Why is this important? Well, every one of you running a computer may be affected by the outcome of this 'battle'.
The article looks at the reasons behind this, with what I feel is mostly sound reasoning. It's quite insightful in places too. There are a couple of areas that fall down, namely the reasoning and maths behind the date for the next hardware transition (presumably 64-bit to 128-bit) - they suggest a date roughly 50 years from now, whereas given previous trends it may well be sooner than that. Presuming (possibly irresponsibly) the article is otherwise mainly accurate, here is my opinion piece based on the article:
The serious contenders for this new 64-bit OS are: Windows Vista, Mac OS-X and Linux.
Reducing this article tremendously, the following principles seem to shine out:
- An OS is cemented in the mainstream when it is adopted by.. the mainstream (i.e., greater than 50% of the market). This typically happens about three years behind the adoption of leading-edge hardware (64-bit processors were available in 2005, which gives us this 2008 date). When the mainstream RAM requirements exceed the limit of the existing mainstream architecture (32-bit limits us at 4GB RAM), there is a chance for a new OS that works best on the new architecture (in this case 64-bit) to rise. There are a number of reasons Windows may not be a 'default' choice any more, due to technical and financial reasons. See the article for full details.
- An OS MUST perform the following functions in order to be accepted by the mainstream:
- Drivers for all major existing hardware.
- 32-bit legacy platform emulation.
- Surviving the killer app.
- Enabling preinstalls.
- Support for all major multimedia formats.
In the article it describes the advantages and disadvantages of each of the serious contenders. One disadvantage of OS-X is the fact that it is only written to run on Mac-specific hardware. The article suggests that changing their business strategy at this point in time can only hurt Apple, so they won't choose to. This is one point I disagree with. If Apple decided to move to selling OS-X as a separate entity that will run on commodity x64 hardware (and x86, although this 'battle' is for x64, so x86 is less important), their revenue stream suddenly becomes big enough to support the commodity hardware (and as OS-X is based on BSD, which has open-source support for hardware anyway, this shouldn't prove too much of a struggle), and they can then achieve the magic 50% market share to ensure their success. Microsoft have played dirty tricks in the pre-install category before now and will no doubt complicate things if Apple decide to sell their OS to run on commodity PCs. But the question for Apple is - do they even want to do this? Steve Jobs is a clever chap, and is keeping his cards close to his chest. But they have been slowly commoditising the proprietary hardware that their OS runs on. Is that just for cost reasons, one has to ask, or is there an overreaching plan here?
Anyway, ignoring Microsoft and Apple for a moment, we come back to the question of Linux. Multimedia is clearly the crux of this argument, as the others are all mostly surmountable. The trouble with multimedia is that the codecs required to access this content are protected fiercely by the codec creators. Microsoft have paid royalties to get these codecs bundled with Windows. So have Apple. Both companies have developed their own codecs as well to further protect their positions.
So what about Linux? Well, this is the problem they face. Because of the fierce protection of the codecs by their owners, It is my opinion that unless a commercial company pays for the right to install them with the OS, multimedia support within Linux will remain as poorly supported as it is today, which will cost Linux the 64-bit OS title in 2008. But wait, a commercial company is already trying to do this - Lindows/Linspire is the company. But it has not met with great success, due to prevailing attitudes within the Linux community, namely if it's free why pay extra for it? Apple would have the same problem except for the cunning strategy of packaging the OS well and making it very simple to use. Now we begin to see why Apple is keen not to break this model.
But why do we want Linux to win anyway? Well, because we would break out of the cycle of paying through the nose for something proprietary that stifles freedom to innovate. What is the only way this is going to happen? If a commercial company takes Linux, packages it up with multimedia capability and therefore releases it as a proprietary product.
Hm. That doesn't sound any better than any of the existing alternatives, which is why no one will go for it, or indeed why no one has been going for it (Lindows/Linspire).
So I believe that Linux appears to be doomed to remain an OS run mainly by geeks and servers, unless multimedia can be built in for free and the whole thing made simple. I don't see this happening very easily, as currently the legal/financial climate prohibits this.
One point that is made in the article is that a killer app might cement Linux's success, which would then force the multimedia codec owners to allow integration of their software into Linux. The article suggests the criteria for a killer app has to be:
- Not deliverable through a browser (otherwise all contenders will support it about equivalently)
- Requires 64 bits of address space (otherwise it could be deployed on 32-bit hardware now, and would make no difference to the transition)
- Can't be ported off its home platform, whether for technical or legal reasons (otherwise there would be no reason for people to support Linux above something else)