There are multiple ways to look at Mac OS X and take it apart. Each way makes its own contribution to your understanding of the OS. In this long and interesting sample chapter of the “Mac OS X Disaster Relief” (updated edition) book, Ted Landau looks at the major ways to “take apart” Mac OS X.
Understanding Mac OS X: In Depth
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Although this is the kind of thing that you used to get in the box as the manual with the OS. 🙂
I do wish he’d touched a bit more on the mach kernel. He incorrectly identifies it as FreeBSD’s kernel….it’s not. In fact, mach is *much* better than a standard BSD kernel for OSX. And, you shouldn’t use native unix system calls (although OSX will handle those perfectly) to write apps, rather, you should use mach’s facilities.
But good article overall.
A agree with the poster above. The article gets *very* light when it turns to mach/bsd. In fact it seems the author has the user in mind. Not a single bad word about this bloody mach/bsd marriage and all its drawbacks.
Apple, do an OS X for Intel. I don’t understand how they can just sit down and watch their market dwindle away when they can potentially fly from 2% market share to 50%. What a waste! What collosal lack of balls!
There will never be an OS X for Intel. Apple is a hardware company, they just happen to write their own operating system. Apple could not survive without selling their hardware.
Apple wont realse a Intel version of Os X due to the simple fact that they are making more money on selling hardware then software. Why should I buy a Mac if I can run Os X on my Intel box… because the apple box look good on my desk ???
Apple makes money with hardware and only some with software.
Their marketshare is messured on sold hardware.
If they port OS X to x86 their hardware sells will fall to zero, which means: “Good bey Apple”.
I beleave that to one sold copy of x86-OS X come 1000 pirated copys. Again: “Good bey Apple”.
So – allways the same end of the story – no OS X on x86.
Concrete that means: No OS X for ordinary x86 hardware.
If might be an option that Apple switches their CPU architecture to x86 in the future (but I beleave that the go with the new PPC970), but I am sure that the OS X build for these machines will not work an any non Apple x86 machine.
Again no luck to install OS X on your Dell or whatever PC.
Is there proof that Apple couldn’t be making even more money from OS X for Intel than hardware Macs? But I don’t think this would happen anyday soon. Apple first needs to open up their market a bit to clone makers. Take steps Palm took.
>> Apple makes money with hardware
>> and only some with software.
Do you think Apple, or anybody in their senses, care how the money comes in? Software, hardware, ipod, hiphop, what do their shareholders care? They just want to make money and stay afloat, that’s why they are branching into things like selling music. Imagine how much money they will make from their excellent software if they crack into the PC market! They can even do a credible Office Suite, with Open formats maybe, and people will actually buy it. The possibilities are infinite!
The mac hardware is a dead end, with an obviously dwindling market share. You can only survive in your own little, expensive, hardware universe for so long. Apple can choose to deal with this now or, like SUN, do so a bit later, when the long term reality starts hitting home.
Again, apple DOES write software, has always written software, writes excellent software as a matter of fact, and so is as much a software as a hardware company. And Apple users don’t typically care about the metal, they care about what they can do with Apple software, be it OS X or itunes.
I don’t get this Apple being tied to hardware thing. I mean, does the average person who buys a mac really buy their mac because they want Mac OS X? I think most people buy macs because of their great looks and hardware integration.
Even if Apple sold OS X on x86, they would probably still be able to shift just as many of their already high end machines. A lot of people buy Apple because they build great hardware.
Consider the IPod. They ported the IPod software to Windows, and it didnt eat into their IPod sales – in fact I bet it improved IPod hardware sales.
I think if they ported to x86, they could still maintain their market share of desktop PC’s, and certainly excel in the workstation market.
I am not sure about the average user, but I bought a mac because I wanted OS X. The hardware integration would be the second reason followed by the great looks. I always kept track of the macs, but never actually wanted to buy one until OS X.
Not really sure what you mean by iPod software. They make an iPod for windows, which uses different software. Of course this improved sales because more people can use the iPod, but no money is made from the software.
You might be right about apple maintaining ther market share, which is small to begin with. However, I am not sure if much is too be gained from selling OS X on regular x86 hardware – especially with pritated software. IMHO, pirated software is a big reason Windows keeps its marketshare.
Apple OS X on x86 hardware is just a matter of making deals with whoever computer manufacturer (Dell/Gateway/HP, etc?) and licensing the OS to them, I mean doing exactly what MS does, selling the machine with the OS.
The other possibility will be manufacturing their own x86 bios/chipset. If Intel/VIA/SYS and whoever else do it, why Apple (who build systems on their own) is not going to be able of making it???
That way Apple does not loose money, and it’s operating system is spread and made popular amongst many people, if it works for MS why it is not going to work for Apple??? I can even imagine an ?easy to pirate? scaled-down version of the OS to attract more costumers to the platform (MS does nothing really useful to prevent piracy at the consumer level)
The answer regarding why they don’t do it is more related to the feeling of being “exclusive” (Anything different than the common) rather than technical reasons. It is a commercial reason.
Don’t doubt that if the water level goes too up Apple will release x86 machines.
I bought a mac for the UNIX like abilites of OSX, and because i could get a small lightweight laptop in the process. I did have a laptop with linux already but i was always “trying” and not actually “doing” anything with it.
The link doesn’t work anymore, either from here or from InformIt home page. They’ve been “osnews’ed”:) BTW, if they wouldn’t running that *$# IIS…
apple is currently trying to minimize its reliance on desktop mac products. That is why they have introduced software, servers, and why they are believed to be eyeing 64-bit computing and work stations.
If apple succeeds in even just some of these markets then they will have the economic means to Vastly reduce pricing on standard desktops which, i believe, could result in 10%-20% of the desktop market share (in terms of sales). they could, at that point, do a palm and license a FEW clone builders.
Diversification in the short to mid term is the winner for apple because that will let me take market share on the desktop. x86 is a quick death trap. no OS X on x86.
Apple did have a cloning program. It almost killed them. Twice. Once, as 90% of their income comes from hardware – so clones undercutting them on price meant Apple’s revenue stream dried up. And secondly, as they killed the cloning program, Motorola (one of the clone manufacturers) took offence and cancelled the G5 project. Which is why Macs have been lacking in the speed department.
So I doubt they’ll take that route any time soon.
releasing OSX for x86 has 2 disadvantages dor apple in the moment:
1: the current mac-hardware ist not fast enough! so the sales of thier own manufactured hardware will fall down dramatically
2: this would mean war with microsoft! and apple coud forget all msoffice support for their os! and compatibility is still a sales-arguement.
BUT:
whith IBMs PPC970 Apple has the chance to lead the speed rally for a while. than an x86 OSX would eaven help to sell ppc hardware, as everybody would see its avantages.
and whith a 100% compatible and stable office package available, they where indipendent from MS.
in this scenario i think, they even should give away OSX for free, at least the first two releases, just to gain more market!
and bigger market would bring more apps and games to there platform – and games they need to sell highend hardware!
so fare the master plan!:)
yeah, apple would losing alot of money if it would let macosx run on clone-hardware, for the reasons already stated.. And there’s another reason, Microsoft has a bunch of shares in Apple. I really don’t think that microsoft would like to see macosx competing on the same platform they are on.
It would be incredibly great for the dozen or so machines that it actually worked on.
There are only a handful of OSes that are really prepared to play Microsoft’s game of support everything hardware vendors come up with. One of the reasons MacOS can be developed as quickly as it is is the fact that they control the hardware and thus have a FRACTION of the drivers to write.
For instance, there are only a couple of video cards to support in the Mac world. This is compared to the hundreds, if not thousands of video cards that you’ll find in the PC world. This is just video cards…now add chipsets, sound cards, SCSI cards, NICs, mice, scanners, printers. It would be a mess! Don’t count on hardware vendors releasing drivers for Mac OS X86 any time soon, they’re just now starting to support Linux and (in a limited fashion) FreeBSD. Hardware vendors have been spoiled by having one platform to write drivers for and they absolutely LOVE the WinTel monopoly.
The point is even if Apple could get past worries about pirating and things of that nature, it would still be a HUGE drain on resources just because of hardware diversity.
Oh and stealing driver code from other open source OSes? Linux is out because of the viral GPL license that all of its drivers are licensed under. The BSDs are an option, but even that wouldn’t be easy because the Darwin kernel is a very different animal from a traditional BSD kernel.
So in short…no Mac OS X x86 unless Apple really feels like committing harikari.
-bytes256
The Apple Music Store sold 1,000,000 songs in a week. An impressive number, but still is only optimistically a half a million dollars revenue. But in the same week, they sold nearly 110,00 iPods. That is almost 35 million is revenue.
Great software to sell expensive hardware.
Also Microsoft does not have stock in Apple. Microsoft bought it five years ago and sold it before the stock dropped 50 dollars a share. And it was non-voting stock, giving them no direct or voice on Apple affairs.
Then they would end up like Digital Research, GEOs, BE, IBM and all the other companies that tried the same thing.
People wouldn’t buy the software, they would may copies and Apple would be dead.
yet another thread devolves into “{Mac, PC}’s are better than {PC, Mac}s” or “Apple should start selling Intel boxes.”
Personally I think Jobs would rather die than see a commercial Mac OS X beige box sold by anyone (whether that is Power PC or x86). Just a hunch.
Anyway, back on topic. I rather liked the chapter. It’s a little light on technical detail but it seems more geared toward another crowd.
Despite rumors to the contrary, x86 is really a closed platform for microsoft. The illusion of it being open is valid only for hobby kit makers and a few bold, but tiny, pc builders like walmart.
windows market share on x86 demonstrates this. No one, except maybe open source (because they have no cost), will ever share x86 with MS. ever. They control too much share and can dictate terms to both PC vendors (dell, hp, etc.), pc equipment makers (a/v cards, networking etc.) and even intel. Yes MS can threaten and control even mighty intel
Any attempt to leverage x86 for anything other than an open source project will end in failure. That includes apple. Apple is far better off expanding its own volumes and working with ibm and perhaps motorola to help push volumes for powerpc.
OS X on x86 = Death of apple.
x86 = closed platform (MS only)
powerpc = future of apple
OS X would be a nightmare on generic x86 hardware. Remember, the reason Windows fails about 90% of the time is due to hardware/driver issues. OS X works perfectly because it only runs on a very limited range of hardware. It knows EXACTLY what is getting connected.
Windows has an HCL longer than the friggin constitution. It’s easy to say “I want OS X for x86,” but when you really think it out, it’s much more ambitious than you’d think.
If Apple ever does move to x86 based chips, expect it to run on a very limited range of hardware – probably Apple manufactured only.
A move by apple to x86 hardware would be a disaster. The main reason cited by most proponents of an apple move to x86 are 1. poor performance of powerpc processors as produced by motorola and 2. decreasing the costs and increasing the customizability of the apple platform. However, Apple is in a position to make both of these issues non-issues and expand the platform.
If the IBM PPC970 is able to hold it own against Intel’s best(which preliminary although uncollaborated reports indicate this is true) then the performance issue would no longer represent a legitimate reason for a switch. Regardless, the existence of a PowerPC alternative is probably enough to hold Apple from making any quick decisions.
Apple should view the second issue as an opportunity to sell more hardware. You can’t easily build a custom mac by heading over to CompUSA or your online parts vendor and putting together the box you want. Now that they have been courting geeks with OS X, why don’t they go all the way by selling you the hardware to build your own box. In particular Apple should sell a mobo/processor combo that you could put into the box of your choice. Supported hardware and the associated drivers wouldn’t necessarily have to increase given the fact that you would be dealing with a more sophisticated consumer that is intent on building a specific box that they want. Apple would need to build a community web site(a la Gentoo, etc.) that would contain all the information needed to build a box using the set of hardware that is supported by apple. In addition, the site could store drivers, etc for unsupported hardware, etc.
The strength of Win/x86 is the ecosystem of interested parties, etc that are there to support the platform. It is not likely that Apple could build a similar ecosystem alone without canibalizing its own dwindling market. However, they could in combination with IBM and Motorola increase the community of the platform and therefore their market share and reach.
Saying that Apple is a hardware company is really dumb. How hard is it to write an operating system? Pretty hard. How hard is it to design a PC? Not hard at all. Would Apple have anything to sell if not for it’s OS? Nope. Is the cost of a Mac based on the cost of desiging the hardware and the OS? Yep. That’s why Macs cost so much. This is really pretty obvious….
Motorola did not kill the G5 because Apple kiled clones. Get real. Motorola sold very few clones and made little effort to do so.
Apple is currnelty minimizing its reliance on desktops becasue they are not selling very well. The servers are a move into the corproate space, the laptops are more of an attempt to make up for slow desktop sales. Once the PP970 Mac is released desktop sales will be the mainstay of their business once again. I don’t know where you get the idea that Apple has, or will have “workstations” as opposed to desktop Macs.
Moving to Intel means every app written for OS X has to be rewritten for Intel processors. tahts a really big deal and many developers who just made the transition from Mac OS to OS X aren’t going to be up for that. If the PP970 is all it appears to be then there is no reason to go Intel anyway.
Apple didn’t port ipod software to Windows. They made the ipod use a Windows disk format and use MusicMatch jukebox as the itunes replacement.
…and your comments are really dumb.
“Saying that Apple is a hardware company is really dumb. How hard is it to write an operating system? Pretty hard. How hard is it to design a PC? Not hard at all. Would Apple have anything to sell if not for it’s OS? Nope. Is the cost of a Mac based on the cost of desiging the hardware and the OS? Yep. That’s why Macs cost so much. This is really pretty obvious….”
Apple makes BILLIONS of revenue on HARDWARE. They make a couple hundred MILLION of revenue on SOFTWARE. More than twice that much software revenue is spent on software R&D. So… Apple makes ZERO money on software and BILLIONS on HARDWARE.
How can you even try to say that they aren’t a HARDWARE company?
Apple uses software to drive their hardware sales. Yes, their innovation, etc.. is in software as the hardware (excepting iPod) is mostly standard stuff. But their software drives the high revenue / high margin hardware sales. That’s where the money comes from.
Apple could make it a lot easier and just lower their prices. Afterall, everyone who wants osx on intel does bc then they could afford to buy a “mac”….
>whith IBMs PPC970 Apple has the chance to lead the speed rally for a while.
Why do you think so? Nothing that reached the public so far suggests that.
Lets get this straight:
The PPC970 is a downgraded version of IBMs Power4 server CPU with Altivec attached. It is one big hack out of desperation because there are no other half-way fast PowerPC chips available.
The PPC970 is a 64bit CPU because it was designed for server use. On the desktop 64bit is still pointless.
Its only useful for a few specific tasks and desktop computers aren’t normaly used for any of them.
All ‘leaked’ benchmarks showed that the current x86 chips from Intel already wipe the floor with PPC970 when it comes to desktop tasks. The few benchmarks where PPC970 kicks ass are mostly unimportant for real world desktop preformance.
The PPC970 is still months away. And at the beginning it will probably only be used in Apples high-end/server machines. At the time you will find it in an iMac the x86 chips from Intel will be nearly twice as fast as the PPC970 for most tasks.
Of course without the PPC970 it would be much worse.
Apples will never be fast again. MacOS X design is preformance-agnostic. The core (Mach) is slow, the GUI ontop is a true preformance killer because of all the
eye candy AND the technical design (displayPDF). And MacOS Xs main programming language – Objective-C – produces slower code than C/C++. Apart from that Intel has much more CPU development power and the Intel engineers are actually trying to create fast desktop chips. IBM, just like Motorola before, doesn’t. They develop server CPUs for their own servers and Apple just gets leftover downgraded outdated hacked up versions of them so they have at least some CPU to put in their shiny boxes.
I doubt that this will change. The cooperation between IBM and Apple is just as half-hearted as the old Motorola-Apple one.
That said: I will buy a 64bit MacOS X Apple as soon as I can afford one 😉
“The PPC970 is a downgraded version of IBMs Power4 server CPU with Altivec attached. It is one big hack out of desperation because there are no other half-way fast PowerPC chips available.”
Baloney. The 970s superscalar/vector unit is designed for this new family–its wider than the Power series and much better than the G4.
It is not a hack–it is a new line that IBM and Apple both have great things planned for.
“The PPC970 is a 64bit CPU because it was designed for server use.”
No, the Powers were designed for server use. The 970 family is designed for workstation, low-end server use. THe whole PPC line was designed for ultimate migration to 64 bit. It is not a remnant nor an afterthought.
“All ‘leaked’ benchmarks showed that the current x86 chips from Intel already wipe the floor with PPC970 when it comes to desktop tasks. The few benchmarks where PPC970 kicks ass are mostly unimportant for real world desktop preformance.”
Untrue: benchmarks show it either in line (slightly behind or slightly ahead of) anything Intel is offering — at 3x the clock cycle speed and twice the die size, or substantially ahead.
I agree, if prices come down, then apple’s share will go up. However, prices won’t come down significantly until apple has a means to offset the potential loses in revenues.
“Motorola did not kill the G5 because Apple kiled clones. Get real.”
There’s no proof of this either way, but it’s not as stupid a supposition as you’re making it out to be. Motorola announced that they were making a “focus shift,” to borrow a phrase somewhat infamous with old-timers around these parts, away from desktop CPUs to embedded systems very shortly after they were forced to pull the plug on their StarMax PC line. While I’m not even sure if the G5 was in the PowerPC road map then (this was when the PPC 604 was high-end and the PPC 620, which became the G3, was in the pipeline), Motorola made it pretty clear they weren’t interested in future development for PCs. In point of fact, all of the information on their website about the G5 when it was in active development was very pointedly focused on embedded systems, and most of the features of Motorola’s PPC line developed in the last few years aren’t useful in the desktop space. The G5 as last described wouldn’t have had a higher clock speed than current G4s do.
It’s clear that Motorola doesn’t value Apple as a customer, and it’s impossible to deny that this represents a serious shift from the days of the “AIM” partnership–and this partnership clearly fractured shortly after Steve Jobs’ return to Apple and the dismantling of the clone program.
Isn’t anyone else reading about the new 970 chip from IBM. It is 64…yes 64 bit…and quite fast. It will blow the doors off of Intel. The Pentium is near the end of it’s cycle and dying a slow death.
Go here: http://www.macworld.co.uk/news/top_news_item.cfm?NewsID=6313
Apple doesn’t need Intel…when it has iPod, iTunes with a music service that so many people are drooling over. Apple is indeed a software company that happens to package it all.
Another aricle recently tells about Microsoft in the early stages of emulating the Aqua look…another Apple patent.
BTW, Apple isn’t lacking in the speed department for most things if you use Safari…and have lots of ram for OS X.
porting OS X applications to intel is a non-issue in most cases thanks to OpenStep, or as apple now calls it cocoa. and the foundation of the operating system is already ported to x86 and lacks only drivers. OPENSTEP, the operating system, and rhapsody, the software that became os x, both had working intel versions. apple only killed intel support right before the release of os x server. it’s a massive non-issue. I just never expect OS X on intel to ever happen, for other reasons.
as for the speed of OS X…. jaguar runs very nice and smooth on a 500mhz imac. quartz extreme made some massive improvements in this area. unfortunately, previous comments are correct about OS X being a ram killer.
I dont want os x for intel…. i want a nice implementation of cocoa on top of GNU/Linux that doesnt look like ASS like gnustep (i’d like it to not be nearly as buggy as gnustep as well)
How quickly one little side note always seems to deteriorate into all the posts by posts by so many of the smartest (?!) people on the net explaining all the errors in Apple’s business practices. I’m sure you all have more money in the bank than Apple and have also all weathered the economy with mass increases in your salaries and assets. I would very much appreciate it, therefore, if you could all just send me a mear 1% of all your millions and tell me where you studied economics so I can go to the same school and learn all those things too! (and maybe even afford a spell checker.) Some of you may want to invest in a spell checker too. It might add a little strength to your postings.)
Cheers!
Partisan politics aside ..
The except contains factual errors. As far I could read
before giving up, they completely misdefined Unicode and
kexts.
As a poor, unemployed geek who can edit copy (as can many college graduates) … I have difficulty stressing how aggravating it is when <em>technical</em> books are issued without technical editing and fact checking. Those errors I noted could’ve been cleaned up by a quick google.
*must restrain Fist of Death!*