Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Why did Apple rewrite Mac OS code?
- Thread starter majordude
- Start date
- Sort by reaction score
Apple from the very beginning never had 100% of the marketshare, sure there wasn't always Windows, but there were other alternatives.
A certain amount of nonsense has crept into this thread. A number of posters assert that Windows has been completely rewritten. No. It has not. Nada. Allow me to demonstrate. Unpack an old windows 3.11 for workgroups machine and watch an avi movie. Once the movie has started playing, move the window. What's this? The window moves but the video doesn't? Now, sit down at your XP Pro box and do the same thing. Try to move or resize a window which contains a playing movie. The window moves but the video stays put. I don't know for a fact this is true on Vista but I have my suspicions. There are countless other areas where this kind of thing is going on.
Yes, M$ made a monumental effort going from DOS to NT. They broke so many things, they had to continue to update their 16 bit OS several more iterations (95, 98, 98se, Me) before they FINALLY managed to come up with a protected kernel OS that could run 16 bit apps and games as well as business apps. That OS was XP. Otherwise known as some flavor of NT, but the NT name hasn't been used since Win 2000.
Now here's the sticking point. Go to msconfig. It STILL has hooks for win.ini. Win.ini is a legacy provision that goes right back to windows 3. Then there's the registry. Why did Vista break so many things so badly? M$ finally tried to lock down the registry and windows system directories and those programmers who ran roughshod over your system before had to learn all over again. This does not qualify as the kind of rewrite that OS X represents over OS 9.
In all fairness, Apple bought OS X. Fine. Let M$ buy their next OS. I suggest they talk to the people over at RedHat or Ubuntu. I would LOVE to see Windows run on top of Linux. Finally the OS would come out of the dark ages IBM/M$ and others caused by trying to make it easier for us end users. Apple finally saw the light and we have a stable 5 9's uptime OS to show for it. XP has excellent uptime but it's not 5 9's. I have a reasonable expectation that 99.999% of the time my Macbook will be available for what I need it to do when I need it. Period. If I had an e-commerce site and I had to process $1 Billion in transactions and a 5 minute outage could mean a loss of $50 Million, there is no way I would pick M$ infrastructure. I might consider Linux, OS X or even Sun, but M$ would not make my short list of OS options in a high availability environment.
There is some speculation that M$ will rewrite the OS for windows 7. I doubt it. It's like trying to turn an ocean liner in less than 5 miles. They are intimidated by the tens of millions of users out there that are using legacy apps. They are afraid these people will bolt. Their fear is justified because today those people have viable options both from Apple and from the open source community.
There is also some speculation that M$ is busy at work on visualization. This could, in theory save M$ from themselves. Finally they could put legacy code in a shatterproof bottle and focus on building a new OS with no regard to supporting legacy apps. Of course, there is a legion of developers who have been dragged down the trail of Microsoft Foundation Classes and most recently Dot Net. Those two API's are abominations. There is no evidence M$ will ever offer an API that is not an abomination. The problem is, M$ would rather support 50 methods of doing a thing than require any of their users to ever learn something. It would be oversimplification to say the fault for all this legacy support lies only with M$. There are fortune 500 companies with IT people who reluctantly moved from COBOL to Visual Basic who threaten to find a new solution if M$ breaks some of their legacy code beyond repair.
So let's get back to the topic of a rewrite. The only way it can happen is if M$ does what Apple did. They need to buy an OS and adapt it. Any in-house solution will hit the same filters that have drowned out M$ innovation all these years. Any in-house solution will yield another exercise in backward compatible mediocrity. To me, for a Windows rewrite to justify the name rewrite, it has to jettison the obsolete chunks that made a rewrite desirable in the first place. For windows it has not happened and in all likelihood it is never going to happen.
Yes, M$ made a monumental effort going from DOS to NT. They broke so many things, they had to continue to update their 16 bit OS several more iterations (95, 98, 98se, Me) before they FINALLY managed to come up with a protected kernel OS that could run 16 bit apps and games as well as business apps. That OS was XP. Otherwise known as some flavor of NT, but the NT name hasn't been used since Win 2000.
Now here's the sticking point. Go to msconfig. It STILL has hooks for win.ini. Win.ini is a legacy provision that goes right back to windows 3. Then there's the registry. Why did Vista break so many things so badly? M$ finally tried to lock down the registry and windows system directories and those programmers who ran roughshod over your system before had to learn all over again. This does not qualify as the kind of rewrite that OS X represents over OS 9.
In all fairness, Apple bought OS X. Fine. Let M$ buy their next OS. I suggest they talk to the people over at RedHat or Ubuntu. I would LOVE to see Windows run on top of Linux. Finally the OS would come out of the dark ages IBM/M$ and others caused by trying to make it easier for us end users. Apple finally saw the light and we have a stable 5 9's uptime OS to show for it. XP has excellent uptime but it's not 5 9's. I have a reasonable expectation that 99.999% of the time my Macbook will be available for what I need it to do when I need it. Period. If I had an e-commerce site and I had to process $1 Billion in transactions and a 5 minute outage could mean a loss of $50 Million, there is no way I would pick M$ infrastructure. I might consider Linux, OS X or even Sun, but M$ would not make my short list of OS options in a high availability environment.
There is some speculation that M$ will rewrite the OS for windows 7. I doubt it. It's like trying to turn an ocean liner in less than 5 miles. They are intimidated by the tens of millions of users out there that are using legacy apps. They are afraid these people will bolt. Their fear is justified because today those people have viable options both from Apple and from the open source community.
There is also some speculation that M$ is busy at work on visualization. This could, in theory save M$ from themselves. Finally they could put legacy code in a shatterproof bottle and focus on building a new OS with no regard to supporting legacy apps. Of course, there is a legion of developers who have been dragged down the trail of Microsoft Foundation Classes and most recently Dot Net. Those two API's are abominations. There is no evidence M$ will ever offer an API that is not an abomination. The problem is, M$ would rather support 50 methods of doing a thing than require any of their users to ever learn something. It would be oversimplification to say the fault for all this legacy support lies only with M$. There are fortune 500 companies with IT people who reluctantly moved from COBOL to Visual Basic who threaten to find a new solution if M$ breaks some of their legacy code beyond repair.
So let's get back to the topic of a rewrite. The only way it can happen is if M$ does what Apple did. They need to buy an OS and adapt it. Any in-house solution will hit the same filters that have drowned out M$ innovation all these years. Any in-house solution will yield another exercise in backward compatible mediocrity. To me, for a Windows rewrite to justify the name rewrite, it has to jettison the obsolete chunks that made a rewrite desirable in the first place. For windows it has not happened and in all likelihood it is never going to happen.
Before NeXTSTEP entered the picture, it looked virtually certain that Apple's post-Classic OS would be built on BeOS:
http://en.wikipedia.org/wiki/BeOS
Apple also provided a very nice transition mode by supporting Classic under OS X on all of their systems until the switch to Intel and on PPCs until just last year (i.e., six and a half years after OS X came out).
Other than the fact that the author thinks media players are the operating system, you mean? Actually, I'm being unfair. That aside they do make really good points. But still...
OS X is returning home -> Intel (x86). On PowerPC OS X was not at home:
(it was just an intermediate stop - os9 -> osx)
"the Mach-O ABI (Application Binary Interface), "which defines how every application in the system executes and calls functions (stack conventions, register usage, and more" was optimized for the CISC architecture of 68k (Motorola) and later x86 (Intel) processors on which the NeXT operating system ran, not the PowerPC RISC architecture Macs use today. "Mac OS X uses an ABI designed for CISC processors, mostly ignoring RISC design principles," says Unsanity, in contrast to the fully RISC-optimized ABI in Mac OS 9."
http://www.powerpage.org/2002/10/geek_report_key.html
That's funny! Every piece of software microsoft has ever sold, they bought from someone else. The only exception is Microsoft Bob. Every thing else, including Word, Excel, PPT, Outlook, Internet Explorer, SQL Server, Exchange Server, and many core pieces of Windows OS itself were all acquired from third parties.
