Why would I want a laptop to have a great graphics card?
That's the wrong question, because the Intel GPU isn't just "not great", it's "bloody awful".
The right question is "Why would I want a laptop to have a complete OpenGL implementation?".
Because this chip's OpenGL is incomplete, and Mac OS X's graphics is built on top of openGL. So one of the two cores in your Core Duo is going to be sitting there running OpenGL emulation... less efficiently than even an entry-level ATI or nVidia GPU... and generating heat. And the extra RAM you'll have to install, both to support this extra code and to make up for the missing VRAM (that's another thing, this chip doesn't have any), will be generating heat too.
If I have to choose between a cheap and power-savvy low-end card and a high-performance but expensive and power-hungry one, I choose the first.
Unfortunately Apple doesn't give you that choice. You get a cheap chip that makes the computer as a whole more power-hungry. A low-end ATI or nVidia chip would have given you the "cheap and power-savvy" option, but just as in the Mac mini you're stuck with what they provide.
So basically they don't have an acceptable low-end laptop, and they don't have an acceptable low-end desktop. So much for the Intel switch making Macs more affordable.
If it wasn't for OS X I wouldn't consider ANY Apple laptop even vaguely acceptable.
Even the Powerbook/Macbook Pro keyboard feels like crap after you've used a Thinkpad, and of course the single mouse button on the trackpad just adds insult to injury.
On the microkernel side we have Minix 3, 15 years after the first not-really-open-source-but-code-available microkernel UNIX systems.
On the monolithic kernel side we have... what? 15 years after the first not-really-open-source-but-code-available monolithic kernel UNIX systems we had... hmmm... things like MiNT, and bits of BSD, but even Net/1 was a few years in the future and Minix wasn't even out.
I think, after you allow for the 20 year head start, microkernels aren't doing that badly.
What really bugs me about Microsoft is not that they're "the evil empire", but because they're not "a competant evil empire".
The problem is that Apple seems to have a lot of the same areas of profound incompetance as Microsoft - style only carries you so far - and while they avoid a lot of the evil they've got their own home-grown kind that Microsoft only seems to have borrowed part of from them.
Any modern day browser would ask the user if they want to open the file, whether it was a click or a redirect.
Just to make clear, though, that has nothing to do with.NET. You seem to be talking about a standard browser open.
I'm talking about the integration of the browser with locally executed unsandboxed code, whether that's implemented using ActiveX,.NET, or even Firefox's XUL.
The browser should not ask the user for permission to open a "file://" URL, because it should not be able to do anything "unsafe" just because the document is in the "My Computer" zone. Or any other zone. The browser shouldn't allow a document to do anything "unsafe", period, regardless of where the file is or what type of file the browser thinks it is.
I had to save a Perl file on the root of the hard drive, so it asked me for the admin password. No problem. Later on, I had to do some work on the file and it asked me for the password when I opened it. Understandable.
Now, I'm working on this file, and I save it a few times. EVERY SINGLE TIME I hit Ctrl-S, it asks me for the password. Why?
Because you're entering a new security domain... running the "save file" procedure in a hidden "root" subprocess... every time you save. The problem is that the system made it too easy to routinely change your security domain. Rather than saving the file to the root directory of your system disk, you should have kept the working copy of the file (the one you're editing) in your own directory (eg, on the desktop or in documents), and put it in the write-protected root of the file system (if you had to) only when it was working.
This dialog isn't really the same kind of thing as the one I'm talking about, though. It's not there to warn the user that they're doing something dangerous, it's to verify to the system that the person who's requesting the action is really the account owner... which is why it asks every time.
The perception that this password dialog has something to do with "keeping you from accidentally doing something dangerous", like the typical approval dialog, is a common and mistaken one. It's got nothing to do with that at all.
The solution isn't to make the dialogs better, or to make the dialogs go away, it's to design the system so that the dialogs aren't necessary, to make it easy and obvious how to work without being interrupted by them. Apple made a mistake here... the editor should have offered to save the file somewhere else, and moving it back into the protected directory should have been a separate process... so it was clear that editing the file in place wasn't a normal thing to do.... so...
Getting back to the Windows example, the privileges granted a document by an application should not be related to the zone the document is in. They should be related to the security domain the application is in and the role within that domain the application plays.
That is, the HTML display component (Webcore in Safari, Gecko in Firefox, KHTML in Konqueror, the HTML control in Internet Explorer) should not have a built-in mechanism to grant local user privileges to a document or a component of a document at all. Not via ActiveX, not.NET, nor active scripts. That mechanism should be managed by the application that is using the control, preferably by installing an extension in the component as Dashboard on the Mac or the KDE desktop do. Any application that is used to display untrusted content (Mail.app, Safari.app, Outlook, Internet Explorer, Firefox, Konqueror) must not provide that extension. Applications that need to execute unsandboxed code (such as Windows Update, the Windows control panel, or Dashboard) should be (and in the case of Dashboard, are) implemented as shells that don't load sources of untrusted content at all.
Do this, and you don't need any "I'm about to do something stupid" dialo
I've read that before, and I can't agree with everything Neal says in there (and I think Cryptonomicon missed the boat in a lot of areas), but it's an interesting read.
Except for iTunes and Quicktime, which are the applications that the iPod and the iTunes Music Store (that seem pretty important to this "platform" thing) depend on.
You listed three generations of two operating systems with the same calling conventions, library formats, and process structure. The biggest difference is between Windows CE and Windows 9x, because Windows CE has dropped most of the DOS shell support, so to run a shell in CE you have to emulate it.
Really.
You can run applications for any of these systems under the Windows NT just by providing the right libraries. I've compiled Pocket PC code for x86 instead of ARM and run it on top of Win32... the "Pocket PC emulator" doesn't even seem to keep you from calling Win32 libraries as well as WinCE ones. These are all the same species of OS... to the programmer, they're like different breeds of dog.
There's more differences between Mac OS X/Cocoa and FreeBSD/Gnome than between any of those systems, even with all the FreeBSD code in Mac OS X. Hell, just the differences between Mach-O and ELF are huge. The iPod isn't even the same phylum.
Perhaps it's time for spybot like program for Macs and Linux just to remove tracking cookies?
Perhaps it's time to quit worrying about cookies?
Safari has a similar option to "Accept cookies only from sites you navigate to".
Camino lets you "Accept cookies only from sites you visit".
And advertisers know about this feature, so they use other tricks like correlating your IP address with referrers or using tagged URLs to gather the information they're looking for.
Internet and Local Intranet don't have that kind of access in.NET.
You mean that if a web page does something like "location = 'file:///c|/...'", or gives the user a link to a local file that the user clicks on, it won't load the specified file? Doesn't matter whether the file is visible in the context of the untrusted document, the question is whether the file can be loaded at all.
The dialog boxes are there so users aren't completely annoyed, trying to find where to grant apps access.
I have had the same people come to me multiple times saying "Peter, I did it again, a dialog bax came up and asked me if I wanted to run something, and now my computer's acting strange. Can you help?"
Technically speaking, this is a social engineering attack, but because users are presented with so many warning dialogs they're trained to automatically approve them. Because they need to approve so many of them just to get their work done. So an approval dialog (no matter how worded) doesn't have enough of an effect on security to matter.
Actually, if you know anything about Apple the iPod is exactly like the Mac.
Did you read the article?
It's a different OS (two different OSes) with a different user interface and different applications base. It's not a "platform" in the sense that Windows is a "Platform".
Apple doesn't own "it's a one stop solution". They don't control "it's a one-stop solution". "It's a one stop solution" for Apple is like "the Internet" for Google.
They're not releasing a new consumer version. They've continued to make the embedded version available, and the Pocket PC and Smartphone systems are based on it, but the Windows CE based "laptop replacement" devices are dead dead dead...
So you say that IBM and Apple don't take open source software, and fund their development whilst giving it away for free? Gee, IBM got sued for doing that. Better let the judge know it never happened!
No, I don't mean a $150 68000 device with 2M running palmOS 3.5 in 2000, when it was the replacement for the Palm IIIe and contemporary with the Visor and the original Clie.
I mean they should have maintained the PalmOS 4 68000 based line and let the price drop and the capacity increase as the cost of memory and chips fell. They didn't have to enhance it and come out with a PalmOS 4.3, 4.4, 4.5, just keep making the black-and-whote DragonBall-EZ based Palms with 8M RAM and selling them as cheaply as they could.
They got down to $100 and stuck there, adding power that the entry level doesn't need, upgrading the screen and processor and memory instead of pushing the price down and going after the grocery stor checkout lines and the educational market until every high school student had one instead of a Ti-83 or whatever this year's sine-qua-non calculator is.
I think he is refering to the closed format of the Apple music platform.
AAC is just another name for MPEG4 encoding of music. That's an open standard.
The only thing closed is the DRM, but DRM has to be closed, because DRM depends on obfuscation to work. DRM involves giving someone an encrypted message, the decryption keys for the message, and an implementation of the decryption algorithm, and keeping them from reading the message except when you want them to.
To quote Douglas Adams, "This is of course impossible".
Microsoft's DRM is less open - the only platform it runs on is Windows (Plays For Sure devices aren't "platforms" - they only run one application) whereas iTunes runs on Windows and Mac OS. Real's DRM isn't any different from Microsoft's. Sony's DRM is dead.
And it doesn't matter if Apple's DRM is open or closed because the first thing any sane person does with the music they buy from iTunes is burn it to an audio CD... and once you do that you can play it on any device.
If someone can find a way of taking open source software, and finds a business model that allows them to fund their development whilst giving it away for free, it's bye, bye Microsoft, Novell and a few other companies who make their livings from pure software licensing.
"browser-integration with.NET".NET's pretty darn secure.
Is it?
Enlighten me.
Is there a mechanism whereby an object in Internet or even LocalIntranet can request the execution of an arbitrarily specified object in the MyComputer zone? If so, does the user get a dialog asking whether this execution should be permitted, or is it unconditionally denied regardless of the user's settings? Is it possible for a user to specify that a specific object (based on any criteria, whether URL or certificate or address or strong name) be granted this right at the time the request is made?
Unless the answers are "no" or "yes, denied, no" then it's not "pretty darned secure".
For anyone who read the article, the author suggests that Microsoft should license Vista and Office for no more than $50.
No, he suggests that Microsoft sell a new OS that's actually usable on existing computers and doesn't have the legacy bloat and security problems of Windows... for $50.
Microsoft can build software for a handheld or tablet computer, a mobile phone or a TV set-top box and even though the wrapper is different, the feel is always very much the same -- that of a fat PC client. Microsoft can't allow a phone to be a phone because they can't dominate and control a plain old phone unless it is more Windows than phone. That's a problem.
It surely is. That was obvious in 2000 when they came out with "Pocket PC", their most successful spin on the handheld, and "Stinger", their fialed attempt to get into the cellphone market.
The Pocket PC meant the end of the Windows CE micro-notebooks and the Windows-CE-based tablets. They were pushing Windows NT as the new tablet... the problem is that while Windows CE felt like a spin on Windows 95, and the Pocket PC felt like a Palm on steroids, the Tablet PC was just an overpriced notebook.
Luckily for Microsoft, Palm had no idea what their product was, and has been trying to turn Palm OS into Pocket PC... and failing, big time. If Palm was smart they'd be selling black-and-white 68000-based Palms for $30-$50 in every grocery store in the USA, and they'd still own the business... because Microsoft couldn't do that. But, no...
But, anyway... Microsoft's platform is Windows. If you're not Windows... even if you look like Windows, Microsoft just wants to make you an annex to the Windows desktop. And if you don't even look like Windows, Microsoft doesn't want you to be a platform. That's why they completely redid the XBox, people were turning it into a platform.
But what's Apple's "platform"? It's not the Mac, and it's not Mac OS, or Mac OS X, because their "handheld/..." is the iPod, and it's nothing like a Mac. It's not even tied in to the Mac. Apple's platform is, near as I can tell, "whatever they can make money selling". That's not something they can control like Microsoft can control Windows. Microsoft isn't Apple's proxy, but what is?
OS X would be a WHOLE lot more secure with them in place.
Not really, mandatory access control adds a lot of inconvenience and, for most people, the kinds of MAC they're likely to put up with can already be implemented in the existing OS.
They're not using groups to separate responsibility for system preferences.
They're not providing a way to use chroot to create stronger internal sandboxes.
They're bypassing traverse checking in the OS-9 compatible "aliases", and probably Spotlight as well.
They haven't ported jails or secure levels from FreeBSD.
They haven't a consistent emulation of the classic Mac file system semantics on top of foreign file systems, so they probably won't be incorporating any non-HFS+ file systems with tighter security (not just secure levels in FreeBSD, but anything from TrustedBSD or any of the Linux file systems).
It's not so easy, removing features, even if they are unsafe.
I know, that's why my recommendation doesn't remove any user-visible features, and even improves the user experience by removing the perceived need for warning dialogs before doing "unsafe" things, and provides a more versatile and flexible tool for managing the whole process.
It would, as far as the user's concerned, add features. And improve security as well.
Slashdot Mirror
Audio in/out, camera, all Core Duos... hard to imagine they could have done this better.
Included a GPU that doesn't suck dirty swamp water through a clogged garden hose?
Why would I want a laptop to have a great graphics card?
That's the wrong question, because the Intel GPU isn't just "not great", it's "bloody awful".
The right question is "Why would I want a laptop to have a complete OpenGL implementation?".
Because this chip's OpenGL is incomplete, and Mac OS X's graphics is built on top of openGL. So one of the two cores in your Core Duo is going to be sitting there running OpenGL emulation... less efficiently than even an entry-level ATI or nVidia GPU... and generating heat. And the extra RAM you'll have to install, both to support this extra code and to make up for the missing VRAM (that's another thing, this chip doesn't have any), will be generating heat too.
If I have to choose between a cheap and power-savvy low-end card and a high-performance but expensive and power-hungry one, I choose the first.
Unfortunately Apple doesn't give you that choice. You get a cheap chip that makes the computer as a whole more power-hungry. A low-end ATI or nVidia chip would have given you the "cheap and power-savvy" option, but just as in the Mac mini you're stuck with what they provide.
So basically they don't have an acceptable low-end laptop, and they don't have an acceptable low-end desktop. So much for the Intel switch making Macs more affordable.
If it wasn't for OS X I wouldn't consider ANY Apple laptop even vaguely acceptable.
Even the Powerbook/Macbook Pro keyboard feels like crap after you've used a Thinkpad, and of course the single mouse button on the trackpad just adds insult to injury.
On the microkernel side we have Minix 3, 15 years after the first not-really-open-source-but-code-available microkernel UNIX systems.
... what? 15 years after the first not-really-open-source-but-code-available monolithic kernel UNIX systems we had... hmmm... things like MiNT, and bits of BSD, but even Net/1 was a few years in the future and Minix wasn't even out.
On the monolithic kernel side we have
I think, after you allow for the 20 year head start, microkernels aren't doing that badly.
I could actually live with that, if it were true.
What really bugs me about Microsoft is not that they're "the evil empire", but because they're not "a competant evil empire".
The problem is that Apple seems to have a lot of the same areas of profound incompetance as Microsoft - style only carries you so far - and while they avoid a lot of the evil they've got their own home-grown kind that Microsoft only seems to have borrowed part of from them.
Any modern day browser would ask the user if they want to open the file, whether it was a click or a redirect.
.NET. You seem to be talking about a standard browser open.
.NET, or even Firefox's XUL.
... running the "save file" procedure in a hidden "root" subprocess ... every time you save. The problem is that the system made it too easy to routinely change your security domain. Rather than saving the file to the root directory of your system disk, you should have kept the working copy of the file (the one you're editing) in your own directory (eg, on the desktop or in documents), and put it in the write-protected root of the file system (if you had to) only when it was working.
... so ...
.NET, nor active scripts. That mechanism should be managed by the application that is using the control, preferably by installing an extension in the component as Dashboard on the Mac or the KDE desktop do. Any application that is used to display untrusted content (Mail.app, Safari.app, Outlook, Internet Explorer, Firefox, Konqueror) must not provide that extension. Applications that need to execute unsandboxed code (such as Windows Update, the Windows control panel, or Dashboard) should be (and in the case of Dashboard, are) implemented as shells that don't load sources of untrusted content at all.
Just to make clear, though, that has nothing to do with
I'm talking about the integration of the browser with locally executed unsandboxed code, whether that's implemented using ActiveX,
The browser should not ask the user for permission to open a "file://" URL, because it should not be able to do anything "unsafe" just because the document is in the "My Computer" zone. Or any other zone. The browser shouldn't allow a document to do anything "unsafe", period, regardless of where the file is or what type of file the browser thinks it is.
I had to save a Perl file on the root of the hard drive, so it asked me for the admin password. No problem. Later on, I had to do some work on the file and it asked me for the password when I opened it. Understandable.
Now, I'm working on this file, and I save it a few times. EVERY SINGLE TIME I hit Ctrl-S, it asks me for the password. Why?
Because you're entering a new security domain
This dialog isn't really the same kind of thing as the one I'm talking about, though. It's not there to warn the user that they're doing something dangerous, it's to verify to the system that the person who's requesting the action is really the account owner... which is why it asks every time.
The perception that this password dialog has something to do with "keeping you from accidentally doing something dangerous", like the typical approval dialog, is a common and mistaken one. It's got nothing to do with that at all.
The solution isn't to make the dialogs better, or to make the dialogs go away, it's to design the system so that the dialogs aren't necessary, to make it easy and obvious how to work without being interrupted by them. Apple made a mistake here... the editor should have offered to save the file somewhere else, and moving it back into the protected directory should have been a separate process... so it was clear that editing the file in place wasn't a normal thing to do.
Getting back to the Windows example, the privileges granted a document by an application should not be related to the zone the document is in. They should be related to the security domain the application is in and the role within that domain the application plays.
That is, the HTML display component (Webcore in Safari, Gecko in Firefox, KHTML in Konqueror, the HTML control in Internet Explorer) should not have a built-in mechanism to grant local user privileges to a document or a component of a document at all. Not via ActiveX, not
Do this, and you don't need any "I'm about to do something stupid" dialo
I've read that before, and I can't agree with everything Neal says in there (and I think Cryptonomicon missed the boat in a lot of areas), but it's an interesting read.
Microsoft's platform, by extension, consists of products designed for people who are programmers and techno-geeks,
I'm a programmer and a techno-geek, and that puppy ain't designed for me.
Apple ipod works only with iTunes.
Which runs on Windows.
Apple applications only work with Apple OSX.
Except for iTunes and Quicktime, which are the applications that the iPod and the iTunes Music Store (that seem pretty important to this "platform" thing) depend on.
No, there's no "platform" here.
You mean like...
You listed three generations of two operating systems with the same calling conventions, library formats, and process structure. The biggest difference is between Windows CE and Windows 9x, because Windows CE has dropped most of the DOS shell support, so to run a shell in CE you have to emulate it.
Really.
You can run applications for any of these systems under the Windows NT just by providing the right libraries. I've compiled Pocket PC code for x86 instead of ARM and run it on top of Win32... the "Pocket PC emulator" doesn't even seem to keep you from calling Win32 libraries as well as WinCE ones. These are all the same species of OS... to the programmer, they're like different breeds of dog.
There's more differences between Mac OS X/Cocoa and FreeBSD/Gnome than between any of those systems, even with all the FreeBSD code in Mac OS X. Hell, just the differences between Mach-O and ELF are huge. The iPod isn't even the same phylum.
Perhaps it's time for spybot like program for Macs and Linux just to remove tracking cookies?
Perhaps it's time to quit worrying about cookies?
Safari has a similar option to "Accept cookies only from sites you navigate to".
Camino lets you "Accept cookies only from sites you visit".
And advertisers know about this feature, so they use other tricks like correlating your IP address with referrers or using tagged URLs to gather the information they're looking for.
Internet and Local Intranet don't have that kind of access in .NET.
You mean that if a web page does something like "location = 'file:///c|/...'", or gives the user a link to a local file that the user clicks on, it won't load the specified file? Doesn't matter whether the file is visible in the context of the untrusted document, the question is whether the file can be loaded at all.
The dialog boxes are there so users aren't completely annoyed, trying to find where to grant apps access.
I have had the same people come to me multiple times saying "Peter, I did it again, a dialog bax came up and asked me if I wanted to run something, and now my computer's acting strange. Can you help?"
Technically speaking, this is a social engineering attack, but because users are presented with so many warning dialogs they're trained to automatically approve them. Because they need to approve so many of them just to get their work done. So an approval dialog (no matter how worded) doesn't have enough of an effect on security to matter.
Actually, if you know anything about Apple the iPod is exactly like the Mac.
Did you read the article?
It's a different OS (two different OSes) with a different user interface and different applications base. It's not a "platform" in the sense that Windows is a "Platform".
Apple doesn't own "it's a one stop solution". They don't control "it's a one-stop solution". "It's a one stop solution" for Apple is like "the Internet" for Google.
They're not releasing a new consumer version. They've continued to make the embedded version available, and the Pocket PC and Smartphone systems are based on it, but the Windows CE based "laptop replacement" devices are dead dead dead...
So you say that IBM and Apple don't take open source software, and fund their development whilst giving it away for free? Gee, IBM got sued for doing that. Better let the judge know it never happened!
By that logic the "apple platform" doesn't include anything the majority of Apple's actual customers are interested in.
The iPod is the most popular MP3 player even among people who use Windows.
"simplicity" is like "internet", you can't "own simplicity" like you can "own windows".
It's not a platform in the sense that Cringely is using the term.
Two words: T-100.
You mean "Palm m100"?
No, I don't mean a $150 68000 device with 2M running palmOS 3.5 in 2000, when it was the replacement for the Palm IIIe and contemporary with the Visor and the original Clie.
I mean they should have maintained the PalmOS 4 68000 based line and let the price drop and the capacity increase as the cost of memory and chips fell. They didn't have to enhance it and come out with a PalmOS 4.3, 4.4, 4.5, just keep making the black-and-whote DragonBall-EZ based Palms with 8M RAM and selling them as cheaply as they could.
Here's what they did:
2000 - Palm m100 - 2M - $150.
2002 - Palm Zire - 2M - $100.
2004 - Palm Zire 21 - ARM-based, 16M - $100
2006 - Palm Z22 - ARM-based, color - $100
here's what they should have done:
2000 - Palm m100 - 2M - $150.
2002 - Palm Zire - 2M- $100.
2004 - (Palm z100) - 8M - $70.
2006 - (Palm Mini) - 8M - $40.
They got down to $100 and stuck there, adding power that the entry level doesn't need, upgrading the screen and processor and memory instead of pushing the price down and going after the grocery stor checkout lines and the educational market until every high school student had one instead of a Ti-83 or whatever this year's sine-qua-non calculator is.
I think he is refering to the closed format of the Apple music platform.
AAC is just another name for MPEG4 encoding of music. That's an open standard.
The only thing closed is the DRM, but DRM has to be closed, because DRM depends on obfuscation to work. DRM involves giving someone an encrypted message, the decryption keys for the message, and an implementation of the decryption algorithm, and keeping them from reading the message except when you want them to.
To quote Douglas Adams, "This is of course impossible".
Microsoft's DRM is less open - the only platform it runs on is Windows (Plays For Sure devices aren't "platforms" - they only run one application) whereas iTunes runs on Windows and Mac OS. Real's DRM isn't any different from Microsoft's. Sony's DRM is dead.
And it doesn't matter if Apple's DRM is open or closed because the first thing any sane person does with the music they buy from iTunes is burn it to an audio CD... and once you do that you can play it on any device.
If someone can find a way of taking open source software, and finds a business model that allows them to fund their development whilst giving it away for free, it's bye, bye Microsoft, Novell and a few other companies who make their livings from pure software licensing.
You mean like IBM and Apple do?
"browser-integration with .NET" .NET's pretty darn secure.
Is it?
Enlighten me.
Is there a mechanism whereby an object in Internet or even LocalIntranet can request the execution of an arbitrarily specified object in the MyComputer zone? If so, does the user get a dialog asking whether this execution should be permitted, or is it unconditionally denied regardless of the user's settings? Is it possible for a user to specify that a specific object (based on any criteria, whether URL or certificate or address or strong name) be granted this right at the time the request is made?
Unless the answers are "no" or "yes, denied, no" then it's not "pretty darned secure".
For anyone who read the article, the author suggests that Microsoft should license Vista and Office for no more than $50.
No, he suggests that Microsoft sell a new OS that's actually usable on existing computers and doesn't have the legacy bloat and security problems of Windows... for $50.
Maybe they could bring back Windows CE?
Microsoft can build software for a handheld or tablet computer, a mobile phone or a TV set-top box and even though the wrapper is different, the feel is always very much the same -- that of a fat PC client. Microsoft can't allow a phone to be a phone because they can't dominate and control a plain old phone unless it is more Windows than phone. That's a problem.
It surely is. That was obvious in 2000 when they came out with "Pocket PC", their most successful spin on the handheld, and "Stinger", their fialed attempt to get into the cellphone market.
The Pocket PC meant the end of the Windows CE micro-notebooks and the Windows-CE-based tablets. They were pushing Windows NT as the new tablet... the problem is that while Windows CE felt like a spin on Windows 95, and the Pocket PC felt like a Palm on steroids, the Tablet PC was just an overpriced notebook.
Luckily for Microsoft, Palm had no idea what their product was, and has been trying to turn Palm OS into Pocket PC... and failing, big time. If Palm was smart they'd be selling black-and-white 68000-based Palms for $30-$50 in every grocery store in the USA, and they'd still own the business... because Microsoft couldn't do that. But, no...
But, anyway... Microsoft's platform is Windows. If you're not Windows... even if you look like Windows, Microsoft just wants to make you an annex to the Windows desktop. And if you don't even look like Windows, Microsoft doesn't want you to be a platform. That's why they completely redid the XBox, people were turning it into a platform.
But what's Apple's "platform"? It's not the Mac, and it's not Mac OS, or Mac OS X, because their "handheld/..." is the iPod, and it's nothing like a Mac. It's not even tied in to the Mac. Apple's platform is, near as I can tell, "whatever they can make money selling". That's not something they can control like Microsoft can control Windows. Microsoft isn't Apple's proxy, but what is?
OS X would be a WHOLE lot more secure with them in place.
Not really, mandatory access control adds a lot of inconvenience and, for most people, the kinds of MAC they're likely to put up with can already be implemented in the existing OS.
They're not using groups to separate responsibility for system preferences.
They're not providing a way to use chroot to create stronger internal sandboxes.
They're bypassing traverse checking in the OS-9 compatible "aliases", and probably Spotlight as well.
They haven't ported jails or secure levels from FreeBSD.
They haven't a consistent emulation of the classic Mac file system semantics on top of foreign file systems, so they probably won't be incorporating any non-HFS+ file systems with tighter security (not just secure levels in FreeBSD, but anything from TrustedBSD or any of the Linux file systems).
It's not so easy, removing features, even if they are unsafe.
I know, that's why my recommendation doesn't remove any user-visible features, and even improves the user experience by removing the perceived need for warning dialogs before doing "unsafe" things, and provides a more versatile and flexible tool for managing the whole process.
It would, as far as the user's concerned, add features. And improve security as well.