Apple is moving in the wrong direction. Closing access to the OS X kernel on Intel processing architecture is not a good idea, it will only cost Apple more time and money to get a good kernel for their processor. Why did they move to FreeBSD in the first place if they are going to make the kernel source on Intel proprietary? What a waste of potential, what a waste of time, what a missed opportunity. Without the ability to modify kernel source serious developers will not touch the Apple platform, maybe that is why the OS X server is not making significant inroads into the server community. Apple has probably shipped tens of thousands of X Server machines, but surely not one hundred thousand. They will never see the growth numbers that linux sees in the server market.
In the end this may be fine, Apple is really a consumer electronics company, but the next killer app lurks out there and while it will surely be platform-independent, what with all the virtualization these days, I bet it will be developed on linux, depriving Apple of an opportunity to profit and innovate.
Perhaps the question is ‘will Apple have enough innovation to maintain a pipeline of consumer products to remain vital?’ On the computing side, they only have a limited number of tricks left to pull out of their sleeve, switching to AMD is likely not one of them.
*now* you're just whining in my opinion. I get it that some people think Apple should keep the kernel open. But a) there has _not_ been a definite word about this yet by Apple (so we _might_ see the kernel sources reopened!), b) this might mean there _is_ something going on in the back and we might see a completely new kernel in Leopard, maybe... and c) this might simply be a measure to keep people from running OS X on the "any PC" right now. But stating again and again that Apple right now doesn't let you hack the kernel on intel Macs really seems like whining to me. And saying that "real" developers won't touch a platform with a closed-source kernel is both pretentious and ignorant, but more importantly: A very closed-minded view. And if open source people close their minds - what's left?
Actually, I don't think Apple is really trying to sell the server platform to your "real" developers as a market at all, are they? How many of those are there? The way I see it, the server OS is an office / small business goldmine. I have installed it in over 30 offices to take the place of badly functioning, support intensive, Windows 200x servers. The point is Apple is making open source tools and applications available to people who don't have time or budgets to waste becoming IT experts - but would rather do things like run their businesses.
I have heard the rest of this arguement before, for years. While it seems like a great idea to just put the OS out there, IT IS NOT THE BUSINESS THAT APPLE IS IN. They make their money selling hardware, and they do it well enough for me to invest my money in.
Closing access to the OS X kernel on Intel processing architecture is not a good idea, it will only cost Apple more time and money to get a good kernel for their processor.
Huh? First off, Apple has not "closed" the source to the OS X kernel, they merely haven't yet released the source. See the statement by Apple manager Prabhakar referenced in Drew Thaler's blog post that discusses this issue and also points out that there are probably good reasons (proprietary code) why Apple is delaying the release:
http://drewthaler.blogspot.com/2006/05/delaying-xnu-x86-source-release.html
And Apple employees are almost entirely the only ones who work on the OS X kernel - this hasn't changed. So where's your "more time and money" statement coming from? Apple has been releasing the source for the OS X kernel so far just for the benefit of its users - as far as I understand Apple has benefitted very little from outside contributions.
Without the ability to modify kernel source serious developers will not touch the Apple platform
Huh?? I guess you are talking about kernel developers - but that doesn't make much sense. Apple is really only interested in application developers. Application developers don't in general care in the least about whether the kernel is open-source - they just care that it does its job and doesn't get in their way. Application developers work at a level far above the kernel.
As far as I read at Arstechnica (http://arstechnica.com/staff/fatbits.ars/2006/6/19/4376), they cannot release the source code now, because it contains licensed code for Rosetta. That makes perfectly sense to me.
This is the biggest load of rubbish I think I've ever seen on the O'Reilly network.
Without the ability to modify kernel source serious developers will not touch the Apple platform...
So what is a "serious" developer? Do you have to be into open source or enterprise systems to be "serious"?
I don't really have a problem with the conclusion that Apple is mainly into consumer stuff, but I'm confused about this article. First it's about the kernel and open sourcing it or not, then it's about consumer electronics, then it's about innovation and AMD or not... perhaps there is a common thread here but unfortunately Jeremiah doesn't pull it all together, in my view.
The reason they are not making progress on the server front is VERY EASY. I am a teacher and my department has six twenty-five machine labs, two x-servs down the hall. They are unusable for running our exams when the simultaneous MySql queries ramp up. Apple decided to screw arund with the way the kernel queues requests and they just won't talk about fixing it. There is no point in trying to use it because their engineers are just trying to protect the company and won't help in any way. I have been supporting my network without district or Apple support for five years (one good engineer in that time). OH! I forgot to talk about trying to get LDAP to function without weird storage path mistakes that have to be fixed by hand. We are going to move to Red Hat. Our network admins said they couldn't allow it becuase they weren't prepared to support it because of training expense. ROTF-dying- We haven't gotten any domain management support in years for goodness sake.
As others have pointed out, Apple has not officially announced that the Intel version of the Darwin kernel will be closed. They just haven't released it yet. Also, I had to laugh out loud at this comment:
he next killer app lurks out there and while it will surely be platform-independent, what with all the virtualization these days, I bet it will be developed on linux
Please. There are no "killer apps" for Linux (respect to Apache though). As awesome as Linux is, in terms of desktop applications, it is strictly a "me too" platform.
I'm a fairly "serious" developer, but I have never, ever needed (or wanted) to re-compile kernels, or tweak them. I agree it should be kept open, but this particular argument is specious.
Wow, what a remarkably uninformative statement of opinion. Virtually every single line in this poorly reasoned, fact-free, piece is based on opinion and conjecture. Why was this published?
You Suck! I had a more interesting comment to make but why waste my time. You obviously don't understand much.
Apple isn't popular in the server space because Darwin is friggin slow, and thus isn't cost effective since it requires more hardware than Linux and BSD to sustain the same load.
Jerermiah - not sure how virtualization helps platform-independence, seeing as it's about running multiple operating systems on one box. You're still writing an app to one of those platforms OR in Java. Of course the next killer app might be a browser-based cross-platform app. And aside from development tools, I've yet to see any killer applications on Linux. Honestly.
Find me one application that would make a non-IT literate Windows or Mac user jealous that it is not available for their machine. And I would say a major reason that will remain the case is the contempt of self-styled 'serious developers' towards CE-style ease-of-use. FOSS developers tend to look at something like iMovie or Keynote and think about replicating the functionality, disregarding the fact that a lot of what makes it 'killer' to people doing presentation is the pre-designed themes, not the flexibility of the software.
For the record, I don't think it's news that Apple is a CE firm. The original Mac was the first CE computer - you couldn't upgrade it, and unlike all previous systems it didn't come with a programming language. NeXT was more developer oriented (like Sun). Merging the two firms has resulted in an interesting hybrid - OS X has been the first time I've been interested in what Apple have been doing since the 80s. Do I care about the kernel?? Not really - I've recompiled the Unix kernel enough times, that I am quite happy never to see one again. And I think that Apple are getting better and better at working with developers. (Equally, lack of access to the kernel has not exactly held back Windows development).
Bob - my understanding is that the problem is at much with mySQL as with Apple. Oracle have optimised their DB to run on OS X architecture and have no performance problems compared to their Red Hat version. No one has done the same with mySQL. Of course, it's open source so you are welcome to do so, or pay someone to do the work, but it seems unfair to blame Apple when it is quite clear that other RDBMS work well.
I think this is one of the big issue with open source cross-platform ports - just because something compiles and runs on multiple platforms does not mean it will perform on all those platforms.
So yes, Apple have mucked with the kernel to suit their own ends, in such a way that mySQL does not execute well. (Isn't the issue down to the cost of starting threads?). However, I understand that the standard Mach kernel and Linux kernels are less efficient in the area of the messaging protocols used by most OS X native apps.
"Apple is really a consumer electronics company" Thanks, this is what the rest of realized 20 years ago.
"Apple is really a consumer electronics company" Thanks for the news, this is what the rest of us realized 20 years ago.
Some good reading:
The 'Mac OS X Closed by Pirates' Myth
The Mac OS X Linux Kernel Myth
Mac OS X Microkernel Myth
MySQL myth
Dude, Xserve is PPC, and the complete PPC kernel is available! Only certain parts of the Intel kernel are not available, obviously due to 3rd party licensing such as Rosetta. Are you just looking for page hits?
BTW, Xserves and XSAN have been a quiet hit and Apple knows this. Look at companies like Oracle, who completely support it. Why do yo think that is?
Apple is a consumer electronics company. I am a consumer, and therefore have bought apples for a really long time. I can do a little bit of hacking, a little bit of work, and a little bit of play. This was true for Apple, which was targeted to the hobbiest consumer, and the Mac which was targeted to the small firm consumer.
The problem was that if one was not somewhat tech savy, it was easier to purchase a PC for home use, even though few PCs, and certainly MS Windows, is not really a consumer OS. What Apple has done is created a situation where a consumer can buy a Mac, but still have MS Windows.
If Apple is a consumer electronics company, then it need not worry about an OS. No one worries about what OS the DVD player runs. No one worries about what OS the car runs. I am not totally happy with the direction of Apple, given that I am a hobbyist, and don't like Intel. But when the next big thing comes around, Apple will be well placed to take advantage of it.
And I will tell you why. Because Apple is flexible. We do not have Mac OS X media edition for the iPod. We do not have Mac OS X lite, pro and enterprise. Apple does not seem to have this need to fit everything in the world into a single master plan and brand it. Apple will make changes as the technology allows, no matter how much the customers complain. And some of those will be a success.
Jeremiah, research the facts more thoroughly before publishing. Else it's a waste of your time (writing), our time (reading expecting some keen insight), and O'Reilly's brand reputation (which influences our next book purchase decision).
[URL]http://www.musica-latina.anticoit.org[/URL]