I have a MacBook Pro r3.1 and can walk into an Apple store and get support at the Genius Bar for it. I have Snow Leopard 10.6 on it, along with Ubuntu 12.04, and I can upgrade the OS to Mountain Lion if I wanted.
Which you might want, if you find that the answer you get at the Genius Bar is "OK, that's a bug we fixed in {10.7.x, 10.8.x}, that's about all the help we can offer". Yes, the Genius Bar may listen sympathetically to your problem report, but the advice they offer might be "you'll have to get a newer OS if you want that problem fixed".
And, no, this isn't different from some other OSes, and I'm quite aware of that, so "butbutbut Ubuntu!" or "butbutbut Windows!" is not a sensible riposte. However, if you want security updates for your OS version for a long period of time, you might get more of that with Windows, as Microsoft are offering "Mainstream Support" for Windows 8 until 2018, or an Ubuntu LTS release, as Canonical has April 2017 as the end-of-life date for 12.04.2 LTS. Ubuntu "STS", not so much, as the end-of-life date for 12.10 is April 2014 and the claim is that the end-of-life date for 13.04 will be October 2014. If Apple continue their every-year release strategy and their "software updates for the current release, security updates for the previous release" policy, you'll get security updates for two years for each release, which is more like the non-LTS versions of Ubuntu than the LTS releases or Windows.)
(Oh, and I'm a Mac user, and have used OS X to work on stuff such as this. Note, BTW, that, after those slides came out, we went with running a single automountd in a global session, but mounts are done by subprocesses that are, in various release-dependent ways, attached to the session that provoked the mount.)
I'm pretty sure you could take any osx variant and install it on any computer in the apple store.
Good luck installing a PowerPC-only OS X version (i.e., anything prior to 10.4, or any of the pre-x86 versions of 10.4) on any computer you can buy now from Apple (none of which have PowerPC processors). Even older x86 versions might not have support for all the hardware available on newer machines.
Microsoft supported Windows 2000 up until 2010. A 10 year support life cycle. The LONGEST Apple has EVER officially supported a release of OS-X was 4 years, and it's generally 3 or less (n-2).
So I'm about to loose support for my Mac running 10.6? Why would Apple do that when they still offer 10.02 downloads? 10.2 was replaced with 10.3 almost 10 years ago. There goes your "4 years".
"Downloads" != "support". If, on a 10.2 machine, you fire up Software Update, as far as I know it'll inform you that there are no updates for your OS. If there's a security issue or serious bug in 10.2, you're stuck with it (unless it's in the open-source part and you can fix it yourself).
Period. More than Windows in fact. Horrible UI. Can't get my head on what makes it so appealing to people? The 'Apple' logo on the side of their machine?
The fact that "horrible UI" is a matter of personal preference, and one person's "horrible UI" is another person's "great UI"?
File system support - certain filesystems on external USB drives cannot be written to by default - I've had to use third party - buggy drivers to enable this again - this is an area where I would expect OSX to "Just Work(tm)" it doesnt very well in this case.
That probably has nothing to do with "external" (or "USB"), except that external drives are more likely to have "non-native" file systems on it, and everything to do with the file system on the drive. You probably won't find plugging an external drive with an ext2 file system on it to Just Work on either OS X or Windows without third-party software, and, if you can put ReFS on an external drive, you probably won't find it Just Working on Linux until it gets reverse-engineered and reimplemented.
Sorry for the A/C...give him a few days with Xcode, and he will be begging for Linux again.
Or just go to the command line and do development from there. Pretty much the only time I fired up Xcode was at Apple when I needed to tweak an xcproj file and the tweak involved adding or removing files from a project (because I'm too lazy to go generate UUIDs from the command line, for the add case, and manually edit all the dependencies and crap); the rest of the time it was Good Old MicroEMACS and the in-house script wrapped around xcodebuild, and other command line tools such as gdb (as per the other reply to your comment).
Once winDOS has finally died its much deserved death
"WinDOS" either on August 24, 2001 or on July 11, 2006", depending on whether "death" refers to "a version of Windows not using the code base derived from the old DOS-based Windows became available for general consumers" or "support for the last version of Windows using said code base ended". It's all NT-based now (except for WinCE, although Windows Phone 8 is apparently NT-based as well).
Unix(R) is a trademark, licensed to anyone willing to pay. Even to something like OS X, which is closer to being a huge binary blob like Windows, than it is to being small tools each doing only one thing.
Yes, the GUI may be "a huge binary blob" (by which you presumably mean "a combination of large binary shared libraries and executables"; neither Windows nor OS X have One Giant Binary File With The Entire GUI), but, well, presumably once you put KDE or GNOME on a Linux distribution, it's not *nix any more, right?
Even more interesting is that the kernel for OSX is FreeBSD - you know, UNIX.
Its not. Its based on Mach. It has some FreeBSD and NetBSD parts, but the kernel is not from FreeBSD. Check http://en.wikipedia.org/wiki/OS_X
("fs" is a script that uses find to find source files, based on extensions.)
$ cd ~/src/macosx/xnu-2050.7.9 $ for i in * > do > if [ -d $i ] > then > echo -n "$i:" > fs $i | xargs wc -l | tail -1 > fi > done EXTERNAL_HEADERS: 11083 total SETUP: 8465 total bsd: 749445 total config: 359 total iokit: 79156 total libkern: 77749 total libsa: 1061 total libsyscall: 9129 total osfmk: 267589 total pexpert: 4526 total security: 18363 total tools: 46325 total
A bit more than "some FreeBSD and NetBSD parts"; the bsd directory has more "lines" of source code (counted crudely as "lines of text in the source file", so it includes comments) than any other subdirectory of the XNU source - more, in particular, than osfmk (i.e., the Mach kernel part).
I guess it's "based on Mach" in the sense that the Mach code doesn't use some other code to make it work while the BSD code relies on the Mach code to provide functions such as the VM system and the lower levels of the process and thread model (BSD processes are implemented atop Mach tasks, and pthreads are implemented atop Mach threads), but it's not just Mach with a little bit of BSD sprinkled in.
Yes I do like Linux and Unix, and I think it is great for Servers and Embedded systems, where it does a few things and does it well. However for the desktop we need to do a lot of crazy things all the time. That is where Macs and Windows excels.
Yup, Unix boxes don't work very well as desktop machines; that's why you're better off running OS X rather than some Unix.
Cisc instructions still have a use where on risc you have to everything yourself
...if you're a compiler writer or an assembler-language programmer. Otherwise you don't have to care that there isn't a POLY instruction or a string-compare instruction or whatever.
Don't even bother. There's a whole contingent of "but it's RISC under the hood" folks around here who don't understand that a single accumulator architecture that has gems like "REPNE SCASB" in its instruction set will never be RISC.
Presumably "accumulator" means something other than "the only register that can participate in load, store, and arithmetic instructions" here (that being what it meant on some older machines, with the exception of multiply and divide instructions, and possibly other arithmetic instructions also involving the MQ register).
mille, as long as Hasbro doesn't kill it (I presume that's why there's no "bornes" in the name). At one point, it was one of my favorite toys for "compiling!"xmille (and the accompanying README) if you want a GUI.
We're not talking about a "time left" bar, we're talking about a progress bar
Actually, the person who submitted the story was talking about a progress bar accompanied by a "time left" message:
How come after 25 years in the tech industry, someone hasn't worked out how to make accurate progress bars? This migration I'm doing has sat on 'less than a minute' for over 30 minutes.
From the article: "Seems to violate law of conservation of momentum". - Yup it does.
And from the article, in more detail: "It seems to violate of the law of conservation of momentum, implied by Newton, which says that no closed system can have a net thrust. However, Shawyer says net thrust occurs because the microwaves have a group velocity which is greater in one direction than the other and Einstein's relativity comes into play."
Is the article and/or Shawyer trying to say here that "Einstein's relativity" magically makes the law of conservation of momentum go away (perhaps the idea is that Einstein's laws replace Newton's laws, or something such as that)? I may be misremembering my physics from ages ago, but I though conservation of momentum was just as much a law in Einsteinian mechanics as is Newtonian mechanics.
TCP/IP was quickly created by some students and released into the world, and then improved on.
Actually, the specs were created by a bunch of people including at leastthreePh. D.'s, and there were several independent implementations, including one done at Bolt Beranek and Newman for UNIX; the latter one was subsequently modified by some students at Berkeley (I have the impression that, in addition to Bill Joy, Sam Leffler worked on the 4.2BSD TCP/IP stack) and offered as part of BSD UNIX, and the rest was history.
ISO OSI covers TCP/IP in that it is a logical break down of the various aspects of the stack. The fact that most implementations shortcut aspects of it for performance does not mean it is invald. It is a useful abstraction even if only in theory.
There's the OSI Reference Model, as described by ISO/IEC 7498-1 (available as ITU-T Recommendation X.200), and there's the OSI protocol stack, with protocols such as the Connectionless Network Protocol ISO/IEC 8473-1 (available as X.233), the Connection-Oriented Transfer Protocol ISO/IEC 8073 (available as X.224), and the Connectionless Transport Protocol ISO/IEC 8602 (available as X.234).
The "ISO OSI" that "covers TCP/IP" is the Reference Model. The person to whom you're responding is talking about the OSI protocol stack.
Slashdot Mirror
Or compile gcc with gcc?
Not if Ken Thompson contributed to gcc, unless you've taken sufficient care.
I have a MacBook Pro r3.1 and can walk into an Apple store and get support at the Genius Bar for it. I have Snow Leopard 10.6 on it, along with Ubuntu 12.04, and I can upgrade the OS to Mountain Lion if I wanted.
Which you might want, if you find that the answer you get at the Genius Bar is "OK, that's a bug we fixed in {10.7.x, 10.8.x}, that's about all the help we can offer". Yes, the Genius Bar may listen sympathetically to your problem report, but the advice they offer might be "you'll have to get a newer OS if you want that problem fixed".
And, no, this isn't different from some other OSes, and I'm quite aware of that, so "butbutbut Ubuntu!" or "butbutbut Windows!" is not a sensible riposte. However, if you want security updates for your OS version for a long period of time, you might get more of that with Windows, as Microsoft are offering "Mainstream Support" for Windows 8 until 2018, or an Ubuntu LTS release, as Canonical has April 2017 as the end-of-life date for 12.04.2 LTS. Ubuntu "STS", not so much, as the end-of-life date for 12.10 is April 2014 and the claim is that the end-of-life date for 13.04 will be October 2014. If Apple continue their every-year release strategy and their "software updates for the current release, security updates for the previous release" policy, you'll get security updates for two years for each release, which is more like the non-LTS versions of Ubuntu than the LTS releases or Windows.)
(Oh, and I'm a Mac user, and have used OS X to work on stuff such as this. Note, BTW, that, after those slides came out, we went with running a single automountd in a global session, but mounts are done by subprocesses that are, in various release-dependent ways, attached to the session that provoked the mount.)
umm, citation? I have never heard this before. I'm pretty sure you could install any osx variant on any computer in the apple store.
Same incorrect claim, same response.
I'm pretty sure you could take any osx variant and install it on any computer in the apple store.
Good luck installing a PowerPC-only OS X version (i.e., anything prior to 10.4, or any of the pre-x86 versions of 10.4) on any computer you can buy now from Apple (none of which have PowerPC processors). Even older x86 versions might not have support for all the hardware available on newer machines.
away from you
Apple still offers PPC Mac and 10.2 downloads.
But they don't offer any fixes for them, not even security fixes.
Microsoft supported Windows 2000 up until 2010. A 10 year support life cycle. The LONGEST Apple has EVER officially supported a release of OS-X was 4 years, and it's generally 3 or less (n-2).
So I'm about to loose support for my Mac running 10.6? Why would Apple do that when they still offer 10.02 downloads? 10.2 was replaced with 10.3 almost 10 years ago. There goes your "4 years".
"Downloads" != "support". If, on a 10.2 machine, you fire up Software Update, as far as I know it'll inform you that there are no updates for your OS. If there's a security issue or serious bug in 10.2, you're stuck with it (unless it's in the open-source part and you can fix it yourself).
The checks and balances were broken with the 17th amendment. That was the first chair to be removed from the game.
So leaving the election of Senators up to politicians rather than to the electorate was a check on the power of the electorate, presumably.
The nice thing about Unix (in my case Linux/BSD) is that they are just as much a workflow toolkit as they are an OS.
Yup, that's why I'm running a commercial Unix, with some lower-level components open-sourced, on my laptop.
Period. More than Windows in fact. Horrible UI. Can't get my head on what makes it so appealing to people? The 'Apple' logo on the side of their machine?
The fact that "horrible UI" is a matter of personal preference, and one person's "horrible UI" is another person's "great UI"?
File system support - certain filesystems on external USB drives cannot be written to by default - I've had to use third party - buggy drivers to enable this again - this is an area where I would expect OSX to "Just Work(tm)" it doesnt very well in this case.
That probably has nothing to do with "external" (or "USB"), except that external drives are more likely to have "non-native" file systems on it, and everything to do with the file system on the drive. You probably won't find plugging an external drive with an ext2 file system on it to Just Work on either OS X or Windows without third-party software, and, if you can put ReFS on an external drive, you probably won't find it Just Working on Linux until it gets reverse-engineered and reimplemented.
Sorry for the A/C...give him a few days with Xcode, and he will be begging for Linux again.
Or just go to the command line and do development from there. Pretty much the only time I fired up Xcode was at Apple when I needed to tweak an xcproj file and the tweak involved adding or removing files from a project (because I'm too lazy to go generate UUIDs from the command line, for the add case, and manually edit all the dependencies and crap); the rest of the time it was Good Old MicroEMACS and the in-house script wrapped around xcodebuild, and other command line tools such as gdb (as per the other reply to your comment).
Once winDOS has finally died its much deserved death
"WinDOS" either on August 24, 2001 or on July 11, 2006", depending on whether "death" refers to "a version of Windows not using the code base derived from the old DOS-based Windows became available for general consumers" or "support for the last version of Windows using said code base ended". It's all NT-based now (except for WinCE, although Windows Phone 8 is apparently NT-based as well).
No, Linux is *nix. OS X is Unix(R).
*nix is a term used for things that works like Unix. You know, init
Oh, well, so much for several Linux distributions.
shell scripts
easily combining small tools that each do one thing well.
See previous examples, another comment of mine for this article, and several of the scripts in question.
Unix(R) is a trademark, licensed to anyone willing to pay. Even to something like OS X, which is closer to being a huge binary blob like Windows, than it is to being small tools each doing only one thing.
Yes, the GUI may be "a huge binary blob" (by which you presumably mean "a combination of large binary shared libraries and executables"; neither Windows nor OS X have One Giant Binary File With The Entire GUI), but, well, presumably once you put KDE or GNOME on a Linux distribution, it's not *nix any more, right?
Even more interesting is that the kernel for OSX is FreeBSD - you know, UNIX.
Its not. Its based on Mach. It has some FreeBSD and NetBSD parts, but the kernel is not from FreeBSD. Check http://en.wikipedia.org/wiki/OS_X
("fs" is a script that uses find to find source files, based on extensions.)
A bit more than "some FreeBSD and NetBSD parts"; the bsd directory has more "lines" of source code (counted crudely as "lines of text in the source file", so it includes comments) than any other subdirectory of the XNU source - more, in particular, than osfmk (i.e., the Mach kernel part).
I guess it's "based on Mach" in the sense that the Mach code doesn't use some other code to make it work while the BSD code relies on the Mach code to provide functions such as the VM system and the lower levels of the process and thread model (BSD processes are implemented atop Mach tasks, and pthreads are implemented atop Mach threads), but it's not just Mach with a little bit of BSD sprinkled in.
Yes I do like Linux and Unix, and I think it is great for Servers and Embedded systems, where it does a few things and does it well. However for the desktop we need to do a lot of crazy things all the time. That is where Macs and Windows excels.
Yup, Unix boxes don't work very well as desktop machines; that's why you're better off running OS X rather than some Unix.
Cisc instructions still have a use where on risc you have to everything yourself
...if you're a compiler writer or an assembler-language programmer. Otherwise you don't have to care that there isn't a POLY instruction or a string-compare instruction or whatever.
Don't even bother. There's a whole contingent of "but it's RISC under the hood" folks around here who don't understand that a single accumulator architecture that has gems like "REPNE SCASB" in its instruction set will never be RISC.
Presumably "accumulator" means something other than "the only register that can participate in load, store, and arithmetic instructions" here (that being what it meant on some older machines, with the exception of multiply and divide instructions, and possibly other arithmetic instructions also involving the MQ register).
Don't feed the trolls.
mille, as long as Hasbro doesn't kill it (I presume that's why there's no "bornes" in the name). At one point, it was one of my favorite toys for "compiling!" xmille (and the accompanying README) if you want a GUI.
Here is your answer: http://en.wikipedia.org/wiki/Kalman_filter
The good news is that, for example, a Google search for
actually finds things.
We're not talking about a "time left" bar, we're talking about a progress bar
Actually, the person who submitted the story was talking about a progress bar accompanied by a "time left" message:
No.
Ian, is that you?
From the article: "Seems to violate law of conservation of momentum". - Yup it does.
And from the article, in more detail: "It seems to violate of the law of conservation of momentum, implied by Newton, which says that no closed system can have a net thrust. However, Shawyer says net thrust occurs because the microwaves have a group velocity which is greater in one direction than the other and Einstein's relativity comes into play."
Is the article and/or Shawyer trying to say here that "Einstein's relativity" magically makes the law of conservation of momentum go away (perhaps the idea is that Einstein's laws replace Newton's laws, or something such as that)? I may be misremembering my physics from ages ago, but I though conservation of momentum was just as much a law in Einsteinian mechanics as is Newtonian mechanics.
case in point TCP/IP vs OSI
TCP/IP was quickly created by some students and released into the world, and then improved on.
Actually, the specs were created by a bunch of people including at least three Ph. D.'s, and there were several independent implementations, including one done at Bolt Beranek and Newman for UNIX; the latter one was subsequently modified by some students at Berkeley (I have the impression that, in addition to Bill Joy, Sam Leffler worked on the 4.2BSD TCP/IP stack) and offered as part of BSD UNIX, and the rest was history.
ISO OSI covers TCP/IP in that it is a logical break down of the various aspects of the stack. The fact that most implementations shortcut aspects of it for performance does not mean it is invald. It is a useful abstraction even if only in theory.
There's the OSI Reference Model, as described by ISO/IEC 7498-1 (available as ITU-T Recommendation X.200), and there's the OSI protocol stack, with protocols such as the Connectionless Network Protocol ISO/IEC 8473-1 (available as X.233), the Connection-Oriented Transfer Protocol ISO/IEC 8073 (available as X.224), and the Connectionless Transport Protocol ISO/IEC 8602 (available as X.234).
The "ISO OSI" that "covers TCP/IP" is the Reference Model. The person to whom you're responding is talking about the OSI protocol stack.