So what's going on here? Are we replacing the Apple/Microsoft rivalry with Linux and Apple? And why would Torvalds go out of his way to criticize Mac OS X unless he felt Linux was threatened?
You must be logged in to the O'Reilly Network to post a talkback.

Showing messages 1 through 13 of 13.

• Torvalds Claims He Was Misquoted
  2001-04-18 13:41:48  Derrick Story | [Reply | View]
  
  

  Here's an interesting note that I received from Ward Meier --DS:

  
  
  

  MacSurfer's Headline News posted this link on April 11th bringing into question the validity of the quote that has everyone up in arms about OS X:

  
  
  

  http://www.itworld.com/Comp/2384/LWD010410maccomments/
  

  
  

  From that article it seems pretty clear that either Torvalds was
  misquoted and people are trying to use it as ammunition toward their
  cause, or that he was not willing to back something that he did say.
  Either way, he is definitely not standing by the statement of which he
  is accused of making. Hope this is helpful.

  
  
  

  --Ward Meier

  
  

• Mac OS X, Linux and Linus Torvalds
  2001-04-18 10:42:21  bosscat [Reply | View]
  
  
  I'm glad someone in the trade press is actually paying attention to what their peers say and calling bs when they see bs. Keep up the good work!
  
  I've been watching posts from various sources concerning OS X vs. Linux since before the OS X beta was released and all I can say is what a bunch of whiners!
  
  I use Linux on my servers and will continue to do so for the forseeable future not because I think its better than OS X, but because the learning curve was so damn steep I don't want to screw with something that works. I probably won't even update to Linux 2.4. Don't fix what isn't broken.
  
  As to Linux itself, I have to work in the real world which means I need to be able to read and generate documents in MicroSloth's office suite. I have yet to find a similar apps suite for Linux that actually works. (Read that as something my wife can use - she's a Luddite when it comes to software). I have this ability on OS X and thus avoid the "bitch factor".
  
  
  I've been hearing a lot of whining about speed - well its interesting that once Classic loads up MS Office apps seem to run as fast and in the case of Excel faster than they did on OS 9. Of course I spent a little money to increase my RAM to 500 Megs, but what the hey.
  
  If you are running slow, its the RAM stupid!
  
  Besides, the OS has been out for less than six months. How well did Linux (or UNIX for that matter) work after only six months on the market? Not nearly as well I'm sure. Give it time. And btw the same holds true for apps native to OS X, they will appear in time.
  
  As to Linus Torvalds, he has my respect as an individual and as a programmer. He's not the messiah or God, however. These are the preserve of Bill Gates, just ask Bill, he'll tell you.
  
  While Linus has probably done more to actually get opensource os's off the ground than anyone else I can think of he hasn't done it alone, and Linux is still a long way from being a household word. It's simply not consumer ready, and until the geeks get off their purity soapbox and make it so, Linux will not be able to stand up to the same scrutiny as either OS X or it pains me to say Windows.
• Mac OS X, Linux and Linus Torvalds
  2001-04-18 13:48:12  Derrick Story | [Reply | View]
  
  
  I think one of the points you make in your excellent note is worth highlighting: Linux is having a hard time on the desktop. Despite the excellent efforts of Nautilus and Ximian, when people think Linux they think backend, embedded, or clustering. On the other hand, Mac OS X probably won't be the backend beast that Linux is, but I think it is finally going to bring Unix to the desktop. And that's a good thing.

• We Really Don't Know What Torvalds Said, But ...
  2001-04-16 10:30:06  Derrick Story | [Reply | View]
  
  

  Here's a thoughtful letter that I received that raises some good points -- Derrick

  
  
  

  One problem with this story, as one of the other posters noted, is
  that we really don't know what Torvalds said. We know what was reported
  by someone who doesn't give us the context of the statement nor indeed
  the full quotation. When the book's published then we'll know the truth.

  
  
  

  What I want to know and what may prove interesting in the long term is
  whether Torvalds is altogether against microkernel architecture in favor
  of his monolithic approach. If you read the usenet postings between him
  and Andy Tanenbaum
  (http://www.dina.dk/~abraham/Linus_vs_Tanenbaum.html#mvsms) from the
  early '90s, Torvalds is willing to admit some of the charges levelled
  against his monolithic approach are true and states that microkernels
  are "nicer" (whatever that really means). In response to Tanenbaum's
  arguments for micro vs. monolithic Torvalds responds:

  
  
  

  "True, linux is monolithic, and I agree that microkernels are nicer.
  With a less argumentative subject, I'd probably have agreed with most of
  what you said. From a theoretical (and aesthetical) standpoint linux
  looses. If the GNU kernel had been ready last spring, I'd not have
  bothered to even start my project: the fact is that it wasn't and still
  isn't. Linux wins heavily on points of being available now."
  

  
  

  
  (In that short paragraph with it's concluding statement Torvalds wrote
  everything anyone needs to read to know why then and now Linux has been
  successful.) What I hope to discover from his book is 1) is it Mach in
  particular he thinks is crap or does he also implicate all microkernels,
  and 2) if he thinks all microkernels are crap, when and how did he come
  to this conclusion.
  

  
  

  
  Is Torvalds' pronouncement the death knell for OS X? No! I've been
  using Macs for 10 years going back 20 years for Apple computers in
  general. I like the new OS and believe over the next year we'll see it
  solidified by Apple and the Macintosh community. I have some gripes,
  especially about performance. But like it's cousin FreeBSD, I find it
  robust with a fit and finish I can appreciate and enjoy. As for OS X
  being crap, that won't be determined based on early-days speed and
  graphics issues nor in academic discussions touting monolithic vs.
  microkernel architectures. Apple will solve the performance issues and
  we'll see better graphics drivers for Quartz. We'll know whether Mac OS
  X is crap when we know whether Apple and the Macintosh community can
  produce the critical mass of desktop applications and systems support
  necessary for the average user to work productively without becoming a
  'nix-head.
  
  

  
  

  
  Finally, I think the headline of your piece is misleading and
  inflammatory. The article from Yahoo UK certainly leads one to believe
  that, but IMHO you shouldn't propagate crap.
  
  

  
  

  
  Thanks for the article and the MacDev forum and the great publications
  your company produces.
  

  
  

  
  --Aaron
  

• We Really Don't Know What Torvalds Said, But ...
  2001-04-19 19:21:55  pyrrho [Reply | View]
  
  
  Agreed with one of your last points... that the subject of "Torvalds thinks OS X is crap" is highly inflammatory and misleading. It indicates that the author didn't bother to investigate the story further, and instead just tried to deliver what s/he felt would provide the greatest return in readership.
  
  This is highly contradictory to what I feel O'Reilly publications have represented so far... highly researched, objective and informative publications. This article casts a black mark upon everything else I'll be looking at on oreillynet.com for some time to come.

• Micro-kernel vs. Monolithic-kernel?
  2001-04-09 17:38:10  joelreed [Reply | View]
  
  
  The impression I took away from this article and from what I had already read elsewhere is that Linus's "design issue" statements were about Apple's decision to use a Mach based microkernel. Linus has traditionally been opposed to microkernel's but I am not sure that he is fully aware of how Mach and Mac OS X are put together. What traditionally differentiated microkernel's from monolithic kernels was what components were in kernel space and what were in user space. Monolithic kernels traditional being implemented with more components in kernel space, while microkernels moved those components from kernel space to user space. Linus of course being on the monolithic design side of the fence (Linux being a monolithic kernel.)
  
  However in a quote from Apple's "Inside Mac OS X: Kernel Environment" document:
  
  "... in Mac OS X, Mach is linked with other kernel components into a single kernel address space. This is primarily for performance; it is much faster to make a direct call between linked components than it is to send messages or do RPCs between separate tasks. This modular structure results in a more robust and extensible system than a monolithic kernel would allow, without the performance penalty of a pure microkernel."
  
  http://developer.apple.com/techpubs/macosx/Kernel/General/KernelEnvironment/KernelEnvironment.pdf
  
  Wether any of that would change the mind or opinion of Linus who knows? Probably not, most geeks are deeply religious and there's nothing wrong with that. However I think it is not fair to assume that Apples just blindly walked down the microkernel path without considering any potential "design issues". From the above quote and the rest of the document it sounds like Apple is aware of the "design issues" and has their own solutions. Wether Apples design choices for its Mach implementation and to a larger extent the Mac OS X kernel are right or wrong who knows.
• Micro-kernel vs. Monolithic-kernel?
  2001-04-09 22:17:50  Derrick Story | [Reply | View]
  
  

  I understand Torvalds' long-standing opinion that microkernels are not the way to go. Fair enough.

  
  
  

  What still befuddles me is that he would (allegedly) make such a blanket statement that he knows someone in the press is going to pick up and run with.

  
  
  

  You could argue that the alleged quote has received way too much airtime. But that's what happens when you start calling the other guy's toys crap.

  
  
  

  There's got to be more here than microkernel bashing.

• Micro-kernel vs. Monolithic-kernel?
  2001-04-10 03:42:55  giorgio [Reply | View]
  
  
  I hope that Torvalds has some solid reason to say something like that. And even in this case, the fact that a smart and highly respected guy makes such an arrogant statement, is beyond me. And makes this interview irrilevant, IMVHO.
• Micro-kernel vs. Monolithic-kernel?
  2003-03-26 01:10:50  muzzle [Reply | View]
  
  
  I recently read a bashing Linus received from the creator of minix outlining the vulgarity of his 'design' of the Linux kernel. Linus may have never quite recovered from this ;-) Linux has hype and support but it in no way makes Linus the highest authority on kernel architecture and it must be noted that in some camps it is believed that Linux kernel began as a hack and is considered quite vulgar in its implementation.

• Will Knight is misinformed
  2001-04-09 15:51:02  Corvus [Reply | View]
  
  
  This issue has been debated/flamed excessively on Mac OS and Linux fora.
  
  Will Knight has taken a single 1997 quote from Linus's forthcoming autobiography "Just for Fun" about the Mach microkernel which has nothing *NOTHING WHATSOEVER* to do with Mac OS X, plus the fact Linus was offered a job after graduation by Apple which he declined (along with 1e09 other offers by practically every technology company on Earth) and grossly misrepresented it, either through technical ignorance or willful perniciousness. Or both.
  
  As best as I can determine, the facts are these:
  
  1. In the 1980s CMU develops the Mach microkernel, a "pure" microkernel in the sense that everything other than process management runs as a user-process.
  
  2. In 1991 Linus Torvalds began work on a x386 Unix clone based on a monokernel, where most OS processes run in a shared privileged process.
  
  3. In 1997, Linus claims "pure" microkernels are inefficient and unnecessary, with low performance. For an example he cites the Mach microkernel.
  
  4. In 2001 Apple releases a new OS which uses a *modified third-generation version* of the Mach mickokernel with many performance improvements specifically but not exclusively incorporating many *OS processes running as shared privileged processes* in order to avoid the performance problems associated with "pure" microkernels, the same way Linux, Solaris, Windows 2000, and just about every OS not used for research does.
  
  5. Linus decides to publish an autobiography including his earlier comment about "pure" microkernels.
  
  6. Will Knight reads a prerelease of the autobiography, reads the comment about Mach, sort of remembers something about Mach being in Mac OS X, puts 2 and 2 together to get -3.14159, and publishes an irresponsible opinion piece masquerading as hard news about how "Linus sez Mac OS X SUX!".
  
  7. Computer sites which confuse advocacy with fact (including but not limited to ZDNet, the Register, and Slashdot) breathlessly report this unsubstantiated vapor as fact, adding their own free-associations about Linus hating/being threatened by/jealous of Apple. No one bothers to verify any of these remarkable assertions or ask the principals involved.
  
  8. Millions of technology professionals realize the technology press is about as unbiased and accurate as the National Enquirer.
• Linus says he was misquoted
  2001-04-11 18:15:18  Corvus [Reply | View]
  
  
  In an interview with LinuxWorld, Linus Torvalds says the quote in his autobiography was taken completely out of context and he "never commented on OS X". A summary is available at http://www.itworld.com/Comp/2384/LWD010410maccomments.
  
  I will expect Will Knight's retraction shortly after the Pope's wedding.

• Seems Unlikely That Linux Folks Would Feel Threatened
  2001-04-09 15:19:59  Derrick Story | [Reply | View]
  
  

  If Linux were thriving on the desktop (via Nautilus and Ximian for example), then I could see how Mac OS X could be perceived as a threat. But the fact of the matter is that Linux is floundering on the desktop while it continues to build momentum on the backend.

  
  
  

  Since Apple is strong on the desktop and weak server-side, it seems like there's plenty of Unix real estate for both entities.

  
  
  

  Am I missing something here?

• Seems Unlikely That Linux Folks Would Feel Threatened
  2001-04-16 12:35:38  gagganator [Reply | View]
  
  
  apple is weak on the server because of the old mac os. with os x, they become a competitor; indeed they have a site up already: http://www.apple.com/macosx/server/
  
  because of the non x86 hardware and the closed os, i think mac os x is more a competitor to solaris than linux