Dojo (HowTo)







  Easter Eggs




  Martial Arts

Carbon and Intel
What is likely?

By:David K. Every
©Copyright 1999

Remember, this is an OPINION piece, on what I think is likely to happen. Treat is as WAG (Wild-Assed-Guessing) -- but so far I've been pretty close on many of my guesses. It is pretty easy -- just learn all the potential paths, then guess towards the ones that are more reasonable. I also cover a few paths, and mention that Apple is likely to follow one of them, and they usually do.

Q: Is Apple going to make stuff for the Intel market?

A: Yes.

There is no doubt that Apple wants to play better with the PC market, in order to expand their business. But not at the sacrifice of everything they have built on the MacOS. Apple does not want to telegraph their punches, and they don't want to scare their customers, so they will be very quiet until release, and even then, probably try to do in a "nichey" way, or one step at a time to not scare off the Mac users. But this does not mean that Apple HAS to offer MacOS X directly on PC platform, or even Rhapsody for Intel. There are many choices.

Q: Why is Apple going to make stuff for the Intel market?

A: Because there is money there.

The PPC is better for many things -- but there are many PC's. Apple wants to bring their ideas to the rest of us -- which includes "them" as well (and make money). But more than that, Apple wants to make it easy for Mac developers to sell to Windows users, so as to encourage as many people as possible to make Mac Apps. Again, that may not mean offering a whole OS.

Q: What are Apple's choices for delivering Intel Products?

Whew, quite a few. Follow this:

  1. Deliver MacOS X on the full spectrum of PC's (Intel).

    This would be the entire OS -- or a recreation of the StarTrek project for the Mac/Carbon side (and of course YellowBox for the Rhapsody side). This would be a huge expense. Not because of the OS itself -- Apple has YellowBox working, and has most of MacOS (and could have Carbon working in an instant). But getting all the Drivers written, and supporting all the legacy hardware, would be hell. For this reason I don't expect it to happen -- even if Apple could get some one else (say Intel) to foot much of the bill.

  2. Deliver MacOS X on the subset of PC's (Intel).

    This would be the entire OS -- but only running on some machines. The PC is an ugly mishmash of crappy add-on and hacks, bolted on to a crappy design in the first place. It was an old architecture in 1981 (it was just a bad copy of S-100's that go back another 8 or 9 years). But Microsoft, Intel and many others are trying to push for "new" PC architectures, that are a subset of the old. They want to lose the ISA bus (and cards), simplify the I/O, simplify plug & play, to re-architect a clean design, to design in things (audio, video, and so on), and to get rid of IRQ problems and other expansion limitations. In other words, their goal is to make something more like the Mac was in '87 (but with newer ports).

    If any of those initiatives succeed, then Apple may try to support that subset. There are timing issues, and partners. But the PC makers, and Intel, all want to get out from under the thumb of Microsoft, and would be more than willing to work with Apple (if they thought it would sell more boxes). Still, don't under estimate the effort required to make this successful -- it would take timing, luck, and a lot of resources.

    If Apple were to do such a thing, it would need to time it just right, and sort of evolve towards the future. MacOS X is still a year out. So they need it to be with an initiative that starts a year or two from now. You know, like say start with the release of Merced Processors, and a new server architecture, and make Mac OS-X for that small subset, and grow with that market (probably licensing to others that side of the business, since it wouldn't pirate Mac sales much, if at all). As that market mainstreams, Apple could as do so as well, on that side. Also the high end niche would be perfect because of the Macs Apps that would fit that market (publishing, 3D, etc.). There is a strong rumor that Intel and Apple are already working on these projects together (supposedly at Intel) -- and this would be the kind of products I would expect them to work on. But any niche entry would work. (Start small and expand). It could as easily be an Intel NC or set-top box running MacOS-Lite, or the same type of Apps as iMac, and so on. Apple just can't swallow the entire PC market in one gulp.

  3. Deliver an MacOS X (Carbon and YellowBox) to ride on top Windows

    Why does Apple have to take all the support nightmares of PC's, if they can make Microsoft deal with that? Apple can just release their OS as a library that rides on top of Windows lowest level API's. Oh wait, Apple is already doing this with YellowBox and Web Objects, and with QuickTime (QTML), and will likely to continue to do so -- so the only question then is how far will they expand this path. The answer seems pretty obvious that they will continue to do so.

    Apple has to charge people for YellowBox because they don't own the imaging engine (Display PostScript). Carbon has a hybrid imaging engine (eQD -- enhanced QuickDraw) that IS cross platform (and is probably really just Bionic reincarnated QuickDrawGX, with a different name). So Apple will want to put Carbons new imaging engine underneath YellowBox anyway, they already have QuickTime ported to Intel -- which includes most of the Mac API's in the QTML layer -- so what is left? Not much except the new API's -- and Apple is writing those API's to be cross platform anyway. So the expenses are QA, and the marketing effort. Not too much.

    Remember, Microsoft can't change things at the low-level without breaking lots of Apps, and getting into even MORE trouble with the DOJ, so I suspect Microsoft would not be able to stop it if they wanted. The ironic thing is that Apple gets to stick Microsoft with paying for most of the OS support -- and Apple just gets to run all Carbon Apps, inside a Boxless GreenBox (remember, BlueBox is for MacOS, YellowBox is for OpenStep Apps -- and GreenBox is a little of both and more). So even if Microsoft is against it, it could succeed.


So I doubt that #1 will happen. I suspect very strongly that Apple will do #3, whether Microsoft wants it or not. I suspect that something is happening along the lines of #2. But #2 and #3 are not mutually exclusive -- Apple can actually use each effort as insurance against the other failing, and play one side off against the other. But wait, there is more.

Since Carbon and YellowBox are going to likely be merged into one mega-API in the future (with YellowBox being the high-level API, on top of Carbon which is the low-level API), I sometimes refer both "Carbon and YellowBox" as just "Carbon" -- as I believe they will be the same thing. A reader mentioned that they should be called "Diamond" (since Carbon becomes a diamond with heat, time and pressure).

What if Intel is helping Apple, and why?

I bet it is happening. Intel wants EPIC support, and they certainly would love for MacOS X to run to run on their platform (and mean more processor sales). They certainly have to fear migration to MacOS (and PPC) a little -- but it hasn't hurt them in the past, so they are probably not too worried about it -- at least not as worried as they are about other things.

Remember, all the processor clone makers are starting to hurt Intel. With Intel falling on their face with the PentiumII for portables, and maybe being forced by the DOJ to license (or give away) their slot-design, they are not in a good place, and they know it. The best thing they can do to guarantee success is get EPIC out, and make it FAST, and get as much support for it as possible. They know this. They want OS's, and support, so that they can build momentum. If they succeed, they know they have a multiyear lead on all the Processor-Clone makers -- and they NEED that lead. They need that far more than they need to worry about a little market erosion to PPC's. Long term (5 - 10 years), I think they will have a different view of the situation -- but Intel is scared shitless about the next 2-5 years, and for good reason.

Of course there are plenty of others who wouldn't mind working with Apple (I would think). Compaq and Dell probably wouldn't mind getting out from under the "thumb" of Microsoft at all -- and they all want to differentiate themselves from the Competition. So while Carbon on Windows is a great plan, Mac OS X on Intel is not exactly a bad plan either.

What if Microsoft was not against Carbon on Windows?

Remember, that Microsoft is a multifaceted company. It is a hydra -- like IBM, and Apple could possibly play one head off against the other -- or the big head may favor one more than the other. Microsoft already wants to get out from under the DOJ's thumb -- and playing nice, might be just the tool to do it. They also don't mind making more money off the Mac market.

Imagine this -- Apple and Microsoft deal is bigger than people realize. What if that convergence idea I expressed a while back (What if?) is not so far fetched? It doesn't have to happen as MacOS on Windows kernel (on the PPC) -- there are many ways to get to the same destination. So Microsoft agrees to put Carbon API's (libraries) in the package with Windows -- when people install Windows99 (or 00) they are really installing the Carbon Libraries as well. In exchange, Apple would have to do the same -- Win32 API's, or part of them, that would ride on top of MacOS Kernel. Apple could make a boxless RedBox to run all Windows Apps in the Mac UI. Either side would have push button compiles to run on the others platform. Going either way there COULD be emulators as well -- remember, that if the Windows API's were compiled native for the Mac, and many of the cards drivers are native (and some would be), then the emulator would only have to be running for 10% of the code -- an emulator would perform pretty well. Think about that.

There is a small chance that Apple could make a boxless RedBox without Microsoft -- and it would definitely be in Apple's best interests -- but it would be a large effort, and would likely run into legal entanglements. It would be far better if Apple could make a deal.

For Microsoft, they would gain getting out from under the DOJ's thumb (a little). They would gain yanking some control (and value) away from Java -- which they really fear. So the question for them is really who is the bigger threat to Microsoft -- Sun and its allies, or Apple and its allies. I'm betting they fear loss of control to Java far more than playing nice with Apple. Remember, that Microsoft also gets to move Apps back and forth, and make more money cross-selling into other markets. So there is some things to be gained for Microsoft -- the question is do those things outweigh the costs. With Microsoft that latter question would be tough.

What about Carbon elsewhere?

Let's continue to do the logically reasoning. Hmmm... Carbon (MacOS X) runs on Unix. Carbon will get a lot of Apps, just because of the Mac ports -- not to mention OpenStep/YellowBoxes rapid Application Development advantages. Who else has OS's that could benefit from these Apps and these API's? Or more importantly, who else does NOT!?!?!

IBM could certainly use Carbon Apps on their workstations. Sun needs the same stuff as well. Not to mention anyone else. Apple could be the OS for the rest of THEM as well. Again, it would still be in Apple's interests to just port the API's on top of the others OS's -- and let those companies support the hardware level. But it is in IBM's and Sun's interests even more than it is in Apple's. If either of them have a brain, they would make these deals with Apple. After all, they are going to with QuickTime and Java -- so it isn't much of a stretch to do it with Carbon as well.

Don't get me wrong -- I think many corporate leaders are not forward thinkers. They fear change. So it is not a given. With all large-corporations there is a lot of little kingdoms that will fight each other -- and something like this could put one group against the other, but it could happen, and would be in the Corporate interest (over all).

But Apple doesn't want to lose Mac sales?

Apple canceled cloning because 95% of the clones being sold were pirating Apple's sales. Carbon on Windows, or MacOS X on Intel, would be the exact opposite. Every OS sold (or licensed) would be revenue -- or if given away it would at least be free advertising for Apple. Either way it would be a way to encourage more Mac development (which is good for Apple's hardware). There would be surprisingly little extra investment -- I'd guess around 5-10% more development costs, and maybe only 10-20% more testing costs, and for Carbon on Intel the support costs would be near nil.

For those that doubt the testing figure (and think it would be higher), remember, that testing comes in surges. If Apple has to employ those people, then you can do some phased testing -- release on Mac first and Windows later. As soon as one test is winding down, you can roll those poor sods onto the next project and never give them a break. Otherwise there can be some real down time and wasted resources. Costs do go up, but not as much as one might initially think.


I think Apple will choose one path or the other, either MacOS X on Intel, or more likely Carbon on Windows (and everything they can). Not because the resources are too high otherwise, they are not, but because of the marketing confusion it would cause to try to do it all at once. (Users would get confused and think it was crazy). If Apple was going to do both, I think they would stagger it, so one concept came out first (probably the Carbon on Windows), so they could get people used to that idea -- and then phase in the other one (MacOS on Intel) once people are comfortable with the first concept.

My message is not changing much. You can read what I was saying six months or a year ago, and the general concepts are staying the same -- it is just which path to take to get there that is evolving a little. Apple is smart enough to see these goals too -- as are the others. So it isn't like am stating things that people haven't thought of -- so the only question is can Apple pull them off? And if so, pull off which things?

Optimally, Apple would pull ALL these things off. But any one of them could be a big win. Two or more and we will see a whole different industry in 5 years. Apple bought NeXT (and OpenStep) because it just made sense. Well this stuff makes sense too. Of course Apple has sharp thinkers, as do the other companies. There are probably other variants that I haven't thought of as well, and the details can (and will) change -- so it is really left now up to chance, ego's and fear. In this twisted Machiavellian industry (with every side trying to outmaneuver the others, and put a knife in their backs, like some Royal Court) the question is still kinda out there -- there are a lot of twisted plots and subplots, and some people do stupid things for reasons that I haven't thought of. So just because it makes sense, doesn't mean it WILL happen either. But it is sure fun to speculate at what could happen -- and dream about what it would mean to me, and for Apple.

Special thanks to Eric Yang at the Mac Evolution web site, for being a sounding board for this article. If you haven't been to his site, he has some interesting PowerPC information as well as Financial and investor information. In fact, his PowerPC stuff has been somewhat an inspiration for me to finally expand, update, and collect my Processor information into a more complete form.

Created: 06/01/98
Updated: 11/09/02

Top of page

Top of Section