In Faster speeds but a slow change Martin Lynch of vnunet.com talks about the problems the multi-core market is facing due to the current shift towards mobile processing. While he is right about some things... I think he is over simplifying the situation.

True: Multi-core processors have largely arrived prior to the software that can most effectively take advantage of it.
False: The difference in timing for application arrival vs multi-core arrival spells bad news.
- Martin talks about this difference being an issue of the cart arriving before the horse. Bad analogy Martin. In this classic analogy, the cart is supposed to represent goods that are spoilable and are unable to be driven to market. In this case... the processors are the horse... with no goods to pull to market.

The true issue is one of development cycles and the state of the economy. With software development being in such shambles due to fragmentation of code base developers (Jim develops for C#, not C++... etc) we don't have a large enough pool of software developers. Well... a pool large enough to keep up with the pace of innovation of hardware. This is evident in virtually everything that exists on the market w/ the problems in driver vulnerabilties that go un-tested and now, the large lag of software behind the processor.

Martin does do well to illustrate the point that Intel and company are doing their part to re-educate the programming mass to go multi-cpu. Too little too late? Perhaps. Somehow, the separating curve between software development and hardware development needs to be closed. What is the problem? Why has it occurred? Simply put, it is an issue of view-points no what software is and how it works in for-profit corporate environments.

Hardware is developed w/ the goal of eventually maybe making a profit on it. The theory with hardware is that even if a given processor development line doesn't result in a salable chip, it does further the knowledge and experience of the chip engineers. Software isn't viewed the same way. Instead, a single product is looked on as needing to be "out the door". If it won't make it out the door... it is looked at as a liability and a failure instead of a natural cost of research and development as it in the hardware world.

For the IT industry to survive... this view MUST be changed... or we need to start breeding mega-coders for our species. Oh wait, to get the genetic understanding to do that... we need software. Oops.

Seriously, the open-source movement is growing and gaining ground because the participants take the hardware r&d model into the software world. Nobody realizes it... but that's what is done. If I'm a programmer in the open-source movement... the attempt to create something has more drive than the end result. Sure, it's satisfying to achieve a working distribution... but the devil is in the details... it's working out the details that is the driving/interesting part of it. If the open-source movement looked upon each hour spent tinkering w/ a new set of code as "wasted" we wouldn't have the class or product that we do there.

I think I hear the thumping on the horse w/o the hoof-beats... soo... I'll close for now. If you read this... be sure to drop me a comment to let me know what you think. It doesn't have to be an essay... just a one/two sentence opinion on my writing. Thanks. :)