Trying to get back on topic: I'd heartily suggest the MacBook Pro as a linux laptop. I run mine with SuSE 10.1.
The only serious problem is internal speaker support (not soundcard, that works perfectly). As soon as that is resolved the only remaining non-functional hardware will be the webcam (which is touch-n-go on Linux in general, and given that its an USB iSight I expect support sooner or later), and the bluetooth module, which, as it is a Broadcomm module, I don't have my hopes up about.
Everything else, OpenGL (Radeon X1600), 802.11g/a, Suspend, Throttling, SMP, etc. . . works perfectly. Even the screen brightness.
You can run OS X if you ever need to, but given that I use Linux as my primary OS I'm very happy with the MacBook. It doesn't hurt that it's got EFI, which I expect Linux to take advantage of in the future, and it doesn't hurt that the BIOS emulation is an Intel Reference implementation, which makes ACPI on Linux (also based upon Intel's reference implementaiton) stable, easy, and functional.
The following list is components which do NOT work on the MacBook pro on linux, in order of severity:
1. Internal speakers (note that the sound card DOES work; I expect this to be resolved shortly). 2. Bluetooth module. 3. iSight USB.
Sleep/suspend works great. All the ports work great. CPU throttling works great. OpenGL works great.
As soon as the problem with internal speakers output is resolved, I'll heartly recommend the MacBook Pro for all linux users.
Furthermore, I expect distributors to focus on the MacBook Pro. Think about it; Apple hardware gives linux distributors a standards-based platform without the kludgy legacy crap that's easy to aim at. Even the BIOS compatibility module is a reference implementation, which probably explains the ease in getting suspend support and the like to work. Also, Apple hardware is high profile, and with a MacBook Pro triboot you can run every piece of software under the sun on one, high powered, sleek notebook.
So, we have a choice between computers manufactured by an American company in China, or computers manufactured by a Chinese company in China.
Given that the actual production houses in China cannot be majority owned by foreigns, I'd say it doesn't make a single bit of difference.
I'd love to buy "American" made products, but when it comes to electronics thats totally hopeless. From the top to the bottom end, _ALL_ of it is manufactured either in China directly, or Taiwan; and although Taiwan is a democracy and a responsible U.S. ally I suspect it would take about 5 minutes for the PRC to conduct clandestine commercial activties in the RoC.
I have no problem with foreign products. I'd love to purchase Japanese manufactured computers (and I'm a huge supporter of Japanese cars, many of whom which ARE produced on the island). But nearly all electronics manufacturing is done in China, sadly.
Though that worked out well for the primaries of stacker ($600 million settlement, or something like that), the company failed during litigation........
One wonders if Microsoft is trying to "phase out" Symantec in the same fashion. Obviously, they believe it is cheaper to litigate/settle than license or buyout.
1. Get a T-mobile phone with bluetooth and EDGE 2. Get a laptop with bluetooth. 3. Enjoy unlimited wireless internet access with (the good) ~120 kbps real-world throughput and (the bad) ~800-1200 ms roundtrip latency, for $19.99 a month.
I know that EV-DO has better latency, but I didn't think that Cingular's HSDSPA or whatever alphabet soup it is was that much better. T-mobile's EDGE service is acceptable over an NX connection, and works while in the car up to about 60 mph.
For an addition $10.00, you can get the "T-mobile Total Internet" package, which gives you unlimited T-Mobile hotspots, which are all over the place, and significantly faster than any of the 2.5G-3G data services.
*shrug*---- I've been tempted by Verizon's EV-DO service, but at 4x the cost, with availability of the high-speed component in metropolitan areas only (my northshort Chicago suburb, near O'Hare airport, at the world's largest industrial park, is NOT served by EV-DO) just doesn't seem worth it.
Much of the world still lives on dialup. I can get used to using 2x dialup (with 2x the latency, har har) while on the road; and the price cannot be beat (I average 30-50 megs of usage per month, and I get the added side benefit of browsing on either my phone or laptop whenever I want).
Yes, and no. I believe in the U.S. that if they subpeona the contents of your safe, and you don't provide it, they can get you for obstruction of justice. They might have to pursue the case at an oblique angle; they aren't investigating you, they are investigating someone else.
I'd be _very_ interested to see a Supreme Court decision regarding encryption and the 4th amendment. I'm sure this lies in the future for some one in the U.S.
This is not an argument against passing such an Amendment.
These are exactly the sorts of issues Congress is *supposed* to hash out. Not the relative merits of gay marriage, nor the sexual relations of the President.
This is referred to as a "catch-all" type of law. Beware the wonders of selective enforcement.
The idea here is that if you find a suspected terrorist, and they use encryption, you don't even need to bust them for terrorism OR for not providing their encryption keys when demanded. You can just go to step A, look up their name in the government encryption key database, find out that no, they did not provide their encryption key to , and take them directly to jail.
Regardless of whether or not the are a terrorist, regardless of whether or not they are willing to turn over their encryption keys when asked, you can find them guilty.
This is not about collecting everyone's encryption keys (at least not at first). Initially, this will be used as a blunt stick to smack anyone the government doesn't like. Think of the way seat belt laws are enforced; cops won't stop you for not wearing your seat belt, but they'll sure as hell issue a ticket for it even if you aren't speed, have all your paperwork in order, and have done nothing else wrong. It's a sort of standby crime they can get you on.
Err... how about ZDnet running OS X on a Toshiba Laptop?
Does that count?
Apple has not, to date, released the XNU sources from x86. My guess is less because of piracy issues, and more because the sources are, "Really Fucking Ugly"(TM).
As time has gone on, more and more of their x86 toolchain has been released. I suspect at some point we'll see 99% of the stuff released, except some proprietary modules a) necessary for booting the system, and b) utilizing some on-motherboard TPM capabitilies to prevent booting on non-Apple systems.
1. XServe is still PPC, so XServe's kernel (XNU PPC) is still avaliable. Don't put the cart before the horse, which brings me to point 2: 2. The PPC->Intel transition for this class of performance computing (high-end science) is going to be a far, far bigger stumbling block than inability to modify the kernel. My guess is that migrating an Xserve cluster from PPC to Intel is going to be a huge pain, at least in terms of getting optimal performance. This isn't a Rosetta situation;-) 3. At worst, a closed source kernel for the intel Xserve renders it only slightly (rather than significantly) more suitable to high-end clustering than Windows. Both are trounced by custom kernels delivered on other solutions, like Solaris, Linux, and AIX.
No email administrator who had worked with the main platforms out there would, in his right mind, choose lotes....
I don't know ANYTHING about administering e-mail services, but....
I know this not to be true. Notes is not the dominate platform out there, but it still has substantial market share, and there are new deployements on a constant basis.
This is from a business perspective, not a technical perspective.
ODF (OpenDoucment Foundation) has already beaten them to the punch.
ODF wisely realized that waiting for MS to provide support would cripple the format. I quote: The plug-in is designed to allow Microsoft Office users open and save files in the OpenDocument, a format supported in other productivity suites but not current versions of Office or Office 2007.
As to who helped the development of the plug-in, Edwards wouldn't say except to joke that "people who use numbers in their names" offered some unsolicited suggestions.
The technical break-through came when developers stopped following Microsoft's interoperability instructions and started considering undocumented APIs, he said.
"Your big mistake is following instructions Microsoft provides. You need to be looking at what Microsoft does, not what they tell you to do," Edwards said.
Same as notes, Real Player's got a bad rap. Of course, the poor implementation has been Real's fault.
For the most part, the new, non-spyware infested real player is a very solid player. I've taken to using it across all clients I can, under the logic that a)powerusers like me will appreciate its new features and dumping all the crap from yesteryear, b) newbies have enough name recognition of the real "name" that they'll continue to use it (these are the people who never noticed the spyware!) and c) normal users will complain once or twice, but then they'll see me using it, and they'll see the integration throughout their client, and tend to be happy using it.
*shrug*. OS X people like to have it pointed out that real player gets them full screen video for free, and that real player is currently the only real option for cross-platform video streaming (forgive the pun). Why do I say that? Although I _love_ quicktime, and deploy it on my on sites, when I'm browsing "the web", I'm generally restricted to Windows Media and Real Player as my options.
As such, I choose Real.
Linux users should appreciate that the helix player framework is open source. Certainly more open source than the quicktime or windows media player frameworks.
My current home linux desktop has not been reformated since inception 4 years ago (including numerous hardware upgrades; two motherboards, three processors).
My "rebuilding" isn't wasteful at all. And I find it to be pretty fun. I've tinkered around with my PowerMac G5, too, but the amount of work done on my Linux Desktops and Linux Servers borders on inane.
Throughout all of this, my personal Linux Desktop has never needed a reformat/reinstall. I'll strip the thing down to components, rebuild it, and then assuming that I put the harddrive on the right IDE port, and have my bios pointing to the right bootup drive, I'll be back at a working desktop upon next boot.
That's rebuilding with a smile.
Honestly, while I know (and love) the type of rebuilding your talking about, many of the Windows users I know refer to the term "rebuilding" as reformat/reinstall. They "rebuild" their desktop environment, possibly after changing some hardware.
Furthermore, your experience of never have to de-virus or de-spyware your system is atypical, at best. I feel that I'm quite competant. I write my own software, I can ticker with aspects of my OSS system, I've been running Linux on a constant basis for 3 years, and I'd used Windows on a constant basis between 3.1 and XP (pre-SP1, when I ditched it). I'm well aware of most computing security concerns, and I'd still have stuff get in; perhaps through the 10 seconds during reboot that the network connection comes up before the firewall, or perhaps through misconfigured UPnP; I dunno.
But my experience, my uncle & his familes experience, the experience of a wide variety of offices that my company does business with, the experience of my relatives/friends who work at allstate (largest employeer in my area) who lost their office network for 3 days because of Blaster, the experience of University of Chicago staff (whom I am friends with) who were not able to work for a week as a result of code red, and even the experience of those in the (2002?) Northeast power outage potentially caused by office computers overloading a safe-override network (and making it "seem" that the control systems had gone offline) due to Blaster overriden computers.... well...
Your experience is atypical.
And this isn't even discussing the variety of driver bugs that can corrupt a system. I've seen a Windows system go down hard from an overheating ATI card, and whatever the hell the driver did before it crashed botched up the NT bootsector, and it bluescreens on boot.
Microsoft's fault? Probably not; blame ATI. But this doesn't happen on my Linux boxes with ATI cards, even though the ATI FGLRX Linux drivers are absolutely, shockingly bad.
As a Linux user, I LOVE to tinker with my hardware. Don't mistake me for some MySpace Mac Ninny; I'm not. I love tinkering with my hardware, I get a kick out of it; I pour sweat/blood (damn sharp edges) into my systems. But when it comes time to boot, I don't mess with my software; reinstall the motherboard, plug it in, and I'll go back to gaming/surfing/working/whatever.
He was talking about relatives in the context of Windows 2000.
Seems more than a little ridiculous for the VP of Security at Microsoft to insist that members of his bloodline replace their still-supported-MS-OS for causal security help.
Doubtlessly he'll say the same thing when Vista comes out.
I only know about the package management tools and they download stuff from the internet and install it automatically. Is it possible to just have it download the files and I can then burn them to cd and install them on the offline system easily? how easily.
Well, both SMART and YaST fundamnetally rely upon RPM. Both can install RPMs offline, from either DVDs, directories on your harddrive, or by supplying a file name. If you download the file from an ftp or http site, you can install it.
All of the updates, and all of the repositories' files are avaliable in RPM format. But yes, I'll admit that it is trickier to download/burn to disk/install on Linux than just using the automated download/install.
Slashdot Mirror
Trying to get back on topic: I'd heartily suggest the MacBook Pro as a linux laptop. I run mine with SuSE 10.1.
The only serious problem is internal speaker support (not soundcard, that works perfectly). As soon as that is resolved the only remaining non-functional hardware will be the webcam (which is touch-n-go on Linux in general, and given that its an USB iSight I expect support sooner or later), and the bluetooth module, which, as it is a Broadcomm module, I don't have my hopes up about.
Everything else, OpenGL (Radeon X1600), 802.11g/a, Suspend, Throttling, SMP, etc. . . works perfectly. Even the screen brightness.
You can run OS X if you ever need to, but given that I use Linux as my primary OS I'm very happy with the MacBook. It doesn't hurt that it's got EFI, which I expect Linux to take advantage of in the future, and it doesn't hurt that the BIOS emulation is an Intel Reference implementation, which makes ACPI on Linux (also based upon Intel's reference implementaiton) stable, easy, and functional.
The following list is components which do NOT work on the MacBook pro on linux, in order of severity:
1. Internal speakers (note that the sound card DOES work; I expect this to be resolved shortly).
2. Bluetooth module.
3. iSight USB.
Sleep/suspend works great. All the ports work great. CPU throttling works great. OpenGL works great.
As soon as the problem with internal speakers output is resolved, I'll heartly recommend the MacBook Pro for all linux users.
Furthermore, I expect distributors to focus on the MacBook Pro. Think about it; Apple hardware gives linux distributors a standards-based platform without the kludgy legacy crap that's easy to aim at. Even the BIOS compatibility module is a reference implementation, which probably explains the ease in getting suspend support and the like to work. Also, Apple hardware is high profile, and with a MacBook Pro triboot you can run every piece of software under the sun on one, high powered, sleek notebook.
*shrug* I dig it, anyways.
So, we have a choice between computers manufactured by an American company in China, or computers manufactured by a Chinese company in China.
Given that the actual production houses in China cannot be majority owned by foreigns, I'd say it doesn't make a single bit of difference.
I'd love to buy "American" made products, but when it comes to electronics thats totally hopeless. From the top to the bottom end, _ALL_ of it is manufactured either in China directly, or Taiwan; and although Taiwan is a democracy and a responsible U.S. ally I suspect it would take about 5 minutes for the PRC to conduct clandestine commercial activties in the RoC.
I have no problem with foreign products. I'd love to purchase Japanese manufactured computers (and I'm a huge supporter of Japanese cars, many of whom which ARE produced on the island). But nearly all electronics manufacturing is done in China, sadly.
I've had several POTS outages (DSL too, of course) in an SBC area of S. Milwaukee, Oak Creek, WI.
They claim its construction problems during upgrades, or just cut cables.
*shrug*
I have not seen this at T-mobile, and it seems to be gaining new subscribers at a rate of around 10% per year.
Not with T-mobile's $19.99 unlimited Tethering plan. EDGE speeds nationwide, too.
Upgrade to an EDGE phone on T-Mobile, and get 3x your current download speed.
Sure, 3x of 40 kbps isn't much, but 120 kbps IS a noticeable improvement. Sadly, it doesn't help much with the latency.
Maybe they prefer the slow death of cooperate for awhile then be crushed to the quick death of defy Microsoft and go out of business tomorrow? :P
Reminds me of stacker.
Though that worked out well for the primaries of stacker ($600 million settlement, or something like that), the company failed during litigation........
One wonders if Microsoft is trying to "phase out" Symantec in the same fashion. Obviously, they believe it is cheaper to litigate/settle than license or buyout.
What's wrong with T-mobile's EDGE service?
1. Get a T-mobile phone with bluetooth and EDGE
2. Get a laptop with bluetooth.
3. Enjoy unlimited wireless internet access with (the good) ~120 kbps real-world throughput and (the bad) ~800-1200 ms roundtrip latency, for $19.99 a month.
I know that EV-DO has better latency, but I didn't think that Cingular's HSDSPA or whatever alphabet soup it is was that much better. T-mobile's EDGE service is acceptable over an NX connection, and works while in the car up to about 60 mph.
For an addition $10.00, you can get the "T-mobile Total Internet" package, which gives you unlimited T-Mobile hotspots, which are all over the place, and significantly faster than any of the 2.5G-3G data services.
*shrug*---- I've been tempted by Verizon's EV-DO service, but at 4x the cost, with availability of the high-speed component in metropolitan areas only (my northshort Chicago suburb, near O'Hare airport, at the world's largest industrial park, is NOT served by EV-DO) just doesn't seem worth it.
Much of the world still lives on dialup. I can get used to using 2x dialup (with 2x the latency, har har) while on the road; and the price cannot be beat (I average 30-50 megs of usage per month, and I get the added side benefit of browsing on either my phone or laptop whenever I want).
Go T-Mobile. I highly recommend it.
Yes, and no. I believe in the U.S. that if they subpeona the contents of your safe, and you don't provide it, they can get you for obstruction of justice. They might have to pursue the case at an oblique angle; they aren't investigating you, they are investigating someone else.
I'd be _very_ interested to see a Supreme Court decision regarding encryption and the 4th amendment. I'm sure this lies in the future for some one in the U.S.
This is not an argument against passing such an Amendment.
These are exactly the sorts of issues Congress is *supposed* to hash out. Not the relative merits of gay marriage, nor the sexual relations of the President.
Sorry, yes. That's what I meant. Given that I was replying to an article title including "Mac OS X Kernel", that much should be obvious ;-)
Pardon my lack of specificity
People; don't say "This can't be done."
This is referred to as a "catch-all" type of law. Beware the wonders of selective enforcement.
The idea here is that if you find a suspected terrorist, and they use encryption, you don't even need to bust them for terrorism OR for not providing their encryption keys when demanded. You can just go to step A, look up their name in the government encryption key database, find out that no, they did not provide their encryption key to , and take them directly to jail.
Regardless of whether or not the are a terrorist, regardless of whether or not they are willing to turn over their encryption keys when asked, you can find them guilty.
This is not about collecting everyone's encryption keys (at least not at first). Initially, this will be used as a blunt stick to smack anyone the government doesn't like. Think of the way seat belt laws are enforced; cops won't stop you for not wearing your seat belt, but they'll sure as hell issue a ticket for it even if you aren't speed, have all your paperwork in order, and have done nothing else wrong. It's a sort of standby crime they can get you on.
http://reviews.zdnet.co.uk/software/os/0,39024180, 39235916,00.htm
Forgot the link.
Err... how about ZDnet running OS X on a Toshiba Laptop?
Does that count?
Apple has not, to date, released the XNU sources from x86. My guess is less because of piracy issues, and more because the sources are, "Really Fucking Ugly"(TM).
As time has gone on, more and more of their x86 toolchain has been released. I suspect at some point we'll see 99% of the stuff released, except some proprietary modules a) necessary for booting the system, and b) utilizing some on-motherboard TPM capabitilies to prevent booting on non-Apple systems.
1. XServe is still PPC, so XServe's kernel (XNU PPC) is still avaliable. Don't put the cart before the horse, which brings me to point 2: ;-)
2. The PPC->Intel transition for this class of performance computing (high-end science) is going to be a far, far bigger stumbling block than inability to modify the kernel. My guess is that migrating an Xserve cluster from PPC to Intel is going to be a huge pain, at least in terms of getting optimal performance. This isn't a Rosetta situation
3. At worst, a closed source kernel for the intel Xserve renders it only slightly (rather than significantly) more suitable to high-end clustering than Windows. Both are trounced by custom kernels delivered on other solutions, like Solaris, Linux, and AIX.
Darwin never contained GPL code.
Darwin was BSD. This is the difference between GPL and BSD.
I'm sad about this, but I continue to use Apple, because if I'm forced into using a proprietary OS (OS X or Windows), I'll choose OS X.
Linux is still my preferred OS of choice.
No email administrator who had worked with the main platforms out there would, in his right mind, choose lotes....
I don't know ANYTHING about administering e-mail services, but....
I know this not to be true. Notes is not the dominate platform out there, but it still has substantial market share, and there are new deployements on a constant basis.
This is from a business perspective, not a technical perspective.
ODF (OpenDoucment Foundation) has already beaten them to the punch.
ODF wisely realized that waiting for MS to provide support would cripple the format. I quote:
The plug-in is designed to allow Microsoft Office users open and save files in the OpenDocument, a format supported in other productivity suites but not current versions of Office or Office 2007.
As to who helped the development of the plug-in, Edwards wouldn't say except to joke that "people who use numbers in their names" offered some unsolicited suggestions.
The technical break-through came when developers stopped following Microsoft's interoperability instructions and started considering undocumented APIs, he said.
"Your big mistake is following instructions Microsoft provides. You need to be looking at what Microsoft does, not what they tell you to do," Edwards said.
Same as notes, Real Player's got a bad rap. Of course, the poor implementation has been Real's fault.
For the most part, the new, non-spyware infested real player is a very solid player. I've taken to using it across all clients I can, under the logic that a)powerusers like me will appreciate its new features and dumping all the crap from yesteryear, b) newbies have enough name recognition of the real "name" that they'll continue to use it (these are the people who never noticed the spyware!) and c) normal users will complain once or twice, but then they'll see me using it, and they'll see the integration throughout their client, and tend to be happy using it.
*shrug*. OS X people like to have it pointed out that real player gets them full screen video for free, and that real player is currently the only real option for cross-platform video streaming (forgive the pun). Why do I say that? Although I _love_ quicktime, and deploy it on my on sites, when I'm browsing "the web", I'm generally restricted to Windows Media and Real Player as my options.
As such, I choose Real.
Linux users should appreciate that the helix player framework is open source. Certainly more open source than the quicktime or windows media player frameworks.
Err... No, not really.
Please note what I said:
My current home linux desktop has not been reformated since inception 4 years ago (including numerous hardware upgrades; two motherboards, three processors).
My "rebuilding" isn't wasteful at all. And I find it to be pretty fun. I've tinkered around with my PowerMac G5, too, but the amount of work done on my Linux Desktops and Linux Servers borders on inane.
Throughout all of this, my personal Linux Desktop has never needed a reformat/reinstall. I'll strip the thing down to components, rebuild it, and then assuming that I put the harddrive on the right IDE port, and have my bios pointing to the right bootup drive, I'll be back at a working desktop upon next boot.
That's rebuilding with a smile.
Honestly, while I know (and love) the type of rebuilding your talking about, many of the Windows users I know refer to the term "rebuilding" as reformat/reinstall. They "rebuild" their desktop environment, possibly after changing some hardware.
Furthermore, your experience of never have to de-virus or de-spyware your system is atypical, at best. I feel that I'm quite competant. I write my own software, I can ticker with aspects of my OSS system, I've been running Linux on a constant basis for 3 years, and I'd used Windows on a constant basis between 3.1 and XP (pre-SP1, when I ditched it). I'm well aware of most computing security concerns, and I'd still have stuff get in; perhaps through the 10 seconds during reboot that the network connection comes up before the firewall, or perhaps through misconfigured UPnP; I dunno.
But my experience, my uncle & his familes experience, the experience of a wide variety of offices that my company does business with, the experience of my relatives/friends who work at allstate (largest employeer in my area) who lost their office network for 3 days because of Blaster, the experience of University of Chicago staff (whom I am friends with) who were not able to work for a week as a result of code red, and even the experience of those in the (2002?) Northeast power outage potentially caused by office computers overloading a safe-override network (and making it "seem" that the control systems had gone offline) due to Blaster overriden computers.... well...
Your experience is atypical.
And this isn't even discussing the variety of driver bugs that can corrupt a system. I've seen a Windows system go down hard from an overheating ATI card, and whatever the hell the driver did before it crashed botched up the NT bootsector, and it bluescreens on boot.
Microsoft's fault? Probably not; blame ATI. But this doesn't happen on my Linux boxes with ATI cards, even though the ATI FGLRX Linux drivers are absolutely, shockingly bad.
As a Linux user, I LOVE to tinker with my hardware. Don't mistake me for some MySpace Mac Ninny; I'm not. I love tinkering with my hardware, I get a kick out of it; I pour sweat/blood (damn sharp edges) into my systems. But when it comes time to boot, I don't mess with my software; reinstall the motherboard, plug it in, and I'll go back to gaming/surfing/working/whatever.
Microsoft wouldn't do the above unless they were going to profit from it or damage the competition.
I think you meant to say,
"Microsoft doesn't do anything unless they are going to proift from it or damage the competition."
He was talking about relatives in the context of Windows 2000.
Seems more than a little ridiculous for the VP of Security at Microsoft to insist that members of his bloodline replace their still-supported-MS-OS for causal security help.
Doubtlessly he'll say the same thing when Vista comes out.
IIRC, it was a slashdot interview.
I only know about the package management tools and they download stuff from the internet and install it automatically. Is it possible to just have it download the files and I can then burn them to cd and install them on the offline system easily? how easily.
Well, both SMART and YaST fundamnetally rely upon RPM. Both can install RPMs offline, from either DVDs, directories on your harddrive, or by supplying a file name. If you download the file from an ftp or http site, you can install it.
All of the updates, and all of the repositories' files are avaliable in RPM format. But yes, I'll admit that it is trickier to download/burn to disk/install on Linux than just using the automated download/install.