the table would pay off big time. Eclipse is out there as well. The goal is to target the code that's widely used.
The pool of projects needn't be limited to Java either: PHP, ruby, rails, python, and friends would be great as well. And there's Mono of course. Sure, you could throw in commercial platforms, but that's a different line of thought than this open source noodling.
What They'd Do
For any given project, the performance team would get the code and scour over it for optimizations. If they found anything, they'd offer some changes to the project teams. Once the performance team was satisfied, they'd give the project a seal like "Performance Proven by Company XYZ." If you saw the seal, you'd know the performance people had looked it over and given the project a thumbs up. (No warranty expressed or implied, I'm sure.)
If there was something particular to their proprietary hardware, sure, they might try to slip that in. But, they'd soon learn that focusing on overall performance would get them more positive brand-reputation than using a narrow, proprietary focus. Besides, if they sent over crap-patchs, the project teams would/should just reject them.
As an example: folks are always saying how multi-core chips add in a layer of complexity/opputunities when it comes to multi-threaded, concurrent programming. I'd suspect that'd be one of the core-compentencies of this performance team. I'm sure there's all sorts of graphics and networking optimizations that companies could find, fix, and/or contribute as well.
More Benefit from Existing Efforts
Now, I know that there are already people and companies doing things along these lines. There's plenty of people all over who work on performance. In fact, this idea could apply to non-commercial teams as well. The opportunity for chip companies is to profit from the positive PR of such efforts.
Why does that matter? Because in chips brand is a key differentiator. Sure, there's always an innovation race, but I'm always suspicious of that being a stable differentiator, it seems more spikey: everyone is 64 bit now and we've got multi-cores out the wazoo. Put another way, you need more than just the technology itself to differentiate. Remember "Intel Inside"? And every company is certainly getting brand-juice out of "being green."
New Revenue Sources
The brand idea behind the performance team is to make people think of the company as focused on performance, the hardware they happen to sell just happens to be part of that...a big, big part. But, if you wanted something performant other than hardware, you might think of going to that company asking to pay for something new. And there's a whole new revenue source, if only indirectly by driving chip sales.
Previous Topic
