Really. Practically everything is made of pieces, but that's missing the point. Since the operating system of a computer is not just a piece, but defines its function, a better question would be: How many car manufacturers license their car designs to other manufacturers?
In the technology business, we've found that the component model, as it is has been defined here, really doesn't work. It worked in Microsoft's case because of an historical fluke, a set of events which are not going to be replicated. Even Microsoft can't seem to manage it, and they've had more experience than anyone. That's the take-away lesson here.
It may have been a set of events which are not going to be replicated (as the PC market seems to have matured beyond the point of consumers wanting to all use the same operating system), but that doesn't mean the component model doesn't work in the technology business as a whole. You can argue that it's just a question of definition, but I think overall, the "component model" has its uses in the technology business.
The basic concept of the component model is, from the "Microsoft" point of view, to sell Product X to be bundled with another product as a solution, that consumers/businesses will buy knowing that Product X is part of that solution, basing their purchase passively or actively on that fact or assumption.
An example of this component model is the constant marketing mention of the use of Carl-Zeiss lens in cell-phones or cameras, as an indication of high quality. Both the Carl-Zeiss brand and the cell-phone/camera brand are pushed to the consumer as part of the solution. This business model works for Carl-Zeiss (they don't have to make the cameras), works for the cell-phone/camera manufacturers (they don't need to make lens), and works for the consumer (they can get a Sony cell-phone with a Carl-Zeiss lens), just as it works for Microsoft - the business can get a Dell computer with the Windows Server 2008 software they need for compatibility with previous purchases.
There is, of course, a crucial difference between caring about high quality and needing to purchase a product for compatibility with previous purchases. But this is not a failure of the component model. Apple's foray into the component model failed, as far as I know, mainly because their profitable operating system + hardware sales were cannibalized by less profitable operating system license sales, because they hardly brought in any new customers.
However, what if Apple was able to bring in new customers, and get 1 billion sales of their operating system? That would have been a success. They would have the volume required to keep developing their operating system. Yes, the end-to-end model lets them create a great user experience more easily, but they essentially need to do that because they need to make more money per system sale, with a far lower volume of sales, especially compared to Microsoft.
If OS X ran on a non-Apple computer that provided the same functionality, except feature X, to a specific consumer as an Apple computer would, and cost $500 less, and feature X was only worth $100 to that specific consumer, then the availability of OS X on that non-Apple computer would be beneficial to that consumer, but not beneficial to Apple.
So, Apple's attempt at using a component model failed not because a component model just doesn't work, but because they didn't have the business to support that model.
I believe the component model can work again in the technology business, and that it is a mistake to write it off, missing possible opportunities in doing so. For example, look at the current battle between the Xbox 360 and the PS3, and how much money Microsoft and Sony are losing in that battle, trying to convince consumers to buy their console rather than the other (on top of convincing consumers to buy a next-gen console in the first place).
To do this, the two companies are doing lots of advertising and battling for exclusives. They are courting 3rd party developers not just for exclusives but also for making a game
not exclusive. Microsoft spent about $50 million to get GTA IV on the 360 as well as the PS3, and who knows how much for Final Fantasy XIII.
The current, actual trend is that more and more games are becoming multiplatform, because 3rd party developers want access to more of the market. This also means that they have to develop for two very different platforms/architectures. All this results in lots of money lost, thrown around for exclusives and multiplatform access and developing and advertising... Sony and Microsoft are simply losing a lot of money that they need to earn back sometime, which points to at least lengthening how long each video game generation is.
Because of this, powerful voices in the industry are predicting a possible convergence in video game platforms, saying that maybe, it's not unlike the beginning of the PC industry where software exclusivity and incompatibility took a toll on profits and overall sales. Example:
Code:
http://www.bdgamers.net/2008/01/13/god-of-war-creator-wants-unified-platform.html
Sounds like a possible new component model to me, albeit not exactly the same as Microsoft's history's, of course. But then again, that would just be a question of definition. Presumably, the platform could be a unified operating system designed for game consoles, plus a variety of hardware configs (like one base system that costs $400, for most consumers, and an "elite" system that has a better cpu, more ram, a better graphics card, 802.11n that costs $800, for enthusiasts, that would play the same games with more graphics detail, but still wouldn't be open to upgrades like a PC is). That would just be like a 30" 720p tv for $300 vs a 50" 1080p tv for $2000 - different quality, but pretty much the same content, depending on who you ask.
Or it could be a single "operating system" + hardware solution (making it an end-to-end model instead).
But I'd say that the possibility is there.