Apple, Linux and BSD: The ‘Other’ Platforms

33 Comments

• How does one install an OS on a headless device?

  Log in to Reply

  • Maybe smug isn’t the right word, but we can feel more secure knowing that BSD has already survived the IP lawsuit filed by AT&T prior to the Unix rights being sold to SCO. I don’t see any reason this issue would need to be re-visited.

    Log in to Reply

    • Whether BSD/Apple are safe from SCO’s wrath is a difficult question. Everyone who’s watched the SCO cases carefully knows that SCO’s leadership isn’t always consistent, and their words should probably not be relied upon as guarantors of their behavior. The question of whether BSD derived Unixes are safe from SCO probably lies in whatever agreements are in the sealed settlement between BSD and USL. Unfortunately, we don’t have access to that document, though anyone who’s interested can find the existing court records here:
      |
      http://cm.bell-labs.com/cm/cs/who/dmr/bsdi/bsdisuit.html
      |
      Currently, the available evidence says SCO won’t survive their encounter with IBM, (or RedHat, or SGI) so it’s probably not worth worrying about, but saying that it either will or won’t happen is almost certainly a matter of too little knowledge.
      |
      Alex

      Log in to Reply

      • <<a) Edit a config file in /etc/inet/
        b) Restart or re-run ifconfig
        c) Set up a user account (can’t login as root via ftp, for example)
        d) Edit /etc/ftpaccess>>
        .
        Use SFTP. Duh.
        .
        We disable telnet and ftp on our machines and only allow access via ssh and sftp. Ftp and Telnet are broken since they transmit passwords in the clear. I can use ethereal to get your password in seconds flat on your network unless they are all connected by switches.
        .
        If you happen to be using 802.11, even with WAP, you’re a big juicy target for a hacker.

        Log in to Reply

        • You make many valid points and illustrate effectively where this article falls short on the facts. I would like to reiterate only one point regarding Apple’s OS X Server. The author says, "The things that make an Apple PC compelling — user interface and industrial design — don’t play well on servers, adding up as simply unneeded cost."
          This is absurd for at least two reasons. First, I never hear the same criticism or concern about Windows 2000 and 2003 servers, even though they have the same GUI overhead and even much LESS in the way of command-line options. Second, OS X Server can run completely headless, without any GUI software running at all and taking up system resources. From the start Apple was savvy to this very concern — and so it gives sys-admins the option of running it only with the command line available.
          So much for that point by the author.
          -Jeff Mincey

        • I grow weary of self-appointed Unix purists who take it upon themselves to determine what is and is not Unix — and, in particular, that BSD is not Unix. Unix was born at AT&T and evolved over time to be SVR4. But at some point in the development of Unix, a different branch emerged — the Berkeley branch. In the beginning, it was in fact a mirror image of the AT&T Unix — but over time it evolved some different conventions.
          .
          There is no single "pure" Unix anymore. Even AM ong the SVR4 Unix variants, a single binary is not guaranteed to run unrecompiled from one platform to the other — even with the same hardware architecture. And consider also that there was a time that X11 was not considered a valid part of Unix, though for years now it has been accepted as a part of the standard Unix package. Moreover, the real issue with respect to standards across the Unix family is Posix. Today the definition of an operating system is larger than just the microkernel alone. An OS includes ALL the layers — and not just some. Thus with full Posix compliance, and with the standard suite of Unix shells and tools, I think an operating system has a plausible claim to the Unix name.

        • SFTP is not installed by default on Solaris, Duh.
          I’d like to see you "get [my] password in seconds flat on [my] network" because the Sun and the Mac were connected to each other directly via a crossover cable. Neither one has access to any network you could ever dream of hacking.
          The point was not about network security (Duh), it was about the ease with which one could administer services on the machine. Duh.

        • <<I grow weary of self-appointed Unix purists who take it upon themselves to determine what is and is not Unix>>
          .
          Unix is a trademark. You can ask the OpenGroup whether Linux or BSD are Unix. They aren’t. They are reimplementations of the POSIX standard which is what defines Unix. Neither of them are Unix derivatives because they are not based on Unix code. BSD at one point could have been argued to be based on Unix code, but the BSD vs USL lawsuit settled that.
          .
          Linux and BSD are Unix compatible operating systems. Although this might seem like nit picking, the Open Group who actually owns Unix would agree with me, and so would a US court of law.
          .
          <<Unix was born at AT&T and evolved over time to be SVR4. But at some point in the development of Unix, a different branch emerged — the Berkeley branch. In the beginning, it was in fact a mirror image of the AT&T Unix — but over time it evolved some different conventions.>>
          .
          The reality is that it was Universities, not AT&T that perfected Unix. AT&T took what Universities created and incorporated it into their Unix. Berkeley was a large contributor and that is what the origin of the BSD USL lawsuit is. Essentially, USL tried to do what SCO is trying to do, steal other people’s work and claim it as their own.
          .
          <<There is no single "pure" Unix anymore. Even AM ong the SVR4 Unix variants, a single binary is not guaranteed to run unrecompiled from one platform to the other — even with the same hardware architecture. And consider also that there was a time that X11 was not considered a valid part of Unix, though for years now it has been accepted as a part of the standard Unix package.>>
          .
          X11 is NOT part of Unix. It’s an application that runs on top of it created by MIT, from the Athena project. Calling X11 "Unix" is like calling Apache Unix. X11 isn’t tied to Unix, in fact, X11 runs on Mac and Windows. You can download XFree86 on top of cygwin and verify this yourself if you so please.
          .
          Unix is nothing more than the kernel. Most applications that run on top of it are actually GNU tools. Without the GNU tools, a SCO Unix box would be little more than an expensive door jam.
          .
          <<Moreover, the real issue with respect to standards across the Unix family is Posix. Today the definition of an operating system is larger than just the microkernel alone.>>
          .
          No it isn’t. Microsoft tried to, unsuccessfully I might add, redefine the definition simply to get around the 1995 consent decree which they signed with the DOJ agreeing not to bundle products to destroy competitors. There is no technical valid reason to incorporate IE into the kernel, it decreases stability, it makes unnecessary dependencies, and takes up resources unnecessarily.
          .
          <<An OS includes ALL the layers — and not just some.>>
          .
          You don’t know what you’re talking about.
          .
          <<Thus with full Posix compliance, and with the standard suite of Unix shells and tools, I think an operating system has a plausible claim to the Unix name.>>
          .
          Try it. The Open Group is suing Apple over this very issue, and they will win. Unix is a trademark, it’s not a thing. It’s a certification, and neither BSD or Linux has it. In fact, Linux isn’t even certified to be POSIX compliant, and Windows is. Pontificate on that.
          .

        • First, let me suggest your position consists of both legal and technical components — and for my part I’m interested only in the latter and not the former. What party holds which trademark and who sold which rights to who on what date — life is too short for that and I don’t think this gets to the heart of what Unix is in the first place. I wish to confine my "lens" only to the technical.
          .
          Second, if indeed the term, Unix, refers to a bona fide OS — then please enlighten me as to which one it is. I ask because I contend there is no such product as the "Unix Operating System." There is AIX, HP-UX, Tru64 (which was born as OSF/1), Solaris, and many other Unix-based or Unix-oriented OS’s. But there is no "Unix OS." So I contend you are hard pressed to make the case that Unix is an OS when in fact you can’t obtain one anywhere.
          .
          What, then, is Unix? I submit it is a term which designates a technology and not a specific operating system. More accurately, it designates a technology FAMILY. Now was it always this way? No. There was a time when Unix was indeed an OS. But times change and terminology has to adapt.
          .
          Third, I never said that AT&T was the chief mover in the development of Unix but only that it was BORN there.
          .
          Fourth, I never said X11 is Unix. I said it is a key _component_ of Unix or Unix technology. When X11 first came out, the Unix "purists" of the day did not welcome it with open arms. But today, just try selling a Unix distro without X11 as part of it and then see how far you get.
          .
          Fifth, I contend that an operating system is more than the kernel alone. If it were just the kernel, then we wouldn’t use the term kernel in the first place but rather we would just call that the OS. I submit that a modern-day operating system consists of multiple layers which is based on the kernel, of course, but also the hardware abstraction layer, I/O and file subsystems, drivers, APIs, and a UI (whether CLI or GUI). All these things together make up an operating system.
          .
          And if this is not so, if these other components are "extras" which are superfluous to the OS, then please advise as to where I can obtain a shrink wrap product of the HP-UX kernel alone (to use one example) — sans all these "extras." Where can I obtain the UFS file system separately — apart from a Unix bundle?
          .
          Foremost, the main point I wish to make is that Unix designates an OS _technology_ and not an OS itself. And this technology has parameters, of course, but they are broad enough to include a number of OS’s which can rightly fall under the Unix rubric.
          Jeff Mincey

        • <<First, let me suggest your position consists of both legal and technical components — and for my part I’m interested only in the latter and not the former. What party holds which trademark and who sold which rights to who on what date — life is too short for that and I don’t think this gets to the heart of what Unix is in the first place. I wish to confine my "lens" only to the technical.>>
          .
          Let me suggest you are simply wrong.
          .
          Neither Linux nor BSD have passed Unix certification, therefore they are not Unix. If you want to talk about API compatibility, talk about POSIX, but only Linux has passed that and only for TurboLinux and like a 1/2 decade ago.
          .
          <<Second, if indeed the term, Unix, refers to a bona fide OS — then please enlighten me as to which one it is. I ask because I contend there is no such product as the "Unix Operating System." There is AIX, HP-UX, Tru64 (which was born as OSF/1), Solaris, and many other Unix-based or Unix-oriented OS’s. But there is no "Unix OS." So I contend you are hard pressed to make the case that Unix is an OS when in fact you can’t obtain one anywhere.>>
          .
          The term Unix refers to a group of operating systems conforming to a standard.
          .
          <<What, then, is Unix?>>
          .
          http://www.unix.org/what_is_unix/the_brand.html
          .
          <<I submit it is a term which designates a technology and not a specific operating system.>>
          .
          It’s a certification and a trademark. The same thing is true of DOCSIS, which if you have a cable modem follows that specification and standard. The reason it’s trademarked is to prevent anybody from willy nilly claiming compatibilty without having it.
          .
          <<More accurately, it designates a technology FAMILY. Now was it always this way? No. There was a time when Unix was indeed an OS. But times change and terminology has to adapt.>>
          .
          In the industry, which you don’t appear to work in and I DO, it means a specific thing. Making up terms doesn’t make them valid terms. Your definition is wrong.
          .
          <<Third, I never said that AT&T was the chief mover in the development of Unix but only that it was BORN there.>>
          .
          Did I say you said that? What you implied is that BSD originated from AT&T Unix, the reality is that BSD code made AT&T Unix what it is today, and so did a lot of other Universities.
          .
          <<Fourth, I never said X11 is Unix.>>
          .
          You said it was a "it has been accepted as a part of the standard Unix package". It’s nothing more than an application on top of Unix. QNX, for example, doesn’t have X11 at all. It’s passed POSIX and may have passed Unix compliance.
          .
          <<I said it is a key _component_ of Unix or Unix technology. When X11 first came out, the Unix "purists" of the day did not welcome it with open arms. But today, just try selling a Unix distro without X11 as part of it and then see how far you get.>>
          .
          QNX is making more money than SCO is, but then again, their product doesn’t suck.
          .
          <<Fifth, I contend that an operating system is more than the kernel alone. If it were just the kernel, then we wouldn’t use the term kernel in the first place but rather we would just call that the OS. I submit that a modern-day operating system consists of multiple layers which is based on the kernel, of course, but also the hardware abstraction layer, I/O and file subsystems, drivers, APIs, and a UI (whether CLI or GUI). All these things together make up an operating system.>>
          .
          Well, I suggest that you read an operating systems book then.
          .
          <<And if this is not so, if these other components are "extras" which are superfluous to the OS, then please advise as to where I can obtain a shrink wrap product of the HP-UX kernel alone (to use one example) — sans all these "extras." Where can I obtain the UFS file system separately — apart from a Unix bundle?>>
          .
          Make a deal with HP. All you’ve said that since it’s bundled for MARKETING purposes, it’s a single unit.
          .
          <<Foremost, the main point I wish to make is that Unix designates an OS _technology_ and not an OS itself. And this technology has parameters, of course, but they are broad enough to include a number of OS’s which can rightly fall under the Unix rubric.>>
          .
          Unix is a certification which neither BSD, Linux, or OS X has – therefore they aren’t Unix. This is all I’m saying, and I’m absolutely right about it and you are completely wrong to disagree.
          .
          You can complain that this definition isn’t correct, but it is. Too bad.
          .
          <<Jeff Mincey>>

        • <<SFTP is not installed by default on Solaris, Duh.>>
          .
          With Solaris 9 it is. What are you using?
          .
          <<I’d like to see you "get [my] password in seconds flat on [my] network" because the Sun and the Mac were connected to each other directly via a crossover cable.>>
          .
          Oh, that’s a really useful network configuration, a cross over cable to another computer 2 feet away.
          .
          <<Neither one has access to any network you could ever dream of hacking.>>
          .
          Your network configuration is unusable. I can similarly secure my systems by locking them in a vault, but it makes work a little bit difficult to do.
          .
          <<The point was not about network security (Duh), it was about the ease with which one could administer services on the machine. Duh.>>
          .
          BTW you could have just logged into the Apple box via FTP directly. Why didn’t you think to do that? You don’t need any servers running on the Solaris box at all. Does Apple automatically allow non secure access to the OS? what you’re complaining about is that you had to explicitly allow access by setting up an insecure server (ftp) with Solaris which is, frankly, stupid.
          .
          I will conceede that if you setup a network in stuch an illogical manner, perhaps the Apple excels in that configuration. Of course, if you knew what you were doing, you could have just started the ftp server up directly with running wu.ftpd as root, and the reason you cannot log into root via ftp is that the root password is sent in the clear. If you can log into any privledged account on Apple, it’s because they sacrificed security for ease of use, something that Microsoft does which makes it extraordinarily vulnerable to viruses.

        • This will be my last comment on this subject. When having a discussion with someone on a topic, it’s important first to endeavor to understand their position before you can respond intelligently to it. The other person may indeed be mistaken in his views, but in order for you to counter those mistaken views effectively, you can do so only when you first understand what they are in the first place. Absent that approach, you will only be arguing with a straw man — i.e., against a position that was never put forward to begin with. And that is what is happening here.
          .
          For example, you keep harping on this issue of a Unix trademark — to the point that anyone privy only to your side of this discussion could only infer that I must take issue on this score. The thing is, however, I have never ONCE taken issue with this. And while I know it’s much easier to win a debate when you create your "opponent’s" position for him, I’ll represent my own position, if you don’t mind. (Besides, for me this was only about an exchange of views and not about scoring debate points anyway.)
          .
          In any event, since you keep missing it, for the record let me spell this out: I acknowledge that Unix is a trademark. I acknowledge that one who uses that term without certification risks a lawsuit. I know all that. I knew it from the very beginning when I alluded to the legal implications of this matter. Indeed, I have known about the Unix trademark for years and I have never made a SINGLE statement in contradiction to your contention about it.
          .
          One can only wonder, then, why like some wind-up doll you keep reciting it. NONE of this information (about trademarks) comes to bear on my position — and the fact that you THINK it does only illustrates that you don’t really understand what my position is.
          .
          Not all positions can be presented in one-liners. Some require several paragraphs to flesh out. And while it’s fine for you to interject replies to excerpts of my comments, if you slice and dice my words without first reading them through at least ONCE in their entirety, then you run the risk of failing to understand where I’m coming from. And that’s what we have here. So I see no point in continuing this — especially since you are on record as saying you don’t care even what developers (of operating systems) themselves think about this matter. Like the religious fundamentalist, you know you are right — period.
          .
          The irony is that indeed you ARE right in the particular position you are taking. You are right that Unix is a trademark and that only one organization has a right to that name (or to decide the terms for licensing its use to other parties). I guess it has just escaped your notice that I never disputed this in the first place, hey?

        • <<This will be my last comment on this subject. When having a discussion with someone on a topic, it’s important first to endeavor to understand their position before you can respond intelligently to it.>>
          .
          What is intelligent about your position? Unix is a trademark and a certification controlled by the Open Group which neither BSD or Linux have. You claim that Linux and BSD in particular are Unixes, but they neither have the trademark or the certification. When I originally responded to Enderle’s erroneous article, that is what I was pointing out and apparently what you are taking issue with. Why you take issue with a known fact I don’t know. If you recall, you started this. If I was right in the first place, why did you start this?
          .
          <<The other person may indeed be mistaken in his views, but in order for you to counter those mistaken views effectively,>>
          .
          All you need to do is present facts when they support your argument. This is why I don’t have to confuse the issue and merely need to point to the definition of what Unix is and what Unix isn’t, and neither Linux or BSD are Unix because 1) they haven’t passed certification and 2) they won’t pass certification, at least Linux won’t because it doesn’t have all the required signals for example.
          .
          Let me refresh your memory by the way,
          .
          http://www.unix.org/version3/pr.html
          .
          <<you can do so only when you first understand what they are in the first place.>>
          .
          You are creating a definition and trying to support you position by saying the definition which YOU CREATED is correct. It’s ludicrous.
          .
          The Open Group owns the trademark and they are the certification authority. Nothing more need be said.
          .
          <<Absent that approach, you will only be arguing with a straw man — i.e., against a position that was never put forward to begin with.>>
          .
          Your position is that Linux and BSD are Unix and I’m a "purist" to state otherwise.
          .
          You are creating a stawman of you own, claiming that BSD and Linux fall under the definition of what a Unix is, when clearly they do not.
          .
          <<And that is what is happening here.>>
          .
          Identify the strawman I’ve supposedly created. What position of yours AM d I putting foward that you never did, pray tell?
          .
          <<For example, you keep harping on this issue of a Unix trademark — to the point that anyone privy only to your side of this discussion could only infer that I must take issue on this score. The thing is, however, I have never ONCE taken issue with this.>>
          .
          You haven’t? Then how is Linux or BSD either a Unix when neither one of them meet the certification requirements to be consider "Unix"?
          .
          <<And while I know it’s much easier to win a debate when you create your "opponent’s" position for him, I’ll represent my own position, if you don’t mind.>>
          .
          Your position is, need I remind you: "I grow weary of self-appointed Unix purists who take it upon themselves to determine what is and is not Unix — and, in particular, that BSD is not Unix.
          .
          "(Besides, for me this was only about an exchange of views and not about scoring debate points anyway.)"
          .
          BSD isn’t Unix, and neither is Linux. The reason I wrote back to Enderle is that his facts are so grossly in error, it wasn’t about scoring debate points, merely correcting the erroneous views of somebody that bills himself as an expert behind the authority of an "analyst".
          .
          <<In any event, since you keep missing it, for the record let me spell this out: I acknowledge that Unix is a trademark. I acknowledge that one who uses that term without certification risks a lawsuit. I know all that. I knew it from the very beginning when I alluded to the legal implications of this matter. Indeed, I have known about the Unix trademark for years and I have never made a SINGLE statement in contradiction to your contention about it.>>
          .
          Yet you don’t admit that you are incorrect to claim that BSD is Unix. BSD isn’t Unix. I believe you are just moving the goalposts.
          .
          You also stated, "There is no single "pure" Unix anymore." which you’re almost correct about, there are 4 pure Unixes at minimum now.
          .
          <<One can only wonder, then, why like some wind-up doll you keep reciting it.>>
          .
          Because you’re wrong. I hate it when somebody is proven to be wrong and won’t admint it. Consider it a pet peeve.
          .
          <<NONE of this information (about trademarks) comes to bear on my position — and the fact that you THINK it does only illustrates that you don’t really understand what my position is.>>
          .
          Your position is that BSD is a Unix and for me to not agree with that is to be a "Unix purist". Engineering is all about technical accuracy. BSD isn’t Unix, period.
          .
          <<Not all positions can be presented in one-liners.>>
          .
          Which is why I always take care to respond to your entire post, and why you don’t bother to extend the same curtesy to me.
          .
          <<Some require several paragraphs to flesh out.>>
          .
          Some just need a definition pointed out to them. Why you persist in your mistaken belief I don’t know.
          .
          <<And while it’s fine for you to interject replies to excerpts of my comments, if you slice and dice my words without first reading them through at least ONCE in their entirety, then you run the risk of failing to understand where I’m coming from. And that’s what we have here. So I see no point in continuing this — especially since you are on record as saying you don’t care even what developers (of operating systems) themselves think about this matter.>>
          .
          Exact quote please? Because what you neglected to mention there is that sometimes developers fail to reach certification even when they are aiming for it. My exact quote was:
          .
          "It doesn’t matter what the developers think. Everybody that goes through DOCSIS certification fails on the first pass, for example."
          .
          And by the way, neither BSD or Linux is aiming for Unix certification. They use the POSIX standard simply because it’s better than reinventing the wheel. There are PLENTY of idiotic requirements to meet to become "Unix", why *should* BSD or Linux meet those? It’s nothing but a waste of time and develoment effort.
          .
          Porting between Unix and Linux or BSD is basically trivial. That demonstrates the uselessness of a lot of the requirements. Let me add that porting between Linux and CYGWIN is also trivial, and there is absolutely no way that anybody would consider CYGWIN a Unix derivative.
          .
          <<Like the religious fundamentalist, you know you are right — period.>>
          .
          I know I’m right because the owner of the tradmark and the holder of the certification process agrees with me.
          .
          http://www.unix.org/version3/pr.html
          .
          I notice you cannot come up with anything that backs of your position.
          .
          <<The irony is that indeed you ARE right in the particular position you are taking. You are right that Unix is a trademark and that only one organization has a right to that name (or to decide the terms for licensing its use to other parties). I guess it has just escaped your notice that I never disputed this in the first place, hey?>>
          .
          Let me refresh your memory. You said:
          .
          "In contrast, in seeking to understand the meaning of this term, Unix, I look not to any single "authority" but instead to how the term is used throughout the industry. "
          .
          The "authority" (ironic quotes yours, not mine) refers to the Open Group who owns the standard and the certification. You went on to argue that I was using argument by authority, when I AM merely using the industry definition and you are simply making a definition up.
          .
          but this is my favorite gem, from the first post you made:
          .
          "Today the definition of an operating system is larger than just the microkernel alone. An OS includes ALL the layers — and not just some. Thus with full Posix compliance, and with the standard suite of Unix shells and tools, I think an operating system has a plausible claim to the Unix name."
          .
          The "ALL" is your emphasis.
          .
          By your defintion, not having BASH on a machine and having SH would create a new OS. Adding IE to Windows created a new OS. The *only* reason Microsoft tried to create a new definition was to weasel out of a legal agreement they made with the DOJ. Nobody in the tech industry accepts that definition (well, maybe in Redmond they do), and neither did Judge Kimball.

        • You make very interesting points. One thing in particular leapt out at me, however, which is this statement of yours, "…Linux is a kernel, FreeBSD is a complete OS…" This is especially interesting in light of my discussion below with another individual regarding just what is and is not Unix (or what constitutes Unix today or how this term is defined). My "debate opponent" contends that the kernel IS the totality of the OS — as if all else is superfluous. It’s as if he makes no distinction between a shink-wrap copy of MS-Office, for example, and the UI routines to draw windows, scroll bars, radio buttons, and other GUI objects. To him, incredibly, they are both just add-ons.
          .
          So it’s good to see that at least one other individual can parse more distinctions than this and can recognize that the kernel is the foundation and core of an operating system but not the whole of it.
          .
          Jeff Mincey

        • <<You make very interesting points. One thing in particular leapt out at me, however, which is this statement of yours, "…Linux is a kernel, FreeBSD is a complete OS…" This is especially interesting in light of my discussion below with another individual regarding just what is and is not Unix (or what constitutes Unix today or how this term is defined). My "debate opponent" contends that the kernel IS the totality of the OS — as if all else is superfluous. It’s as if he makes no distinction between a shink-wrap copy of MS-Office, for example, and the UI routines to draw windows, scroll bars, radio buttons, and other GUI objects. To him, incredibly, they are both just add-ons.>>
          .
          It is AM using that you feel that you have to attack me by proxy. Tell me, do you know what an
          .
          OPERATING
          .
          SYSTEM
          .
          Is?
          .
          Just for the pure AM usement of it, why don’t you define exactly what an operating system is?
          .
          Just because somebody makes the same errors you do does not indicate that you’re correct. An operating system provides scheduling, an abstraction to the hardware, and basically allows you to operate the system. That is why it’s named an "Operating System" coincidentally enough.
          .
          What you’re doing is trying to use the marketing term that’s newly invented. We all know how expert marketers are what with their paradigms, and their proactiveness. You now marketers, peole that sell you BS and call it fertilizer, people that invent words to make them feel "smrt".
          .
          <<So it’s good to see that at least one other individual can parse more distinctions than this and can recognize that the kernel is the foundation and core of an operating system but not the whole of it.
          .
          Jeff Mincey>>
          .
          We can debate here if you prefer. It beats my current contract job.

        • Very well, I’ll bite. The Jargon Lexicon defines it well: "The foundation software of a machine; that which schedules tasks, allocates storage, and presents a default interface to the user between applications." Note that last part. That’s where the difference between an OS and a kernel becomes clear. Let’s take a simple example: MS-DOS was an operating system… the MSDOS.SYS file that was part of it was not. Instead, the combination of COMMAND.COM (to present the user with a prompt and basical internal commands) in conjunction with an established set of external commands (format, debug, xcopy, etc) made up the "userland" which was an accepted part of an operating system.
          Linux is a kernel. In and of itself, it is of little use. Wrap it up with a proper userland and you have a complete OS, but it’s not just Linux anymore, because there is no de-facto Linux userland. Instead, to get a complete Linux OS you need to have a Linux distrobution, of which there are many (Red Hat, Gentoo, etc).
          One could argue semantics I suppose but the spirit of the whole kernel vs. OS think when it comes to Linux vs. FreeBSD holds true no matter how you slice it. Call it engines and cars if it makes you feel better. Linux is an engine, which requires someone like Red Hat to build the rest of the car around the Linux engine. Other companies can use the Linux engine in other cars which drive completely different than Red Hat’s. Meanwhile, FreeBSD makes the engine AND the car AND everything else. My FreeBSD 5.1 is like my friends. Your Linux 2.4-based OS might be rather different than your cousin’s Linux 2.4-based OS however.
          Is that better?

        • Thanks for taking the bait. I think you make a good case. To further elucidate your car analogy, I would say that FreeBSD is not the whole car but is rather the whole mechanism or "operating system" required to make the car function AS A CAR (which is to say as a vehicle for motor transport). By this definition, such things as a radio, power windows, seats, ash tray, etc, are NOT part of the operating system. But there are still plenty of non-engine ("non-kernel") components which nonetheless qualify as part of a car’s "operating system." Examples include the drive shaft, tires, brakes, differential, chassis, axles, and dashboard. (The dashboard is essentially the UI.) Even though these things are not part of the internal cumbustion engine, without them the car cannot function as a car.
          .
          Moreover, it’s not only the question of a UI that makes your point (in the definition you supply above); it’s also the matter of allocating storage which does as well. The allocation of storage and the file system used thereto (and the device drivers which manage the I/O to disk) are NOT part of the kernel. And yet these things are not options on a computer (in the sense that a radio or power windows are options on a car). Indeed, the kernel itself requires storage from which it is loaded into ram at boot time. Now, sure, there are diskless workstations without local storage, but they only substitute network I/O (and TCP/IP transport layers) for disk I/O — and these subsystems — whether disk or network — are not baked into the kernel. Does this mean they are "add-ons"? I hardly regard disk and network resources to be add-ons to an operating system.
          .
          Also, let’s consider the term, operating system, itself. I submit it is more consonant with our holistic and systems definition than it is to a mechanistic (or "kernel only") one. "Operating" implies that which is necessary for the operation of a device. I submit that a computer cannot "operate" without storage or network subsystems or I/O drivers of some kind, any more than a car could operate without a drive shaft (even though it’s not part of the engine itself). And "system" refers to the coming together of multiple disparate parts to function as a cohesive whole. A "system" does not consist of only one part — i.e. the kernel alone.
          .
          Finally consider that the kernel itself is not some inviolate and sacrosanct component about which all systems engineers agree. Anything can actually be installed in the kernel, (hence the distinction between monolithic kernels and microkernels). Merely because a component in a given configuration may not be installed in the kernel doesn’t then mean it’s no longer part of the OS. Consider Windows NT 4.x. With this version Microsoft made the decision to bake the video drivers into the kernel itself (for the sake of enhanced performance). For this advantage they were willing to sacrifice stability, but it was a judgment call on their part (an unwise judgment in my view, but that’s another subject). Does this mean that in Windows NT, the video subsystems were indeed part of the OS, but that in all other operating systems, it was not?
          .
          I contend that the definition of an operating system must include the standard of that which is necessary for the functioning of whatever device it is intended to "operate." And by this measure, the function of managing system resources and the interfaces thereto, (such as UFS, TCP/IP, USB, Firewire, SCSI, ATA, PCI, PCI-X, DDR ram, Samba, etc), pass the test. These things are not optional. Oh, sure, a given implementation of a technology is optional — (for example, one might use one interface to storage and not another, such as SCSI instead of 1394) — but having an interface AT ALL to system resources is NOT optional but essential. And once you establish that it is essential for the operation of a device, then you are hard pressed to make a case that it’s not part of the "operating system" of that device. And it’s hardly "marketing speak" to say so.

        • <<Very well, I’ll bite. The Jargon Lexicon defines it well: "The foundation software of a machine; that which schedules tasks, allocates storage, and presents a default interface to the user between applications. Note that last part.>>
          .
          Then by that definition, Linux isn’t an operating system at all because it doesn’t provide a default interface to the user between applications at all. That is the weakness everybody complains about in X11.
          .
          <<That’s where the difference between an OS and a kernel becomes clear. Let’s take a simple example: MS-DOS was an operating system… the MSDOS.SYS file that was part of it was not.>>
          .
          By the definition you gave, MS DOS wasn’t an operating system, and really it wasn’t. It was a collection of software interrupts.
          .
          <<Instead, the combination of COMMAND.COM (to present the user with a prompt and basical internal commands) in conjunction with an established set of external commands (format, debug, xcopy, etc) made up the "userland" which was an accepted part of an operating system.>>
          .
          Is Linux and operating system or not?
          .
          <<Linux is a kernel. In and of itself, it is of little use. Wrap it up with a proper userland and you have a complete OS, but it’s not just Linux anymore, because there is no de-facto Linux userland. Instead, to get a complete Linux OS you need to have a Linux distrobution, of which there are many (Red Hat, Gentoo, etc).>>
          .
          You mean wrap it up in KDE or GNOME and it becomes an operating system.
          .
          Also, by your definition, eCos isn’t an operating system, WinCE isn’t an operating system, vxWorks isn’t an operating system, pSOS isn’t an operating system, Lynx isn’t an operating system and Nucleus isn’t an operating system.
          .
          <<One could argue semantics I suppose but the spirit of the whole kernel vs. OS think when it comes to Linux vs. FreeBSD holds true no matter how you slice it. Call it engines and cars if it makes you feel better. Linux is an engine, which requires someone like Red Hat to build the rest of the car around the Linux engine. Other companies can use the Linux engine in other cars which drive completely different than Red Hat’s. Meanwhile, FreeBSD makes the engine AND the car AND everything else.>>
          .
          What is FreeBSD’s default interface? X11? The differences between BSD and Linux at the user level are so minimal that getting entagled into a war about how one differs from the other is pointless. BSD will eventually die simply because the GPL is more attractive to coders and cannot be proprietarized like BSD’s codebase can be, which is why companies love it and as a result take from it but do not contribute back. The GPL license prohibits this type of exploitation which is why Linux has grown so quickly, and although it’s less stable as a result of it’s faster development, it’s also more advanced because it doesn’t only encourage that modifications and enhancements are to be made public, it forces them to be. From the standpoint of 30 years out, it’s pretty plain to see what the final result will be.
          .
          <<My FreeBSD 5.1 is like my friends. Your Linux 2.4-based OS might be rather different than your cousin’s Linux 2.4-based OS however.
          Is that better?>>
          .
          All you said is that there is less variety in BSD, something I’m fully aware of. This is why Linux is taking over the embedded space and BSD doesn’t even exist there except as parts like the TCP/IP stack of vxWorks. I can’t really see that as an advantage.
          .
          I’m also curious, are there any BSD variants doing the equivalent of Beowulf or super computers?

        • "Instead, to get a complete Linux OS you need to have a Linux distrobution, of which there are many (Red Hat, Gentoo, etc)."
          Actually you don’t need any of these linux dists to get a linux system running. You can compile and build your very own customised version without all the bloat that comes with other vendors. True, it’s much easier to make this customised version using one of these distributions as a build platform but it’s not totally nessesary (otherwise how did anyone build linux before they came along??).
          You can quite happily get along with the kernel and a few GNU apps.
          Oh, and the linux kernel on it’s own CAN be the whole operating system but it wouldn’t be much use since you need at least a shell (or equivalent) to communicate with it and tell it what you want it to do, hence the name GNU/Linux which the Richard Stallman followers would tell you is more accurate (not that i’m in total agreement).

        • First of all, a debate over the definition of "operating system" defeats the whole point of my original post. So I’ll throw it out the door. Yes, of course Linux can be made into an "operating system", but there is no "THE Linux Operating System". It’s really that simple. Here’s another way of putting it: Linux in-and-of-itself does not create a complete user-experience that brings an interface to a user at the console. FreeBSD does. I AM not saying Linux is bad, I’m saying the Linux school of thought is DIFFERENT. Why do you take such offence at that? Forget debates about the definition of OS. My point is simply a statement of fact. Geesh.
          .
          And no, Red Hat or any other distribution is far more than taking the Linux kernel and "wrapping it up in KDE or Gnome". I would assume as a Linux advocate you understanding of your own OS-of-choice would be more thorough than that. Of course you don’t need a GUI to be an OS.
          .
          FreeBSD’s default interface is the Bourne shell (/bin/sh). Where you go from there is up to you, but you have a default usable interactive interface to work with.
          .
          Meanwhile, I’d advise against subscribing to the typical "BSD-is-dying" troll commentary. It’s used so much by Linux zealots that it’s become humorous because of course it’s wrong. BSD is different, and not dying. In July Netcraft produced a report that stated that almost 2 MILLION active sites were using FreeBSD, and also said "FreeBSD secured a strong foothold with the hosting and internet services communities at the genesis of the web and has anything but gone away. Indeed it is the only other operating system [besides Windows and Linux] that is gaining, rather than losing share of the active sites found by the Web Server Survey." And note that only applies to web hosts and does not take into account the growth of it on the desktop as well.
          .
          http://news.netcraft.com/archives/2003/07/12/nearly_2_million_active_sites_running_freebsd.html
          .
          Your arguments about the benefits of the GPL license over the BSD license are pointless, because both have their appeal and disadvantages, often depending on the community. There are plenty of places that choose the BSD software license specifically for its merits. Apple, for example.
          .
          Your arguments about there being less variety in the BSDs are also pointless. I see that as an ADVANTAGE, not a disadvantage. As it is, you often can’t just download the Linux version of a close-source piece of software. You have to download the Red Hat version, or the whatever-your-distro-is version. A complete-and-defined OS from kernel to userland makes for a predictable development environment, which is one of the many reasons I like FreeBSD.
          .
          You hear more about Linux and clustering than FreeBSD because Linux-the-kernel is more of a sub-component so it can be reworked and stuck into more different places. The end-result may not resemble much of what a typical Linux-user would come to expect, however. The whole car/engine analogy again. But since you asked:
          .
          http://www.freshports.org/net/pvm/
          http://www.freshports.org/net/mpich/
          http://www.freshports.org/net/clusterit/
          http://acme.ecn.purdue.edu/
          http://www.canonical.org/~kragen/beowulf-faq.txt (see the answer to question #1)
          http://www.high-availability.com/open/Solutions/index.php
          .

        • <<First of all, a debate over the definition of "operating system" defeats the whole point of my original post. So I’ll throw it out the door. Yes, of course Linux can be made into an "operating system", but there is no "THE Linux Operating System". It’s really that simple. Here’s another way of putting it: Linux in-and-of-itself does not create a complete user-experience that brings an interface to a user at the console.>>
          .
          Neither does WinCE.
          Neither does vxWorks.
          Neither does Nucleus.
          Neither does pSOS.
          Neither does eCos.
          Neither does Lynx.
          Neither do most supercomputers.
          .
          What are those then? Ham sandwhiches? Apparently, they aren’t operating systems.
          .
          <<FreeBSD does. I AM not saying Linux is bad, I’m saying the Linux school of thought is DIFFERENT. Why do you take such offence at that? Forget debates about the definition of OS. My point is simply a statement of fact. Geesh.>>
          .
          Here’s a question. Does FreeBSD, NetBSD, and OpenBSD use different kernels?
          .
          <<And no, Red Hat or any other distribution is far more than taking the Linux kernel and "wrapping it up in KDE or Gnome". I would assume as a Linux advocate you understanding of your own OS-of-choice would be more thorough than that. Of course you don’t need a GUI to be an OS.>>
          .
          I AM not a "Linux advocate". I’m simply stating facts. I can see which direction the wind is blowing. Linux will wipe out Microsoft not because it’s better, in fact, the X11 interface is far inferior to DirectX – Linux will kill Microsoft in the same way Microsoft killed all it’s competitors – it’s "good enough" and it’s cheaper. That is all that matters.
          .
          It will kill the BSDs as well, but will take much longer to do so. This is because it has a shorter development cycle. Not because it’s faster, not because it’s better, it’s because the GPL forces enhancements to be included back into the public domain and the BSD variants do not. This encourages fragementation. BSD code and also be incorporated into GPL code so any really whizbang features that show up in BSD can be easily included into Linux, but not vice versa.
          .
          <<Meanwhile, I’d advise against subscribing to the typical "BSD-is-dying" troll commentary.>>
          .
          It’s not a religious war with me. Why people get all upset about this I don’t understand. BSD is superior in stability, but that’s about it. It will be occluded by Linux, again not because Linux is better, but because Linux can borrow from more projects than BSD can.
          .
          Take for example ReiserFS. It’s under a modified GPL license. If it was full GPL, it couldn’t be included into BSD. Say this FS is all it’s cracked up to be. BSD developers would have to re-implement it to use it or stay behind. This will happen over and over and over again. Meanwhile, corporations who are more than happy to take free work and sell it won’t contribute back enhancements to the public.
          .
          It’s not a problem of the OS itself, it’s a weakness of the license it’s under. For better or worse, and both are easy to argue, the GPL fosters faster growth but it also discourages companies from donating to it.
          .
          BSD is very attractive to companies, because you can modify it, then sell the entire thing as a private proprietary product. This is a weakness in development though. BSD isn’t going to die tomorrow, or even in the next decade. GPL will either flourish or perish in a cataclysm, BSD can just fade away. I don’t think this was an intended effect of the GPL license, but it is a result of it.
          .
          <<Your arguments about the benefits of the GPL license over the BSD license are pointless, because both have their appeal and disadvantages, often depending on the community. There are plenty of places that choose the BSD software license specifically for its merits. Apple, for example.>>
          .
          Apple chose BSD precisely for the reason I stated they did, it allows mixing of proprietary technology with "free" technology. If Apple used Linux instead, there would be hackers running Quartz on x86 and Mac software there as well. Apple is still (stupidily) a hardware company.
          .
          I recognize the advantage it has for companies, but it’s at the detriment of software development.
          .
          <<Your arguments about there being less variety in the BSDs are also pointless. I see that as an ADVANTAGE, not a disadvantage.>>
          .
          My belief is that variety breeds innovation and that was Darwin’s belief as well. It’s not a matter of design, as much as it’s a matter or chance. Good ideas get sucked up by the more mainstream ideas. Knoppix and Gnoppix for example may actually produce a red hat clone. Does BSD have anything like that? There is Mame Knoppix too. There is a Knoppix for DVD that boots up and then plays a movie with Xine or MPlayer – no need for HW codecs anymore… Does BSD enjoy similar ideas?
          .
          <<As it is, you often can’t just download the Linux version of a close-source piece of software. You have to download the Red Hat version, or the whatever-your-distro-is version. A complete-and-defined OS from kernel to userland makes for a predictable development environment, which is one of the many reasons I like FreeBSD>>
          .
          Like what? I’ve never run into a problem getting any closed source code running under any version of Linux, simply because everybody that is trying to make money on Linux is eager to be compatible with closed source projects to proliferate sales of the OS.
          .
          <<You hear more about Linux and clustering than FreeBSD because Linux-the-kernel is more of a sub-component so it can be reworked and stuck into more different places. The end-result may not resemble much of what a typical Linux-user would come to expect, however. The whole car/engine analogy again. But since you asked:
          .
          http://www.freshports.org/net/pvm/
          http://www.freshports.org/net/mpich/
          http://www.freshports.org/net/clusterit/
          http://acme.ecn.purdue.edu/
          http://www.canonical.org/~kragen/beowulf-faq.txt (see the answer to question #1)
          http://www.high-availability.com/open/Solutions/index.php>&gt;
          .
          The reason you hear more about Linux clustering than BSD clustering is that Linux has been doing it at NASA since 1994 and it’s simply done more often with Linux than BSD. For example: the University of Illinois built a supercomputer out of playstations. The US government tried to put restrictions on Linux for that reason. Linux has a higher name brand recognition.
          .
          Again, this isn’t because Linux is better, but it’s the state of affairs.

        • Last I knew, we were talking about general-purpose operating systems for personal computers meant for use by the public to run a variety of applications. This is a certain category of OS, of which FreeBSD, Windows, and Mac OS are a member of. It would be helpful if you kept your points within the context of the topic at-hand, lest you be debating religion in your next post. I’m not talking about embedded operating systems designed to run a single application and be unknownst to the end-user. I suppose my DVD player might have an OS but it’s not relevant to a comparison of Windows, Mac OS, FreeBSD, and Linux-based OSes such as Red Hat, SuSE, Gentoo, and so on.
          .
          If I really cared about embedded OSes as part of this, I’d go on about http://people.freebsd.org/~picobsd/ but since I don’t, I’ll simply provide the link for you own curiosity.
          .
          Within the subject-matter of the article and the context of the points I’m making, Linux is just a kernel compared to the likes of Windows, FreeBSD, and Mac OS. I don’t really see where the debate is on this point.
          .
          FreeBSD has many strengths besides its stability. The development style, the social culture, the organization of the project, the school-of-thought which is applied to how things are done, the ports system, the sensible filesystem hierarchy, the ease of install, beyond the other points which I see as benefits although you view them as negatives.
          .
          I’m not really interested in a debate of licenses. You like the GPL, fine. I really don’t care and license holy-wars are way off-topic from the points I AM making. You are free to feel that Linux will rule the world in 20y. I disagree, and it really doesn’t matter to me that your predictions of the future a generation-out differ than mine. While you’re running Linux me and my friends and family will be running FreeBSD and enjoying the benefits of that.
          .
          Not sure why you bring NetBSD and OpenBSD into this. I have only been talking about FreeBSD. I do not collectively advocate on behalf of all *BSDs nor have any of my points been about or applied to *BSD. No, they don’t use the same kernel. So what? FreeBSD doesn’t use the same kernel as Windows either. It has nothing to do with what I’m saying. You’re advocating a kernel which is used in many operating systems. I’m advocating a SPECIFIC operating system with a SPECIFIC kernel. OpenBSD and NetBSD are fine OSes and have their strengths and place, but I do not use them or speak for them.
          .
          I’m sorry but I do not have a running list of sites where I’ve seem multiple Linux binary downloads depending on your distro. I do not run Linux so it hasn’t been worth noting, but I’ll certainly start making note in case this topic comes up again. I can say though that it’s been several times and most commonly the odd-ball is Red Hat.
          .
          As you say yourself, having more name-recognition and a bigger installed-base does not make ones OS a better choice. One could use the same argument with Windows against your Linux and you’d take offence. People can and do use FreeBSD for clustering. You simply asked if it was possible and if people did it. I provided the answer.
          .
          You also ask if there’s something like Knoppix using FreeBSD. The answer is yes:
          .
          http://www.freesbie.org/
          http://www.freshports.org/sysutils/livecd/

        • <<Last I knew, we were talking about general-purpose operating systems for personal computers meant for use by the public to run a variety of applications. This is a certain category of OS, of which FreeBSD, Windows, and Mac OS are a member of. It would be helpful if you kept your points within the context of the topic at-hand, lest you be debating religion in your next post. I’m not talking about embedded operating systems designed to run a single application and be unknownst to the end-user.>>
          .
          Embedded operating systems are NOT limited to running a single application and "be unknownst to the end-user". An embedded OS is an OS, although it doesn’t fit your silly definition of what an OS is. An OS is nothing more than an abstraction of the hardware that presents a consistent interface to the PROGRAMMER, not the user.
          .
          <<I suppose my DVD player might have an OS but it’s not relevant to a comparison of Windows, Mac OS, FreeBSD, and Linux-based OSes such as Red Hat, SuSE, Gentoo, and so on.>>
          .
          What about your Tivo? That runs Linux. Replay TV does too. Palmtops run anything from vxWorks to WinCE.
          .
          <<If I really cared about embedded OSes as part of this, I’d go on about http://people.freebsd.org/~picobsd/ but since I don’t, I’ll simply provide the link for you own curiosity.>>
          .
          I’m not really curious as I’m well versed in real time embedded operating systems and BSD has absolutely no presence in the space. It never will because support is what is paramount, not software availability. Since there is such a high variability in platforms, free OSes always are going to be playing catchup while companies that distribute RTOSes will simply take the code, make their modifications to it, and sell it.
          .
          <<Within the subject-matter of the article and the context of the points I’m making, Linux is just a kernel compared to the likes of Windows, FreeBSD, and Mac OS. I don’t really see where the debate is on this point.>>
          .
          Do me a favor and answer just ONE of my questions, if Linux isn’t an OS, and vxWorks is, how can that be possible? Are you saying and embedded OS isn’t an OS at all?
          .
          <<FreeBSD has many strengths besides its stability. The development style, the social culture, the organization of the project, the school-of-thought which is applied to how things are done, the ports system, the sensible filesystem hierarchy, the ease of install, beyond the other points which I see as benefits although you view them as negatives.>>
          .
          None of these are advantages from a technological standpoint, except for the ease of install which really isn’t any better than Linux. You might as well have said BSD is cooler than Linux. It is similarly ineffable.
          .
          <<I’m not really interested in a debate of licenses. You like the GPL, fine. I really don’t care and license holy-wars are way off-topic from the points I AM making. You are free to feel that Linux will rule the world in 20y. I disagree, and it really doesn’t matter to me that your predictions of the future a generation-out differ than mine. While you’re running Linux me and my friends and family will be running FreeBSD and enjoying the benefits of that.>>
          .
          Whatever. I brought it up simply because you consider BSD to be an "OS" and Linux isn’t. A viewpoint that is utterly ludicrous.
          .
          <<Not sure why you bring NetBSD and OpenBSD into this.>>
          .
          I’m curious, do they run different kernels or not?
          .
          <<I have only been talking about FreeBSD. I do not collectively advocate on behalf of all *BSDs nor have any of my points been about or applied to *BSD. No, they don’t use the same kernel. So what?>>
          .
          So you end up with 3 basically identical things with 3 different codebases which fragment. I code for a living, maintainence is the nightmare of every project. *BSD has at least 3 of those.
          .
          <<FreeBSD doesn’t use the same kernel as Windows either.>>
          .
          What is your point? The BSD’s all have a common ancestor, but they fragment. For what purpose? The BSD community is up to their eyeballs in some stupid religious war. That kind of crap in the Linux community is relegated to distribution wars.
          .
          <<It has nothing to do with what I’m saying. You’re advocating a kernel which is used in many operating systems.>>
          .
          I’m not advocating ANYTHING.
          .
          <<I’m advocating a SPECIFIC operating system with a SPECIFIC kernel. OpenBSD and NetBSD are fine OSes and have their strengths and place, but I do not use them or speak for them.>>
          .
          I just wanted to underscore with a clear example a problem with BSD’s license. As far as I’m concerned the difference between FreeBSD, OpenBSD, and NetBSD are about as important as the differences between RedHat, Suse, and TurboLinux EXCEPT BSD doesn’t have a unified code base.
          .
          <<I’m sorry but I do not have a running list of sites where I’ve seem multiple Linux binary downloads depending on your distro.>>
          .
          There are certainly different versions of the kernel and vendors do apply patches to it, but there is only one place that holds what is considered Linux, that’s at kernel.org. The changes are minor, if they exist at all.
          .
          <<I do not run Linux so it hasn’t been worth noting, but I’ll certainly start making note in case this topic comes up again. I can say though that it’s been several times and most commonly the odd-ball is Red Hat.>>
          .
          Yes. Red Hat tries to distinguish itself. Red Hat’s Linux is, however, based on the newest version of the the official Linux each time, it doesn’t branch. BSD does, and it will continue to fragment.
          .
          <<As you say yourself, having more name-recognition and a bigger installed-base does not make ones OS a better choice. One could use the same argument with Windows against your Linux and you’d take offence.>>
          .
          No, I wouldn’t. My point is that you assume that technology even matters, my point is that it doesn’t and it’s been proven it doesn’t. The Amiga in 1986 was probably the best computer available on the home market. It could emulate a mac and with a plug in card run x86 software. Now it’s dead, and yes, I realize there is a new computer called the AmigaOne which is nothing more than a living fossil. It’s dead. Why? Apple is only existing still because Microsoft gave them a few 100 million to keep them in business a few years ago to get around the DOJ. Atari ST is dead.
          .
          My point is that in terms of technology, Windows sucked, but it still prevailed, because it was cheaper. Linux has a lower standard to meet in terms of maintenance and the license makes fragementation illogical. Linux can also incorporate BSD technology, but BSD cannot do the same.
          .
          <<People can and do use FreeBSD for clustering. You simply asked if it was possible and if people did it. I provided the answer.>>
          .
          Yes. But BSD doesn’t lead there – it’s still a curiosity there. BSD, by all rights, should be farther along than Linux is in ALL respects, but it’s not – is it only because of the BSD USL lawsuit? That ended in 1994, and that was about the time Linux reached version 1.0 – stability.
          .
          Don’t you ever pontificate as to why BSD was marginalized by Linux?
          .
          <<You also ask if there’s something like Knoppix using FreeBSD. The answer is yes:
          .
          http://www.freesbie.org/
          http://www.freshports.org/sysutils/livecd/>&gt;
          .
          Hunh, didn’t know that.

• I’m glad FreeBSD got mentioned, but the author should do his homework and make some valid points for comparison purposes.
  One major piece of information that he neglected to include was that FreeBSD can run Linux apps, often faster through emulation than the same apps run natively on Linux. This is a testament to just how efficient and clean the FreeBSD code is. Not that I have to run Linux apps often, but if I want to run RealPlayer for Linux instead of mplayer I can do so fine. In the past I’ve also run WordPerfect for Linux (no longer necessary, with programs like OpenOffice and AbiWord which can compile natively) on FreeBSD. This allows one to gain all the benefits of FreeBSD while still having access to all those Linux apps out there.
  And note I’m only talking about close-source Linux apps. Many people who move to Linux also do so because they want free software. Well, if it’s free open-source software, it should compile just fine in FreeBSD making Linux-compatibility a moot point (and chances are it’s already one of the 9000+ programs in the FreeBSD ports system). I run FreeBSD as a desktop (yes, people do this all the time, it works great) and the only Linux apps I have are RealPlayer (which I hardly ever use) and the JRE I used to bootstrap the compilation of a native FreeBSD JRE (hopefully not necessary for too much longer… blame Sun for this one).
  Beyond Linux-emulation and the famous ports sytems, FreeBSD has many other unique strengths. One of which is its notorious stability. In fact, Netcraft just compiled their stats for Sept and found that 7 of the top 13 internet hosts in Sept run FreeBSD (only 2 ran Linux). Yahoo runs on FreeBSD. Hotmail did too before MS corrupted it. The development process of FreeBSD is vastly different from Linux (and remember, Linux is a kernel, FreeBSD is a complete OS) and this results in stability and performance being priorities over "whiz bang" features and having drivers ASAP (which may be buggy, or poorly-written). Quality still counts for something.
  If you decide to learn everything from magazines you may figure Linux is the only open-source OS worth running. But those who do their homework and discover FreeBSD never go back…

  Log in to Reply

• Re: "The biggest long-term problem with moving to an Apple platform is that the company is in decline, which means you might have to migrate again at some point to another platform."
  With regards to Apple, the author must have been living in a "Microsoft" cave somewhere for quite a while.
  Per InternetWeek.com on Oct. 7/03 at
  http://wagblog.internetweek.com/
  which discusses an InternetWeek Reader Poll of dissatisfied Windows users,
  "Many users are now unhappy with Microsoft, but don’t see any alternative. They will likely see alternatives over the next few years, as Linux gets better.
  "A surprising number of respondents said they switched to Apple. As of about six years ago, Apple appeared to be nonexistent on corporate desktops, except as a niche product used by visual workers such as graphic designers. That appears to be changing; the Mac is back."

  Log in to Reply

• He doesnt’ know at all mac products. There are no cheaper than PC products for tremendous power : i think about the e-mac. The X-server is simply to use and its not infected daily by more than 70 000 virus as PC servers do. And next week, finally, its no necessary to explain "how microsoft works successfully". Write rather "how not to get tired of this dirty mess". "Fix the problem": you joke !’. Billl Gates himself doesnt know how to do : "if people get viruses, its because they dont know how to use microsoft patches". Jokin, i say… ridiculous, i maintain.

  Log in to Reply

• Hoo boy, what a mess.
  Mr. Enderle, get your facts straight. Apple and Dell are the only two companies making money building PCs. Apple is NOT in decine and in fact even with the lack of a new 15" Powerbook increased their market share in the laptop world by 1.9 percent in the past quarter. Just wait until the G5 (I just got a dual 2GHz, this thing is going to win converts!) and new Powerbooks have a full quarter to sell in.
  Apple has 3-4 billion and little debt. That’s decline?
  As for the GUI tools on their servers, I use an XServe at work. I AM not an IT professional or a UNIX guru, and the gui tools are invaluable for someone like me who knows enough to to the job with the help of the GUI tools and has no desire to waste time doing things on a command line when I don’t need to!
  he XServe’s performance, even with lowly G4 processors is AM azing.
  TCO for Macs? A research scientist I worked with used to work for Raytheon. They did the research and found Macs were much less expensive to support. Many fewer IT folks required to support them. (No wonder IT people bad-mouth Macs!)
  Macs aren’t for everyone – yet. But they are getting so much better by leaps and bounds that they are anything but in decline! Please, spare us the "unique’ point of view when it has nothing to do with reality.

  Log in to Reply

• From: HTTP_USER_AGENT: Mozilla/5.0 (Macintosh; U; PPC Mac OS X; en-gb)
  AppleWebKit/85.7 (KHTML, like Gecko) Safari/85.5
  Hello
  I USE computers, to do things
  Can you help me? I use Mac OS X and Linux Mandrake. I AM having a trouble. My computers do work as they are designed to do, i.e. they run, run, run and run, every day.
  I cant see any alternatives out there for me. MS Windows does not seem to be an alternative for anyone these days, and it does not run, run, run and run.
  It appears to me that MS Windows has stopped working the way it is designed to work. It only seems to work effectively together with MS Software, viruses and worms, and not together with third party software. It appears to me that MS Windows is not safe to use where it is designed to be used, – on novice John Smith’s desktop, connected to the outside world. As a stand alone computer with no Windows doctors and nurses i the building it is pretty useless, i.e. if you use it as it is designed to be used, connected to anything as stupid as The Internet.
  Can you help me ? What platforms should I shift to ?
  Now, back to work
  Have a nice platform for the rest of your day
  Good-by

  Log in to Reply

• "BSD is Unix, the new Mac OS has a BSD core, and Linux is a Unix derivative."
  .
  BSD IS NOT Unix, and Linux is NOT a Unix derivative.
  .
  "On the other hand, there are few viruses that attack Mac OS X Latest News about OS X, and the platform generally is as reliable as the other Unix variants. "
  .
  I suggest you try using a Mac before writing about it’s virtues.
  .
  "And having recently overcome its multithreading issues, Linux can scale very well up to relatively large-capacity, clustered configurations or down to departmental implementations."
  .
  Linux was being used in Beowulf clusters at NASA in 1994. Beowulf clusters turn stock hardware into supercomputers.
  .
  "The only major negative with a Linux deployment is the license, which is currently going through it first real litigation test."
  .
  What a joke. SCO is an ongoing joke in the IT sector and SCO has already had their claims defeated in Germany, and is now actively trying to stay out of court for Lanham violations with both IBM and Red Hat.
  .
  "In the long term, there are significant intellectual property risks with any product that is not owned by a deep-pocketed vendor."
  .
  Um: didn’t you realize that the newest Mac OS is based on Free BSD? BSD isn’t owned by a deep pocketed vendor.
  .
  "BSD Unix has many of the advantages of Linux, but the BSD founders went through much more trouble to eliminate intellectual property issues. As a result, BSD users currently are feeling safe from the kind of threat that is moving against Linux."
  .
  You mean when they went to litigation against USL and beat them to a bloody pulp because it was discovered that AT&T was regularly stealing code from BSD?
  .
  Hey, this is the same thing that SCO did when they took the BPF code and claimed Linux infringed upon it. BPF stands for Berkeley Packet Filter.
  .
  "BSD has the least number of packaged off-the-shelf applications available for it, but it is still Unix"
  .
  No it isn’t.
  .
  "and there are several companies and education institutions using it[BSD] very successfully. It doesn’t have the application support that Linux does"
  .
  BSD can run Linux binaries.
  .
  "The BSD platform is likely the best long-term value of the alternate platforms because it doesn’t enjoy the potential open-ended risk of disruption that could occur with Linux"
  .
  SCO has threatened Apple and BSD both as well, but their claims are just as unfounded there as with Linux. Also, BSD code can be incorporated into Linux, although not vice-versa.
  .
  "All three alternate platforms require strong Unix skills to deploy, manage and secure successfully in an enterprise Relevant Products/Services from Cisco environment, although Apple has done the most work to lower this skills requirement."
  .
  A drooling moron could run a Mac, it requires no special skills. It requires no special Unix skills. That is the whole point of the Mac.
  .
  "Deploying Unix, according to what I’m told, is reasonably easy."
  .
  You’re "told"? What does it take to become an analysts anyhow?

  Log in to Reply

• so-called tech wizards….
  I really wish that the folks purportedly ‘in the know’ would get their facts straight before asserting that Macs total cost (whether desktops or servers) is more than their counterparts (that have the same/similar features and software advantages). Frequently, it simply isn’t true. Yet one more myth that this article’s ‘tech wizzie’ and others get wrong over and over again.
  Some of the things we spend a great deal of time on in Clinical Psychology doctoral training (and, for that matter, in graduate studies in the philosophy of science’ ) is getting your data (and implications) right. Too bad these are often ignored by others.
  Sincerely,
  Dr. Richard Welser
  Clinical Neuropsychologist

  Log in to Reply

• Sigh, here we go again trotting out the same old tired arguments. Kudos for at least getting the article online with the idea that there are alternatives, but…
  DanKnight covered most of the important points regarding current Macs and OS X but a little more detail might be useful here in considering things.
  1. Apple is not in ‘decline’. This is a sorry excuse of knee jerk journalism, repeating the same old tired line because it’s easier than actually looking at the current state of affairs. They are shipping more product than ever before, and have been consistently profitable over the last two years.
  2. "This platform doesn’t use industry-standard AMD- or Intel-based hardware. While the hardware is generally better looking, you’ll pay a premium to get high-performance machines". A few points to consider here: yes you are looking at hardware replacement, rather than being able to run OS X on your existing hardware. However, if Microsoft’s track record is any indication, you’re probably going to want to end of life anything you’re currently using on the desktop today when Longhorn hits in a couple of years. Apple machine’s are some of the most connectable machines out there in their current release so you can easily integrate a Mac into an otherwise Windows infrastructure, including Active Directory authentication, so you don’t need a massive migration process, just a replacement policy. This brings us to the premium pricing question Why is it that the Virginia Polytechnic supercomputer is being around Apple G5 systems? Price. They came out ahead of DELL, Sun, HP, and IBM. Clearly stated was that price was their main consideration. ref: <http://computing.vt.edu/research_computing/terascale/&gt;
  Current Mac sysetms are actually price competitive across the board, with the exception that they don’t have a bargain basement under $500 stripped down model.
  3. "user interface and industrial design — don’t play well on servers, adding up as simply unneeded cost." Ummm, not sure where to start on this one, so the user interface comes first. Well, Microsoft servers only come with a console right? Because user interfaces are not important on servers. Riiighhhht. Effective management interfaces are a necessity on today’s complex servers. The advantage here is Apple’s excellent set of server administration applications that simplify the tasks of managing things like BIND, Apache etc. But they don’t preclude you taking advantage of a huge library of BSD oriented tools, either command line or X11 based. Industrial design is exceedingly important in the server room, where space costs money. Apple’s industrial design lets you squeeze a dual CPU server, along with a 2.5Tb fibre channel array into 5U of rack space, all including the latest serial ATA, and SMART drive and system monitoring for under $15K. Remote SNMP based management ensures that you can integrate these systems with your current enterprise monitoring tools like HP Openview, which the last time I checked could run on a UNIX box or a Windows system. Not to mention the flat fee licensing costs.

  Log in to Reply

• The author makes a few major flaws in his argument.
  First, the allegation that the Mac OS platform "might not have as many applications in total as Linux does". Not only can the MacOS run most legacy MacOS applications under "Classic", not only can it run all recent OS X apps via "Carbon", and recent OS X apps via "Cocoa", it can certainly run a gamut of standard open source apps compiled via standard compilers such as gcc, etc, including standard Unix-oriented utilities, as well as standard X11 oriented utilities. In fact, the list of applications for Macintosh that won’t run on Linux are far greater than the list of Linux apps that won’t run on Macintosh.
  Second, the author commits a foul in discussing the cost of a Mac when he says "you’ll pay a premium to get high-performance machines." To the contrary, the high-performance machines are really where the Apple value shines. Dollar for dollar, few (if any?!) PCs can match the performance per dollar value of the G5 or the PowerBook offers.
  The only criticism of Apple on price that comes close to sticking is at the entry level area, where few Macs retail at less than $1000, whereas cheap (as in quality AND price) are a dime a dozen.
  Third the author stumbles when he says "The things that make an Apple PC compelling … don’t play well on servers, adding up as simply unneeded cost." To the contrary, Apple’s OS X server is MUCH easier to administer than a traditional Linux or Solaris box. Take for example transferring files around between two otherwise never-before configured computers, which I did just yesterday. I needed to transfer files from the Solaris box to the OS X box. To get the Solaris box ready, I needed to:
  a) Edit a config file in /etc/inet/
  b) Restart or re-run ifconfig
  c) Set up a user account (can’t login as root via ftp, for example)
  d) Edit /etc/ftpaccess
  Doing the same thing on a mac:
  a) Open System Prefs and fix the Network
  b) Then hop over to "Sharing" and check the FTP service.
  c) Click apply
  My time is worth a lot to me and my company. Each 15-30 minutes I waste on configuring a Solaris box, Apple has already done for me. if I save a 25-30 minutes a week (that’s only 4-5 minutes a day) thanks to the awesome tools Apple included in OS X (not to mention OS X Server) that adds up to at least a two hours a month. I bill my time at an average of at least $100/hr, and over the period of a year, I can easily recover $2400 in the period of a year.’
  If saving $2400 in time a year adds up as "unneeded cost" then I’d love to see how you value your time. For me, XServes and G5 servers make a heck of a lot of sense.
  In fact, the author makes this point for me when he says " for enterprise-class tasks, you are generally limited to Unix management tools, which tend to be expensive". To the contrary, the management tools included with OS X Server are inexpensive, and provide a lot of value for the cost of the license.
  The author continues to make a hollow argument when he says "The biggest long-term problem with moving to an Apple platform is that the company is in decline, which means you might have to migrate again at some point to another platform."
  Not only is this not backed up with facts, the company is doing quite well, has plenty of cash in the bank, is producing AM azing products for very good value in the mid and high end, is and has been profitable, and isn’t saddled with concerns and lawsuits unlike, say, many Linux vendors about now. So yes, I’m saying that Apple is in as much of a decline as Linux is thanks to SCO… It’s a red herring that avoids the facts for sensationalism.
  The author doesn’t restrict his mistakes to Apple however. "Linux is a solid server platform that doesn’t require expensive hardware. In fact, some of the biggest Linux deployments I’ve seen are on AMD servers — one of the places where the value of AMD exceeds the perceived risk of not using a major server brand."
  Ironically, the server is where Linux requries the greatest cost. Feature per feature, compared to an Apple solution for equal quality hardware, Apple’s 1U XServe certainly stands up to the Dell 17xx 1U servers which are comparable in performance and features, and when equipped with Red Hat Advanced Server, the Dell is marginally more expensive.
  On the departmental server front, the G5 server blows a Dell 2600 out of the water in performance, and the price is actually a little less expensive similarly configured.
  Then the author moves to making comments directly out of his ass: "As with Mac OS X, you are generally limited to Unix management tools when it comes to enterprise-class tasks." Clearly the author has never seen or heard of MacOS X Server or Apple Remote Desktop.
  The author demonstrates his bias when he says that the Linux vendors "targeted by SCO as part of its legal action, the near-term risk is minimal" while decrying gloom and doom for Apple’s performance.
  Curiously the author doesn’t identify that OS X has BSD roots, which indemnifies OS X from SCO-type attacks as much as it does standard BSD Unix, but that wouldn’t confirm the author’s bias regarding the Apple gloom and doom (which, incidentally, has been pervasive in tech "journalism" for close to twenty years now).
  And, of course, to attribute "low-cost hardware benefits enjoyed by Linux" only to BSD only continues the fallacy that Apple products, by insinuation, are not price competitive, which I have shown to be false for similarly configured systems including the software cost…
  The errors continue unabated as the author alleges "all three alternate platforms require strong Unix skills to deploy, manage and secure successfully in an enterprise environment, although Apple has done the most work to lower this skills requirement."
  Not quite. If the author had spent any time with OS X Server (and I certainly have, I AM a certified Apple technician) there isn’t a single thing I HAVE to launch the Terminal for in order to execute ANY "enterprise level" administration task, and this includes BIND, Sendmail, FTP, SSH, Directory Services, Web, SSL, etc. etc. etc. OS X is as point-and-click to administer as is Windows, but OS X makes more sense thanks to its Unix heritage with the Apple user-friendlyness engineered in.

  Log in to Reply

• SCO has stated that Linux is only the first target in their "intellectual property" jihad. In fact, McBride has mentioned Apple and the BSD’s specifically when asked what other targets he might go after (eg: http://news.com.com/2100-1001-981569.html). If, for some asinine reason, SCO were to somehow win their current court case, they will almost definitely go after Apple next, and then probably BSD. Luckily SCO has an extremely weak case. Nevertheless, the BSD and Apple communities shouldn’t be feeling smug that they are somehow immune from these kinds of extortion style lawsuits.

  Log in to Reply