One of the biggest problems (if not the biggest problem) here is in the filesystem. FreeBSD at least has had this fixed for several years since UFS2 uses 96-bit time structures: 64 of those for seconds and 32 for nanoseconds. The kernel, on the other hand, and from a cursory glance, seems to have support ready (i.e. functions to convert and utilize 64-bit time_t values) but not in place by default [e.g. sys/kern/subr_clock.c has a section to return EINVAL when the year > 2037]. Please correct if I'm wrong.
Distributing a bunch of free software for use on any platform, both free and non-free, hampers the adoption of free software by providing less incentive to use a free system to access said software.
All different, relatively parallel types of hardware have their own properties that make them unique and interesting. I always thought it'd be cool to have an IBM PPC or Sun SPARC box, just to see if there was any difference in the processors and their capabilities. Seems like more companies (probably out of cheap-ness and brand-identification) are moving to Intel x86 procs, which is sad. At least companies like Nintendo and Sony are still showing some love for IBM processors.
Let us not scoff at humble beginnings, assuming they even are such. As would seem particularly apt for Ruby: "A journey of a thousand miles must begin with but a single step."
I hope you were not too offended by my gentle mockery of your post. If you were, I apologise.
Eh, just a little bit. But that's OK; it's bound to happen when two people obviously have such strong feelings about something. Plus I wasn't expressing my thoughts accurately in the beginning, so I understand. But it's interesting to have caused something of a stir.:) Despite the fact that I've been watching Slashdot for a couple years, I've got about three total posts; the environment is somewhat different from what I'm used to. Oh, and I apologize for mis-spelling your username; that's a mistake I often try not make.
Anyways, I'm only a freelance web programmer by trade, scripting mostly in PHP with some Perl thrown in when necessary. But to restate the conditions you posed:
What is the alternative to Java? If you want to use something that has support from many sources, is portable, really high performance, has a cross-platform GUI, has a safe memory model, and is free or inexpensive?
Most of my work I do for small businesses who don't have any very specific requirements (in terms of structure) - they just want a website that will do what they want it to do. And for these types of jobs I can figure out a way to coordinate PHP and MySQL to get what they want.
But I guess that doesn't quite answer your question, so as far as my own personal programming goes, I use whatever seems to fit best with the project. If there's a lot of text processing to be done, it's Perl. Most projects where it's going to be highly modularised (a somewhat complex object-oriented design) I'll do in Ruby. [Actually, Ruby uses a lot of the ideas found in Smalltalk, but I find the syntax to be much easier.] If it's got to interact with the OS at a low level, C or C++ is my choice. I also use Python when there's a builtin library that can really take care of what I need, and also for curses programming.
Performance isn't a hot issue for me with the sort of small/medium projects I do, but you can extend these languages in C (just like using JNI) if you've got a severe code bottleneck. And as far as the GUI goes, all of these languages support GTK2, which looks pretty and runs in a lot of different places. And if GTK get's too complicated, there's always just plain Tk.:)
I guess the biggest problem I have with Java is the centricity of Sun's implementation; it's the industry standard, as they say. [Or at least, so it seems to me, but I admittedly don't know much about IBM's efforts.] But Sun didn't support my workstation (FreeBSD) until about a month ago, and they still don't natively support my other systems (NetBSD and OpenBSD). [As if it weren't obvious; I'm a follower of the BSD "copycenter" (as opposed to both copyright and copyleft) ideology.] There are, as you said, alternative "clean-room" implementations available, but in my experiences they don't work that well, and they aren't up to date with the industry standard. Maybe it's just my platform, or maybe they've improved in the last few months and I haven't realised it yet.
I really do appreciate the flexibility and power of the Sun's Java and classpath, but it's unfortunately not under a liberal enough license. So until a relatively comparable open source (GPL or better) version comes around for my systems, I'll just substitute with a more focused approach.
Just to make mention on your last comment:
I use Java because Smalltalk was too slow and too fragmented between different incompatible versions, and because C++ a maintenance nightmare (I find Java's strack traces on exceptions to be a huge benefit).
If you liked some of the ideas behind Smalltalk (like full object-orientation and the "do:" blocks) you might want to give Ruby a try. I really like it; it's very clean and easy to write, and it scales very well (i.e. doesn't get insanely complicated when you write a big project). I agree that C++ is really hard to maintain; or at least, i
After some more careful thought, I concede, Decaf, that you are correct on nearly all points - at least, at this moment in time. For my own part, however, as long as Sun's relationship with Java remains as it is now, I will continue to use Java only when all other more open options are exhausted. And I will also continue to work towards the ultimate success of open source, no matter what others may say, because I believe in its cause and potential. To me, it is more important to do what I think is the Right Thing, than to do what is most convenient.
I would lay claim that Perl and PHP power more online transactions than Java does. Look at the world's most popular forum system, PHPBB. Yes, that runs on PHP. And people don't run LAMJ servers, they run LAMP servers: the P stands for PHP or Perl or Python, all of them open languages. I didn't say that Java is now unpopular in all domains, that is false. But I think that it will degrade because it will not have the ability to adapt like other languages can. If Sun goes down, methinks it would be all over for Java.
And if you want to chide somebody for wanting to overcome the competition, fine. But don't forget that the origin of all open projects is the desire to build a better product, and it's only because we want to be better that we can achieve that. Wanting something is the root cause for it happening. That's not a guarantee, but it's as close as we can get.
Closed products can only compete with open projects if they are significantly superior in quality and available on a sufficient number of platforms. While I think Java fills the latter requirement, it does not the former; it is at least on the same level as equivalent products, and perhaps lower than some others. No amount of marketing can change this: if Java is not sufficiently opened, it will remain on the path to obscurity. Without new ideas being able to add to the product, it will decay.
Why is it that Sun never offers anything for BSD? The FreeBSD community only weeks ago finally got access to official Java binaries, after years of tweaking and weedling from the FreeBSD Foundation. Could it be that Sun has forgotten their heritage, that the only reason they exist is because of BSD Unix? Such arrogance towards the BSD community from a company founded by one of the original BSD developers is astonishing and shameful.
People don't use kernels, people use applications. In other words, the OS is the applications. If the distro gives you the options of KDE or GNOME for desktop management, one important option is missing: None. You know, there are plenty of decent window managers out there that get a lot of use: FVWM2 is my fave, but Fluxbox, BlackBox, WindowMaker... the list goes on and on. And plenty of people use mv/cp/ls rather than Nautilus. The point is, a lot of us still use and work on projects that don't require a boatload of resources.
As far as switching the project to NetBSD goes, I think that'd be a great idea. But one of the things that would make NetBSD so good to fill this position is the fact that you have to choose exactly what to put on it. Don't want GNOME or KDE? Fine, don't pkg_add them; try FVWM2 instead. Unfortunately, an increasing number of Linux distros make that choice for you, and just assume what you want.
Of course, this isn't really about Linux itself, but about the many Linux distributions that have bloat problems, notably Fedora and Suse. These are the guys who put everything, including the kitchen sink, into their basic installs. In other words, the Emacs of Linux distros. Whatever happened to keeping the number of required packages to a minimum? On the other hand, Ubuntu and Slackware manage to install a great OS with small footprints.
BUT... since we're talking about the exotic hardware here, why not give NetBSD a try; it's very small and has great support for all kinds of strange hardware. And it's more free than Linux! It would be very difficult to go wrong with a BSD.
With everything that's been going on lately, it sounds like the American government really wants to take Google down in the war of public opinion. The gov't just keeps trying to make them look worse and worse. And since the American courts typically just allow the gov't to do whatever it wants, they're winning.
Yes, if there's anyone who can stand to make a quick buck off security worries, it's the so-called "security software" businesses themselves.
Of course, they all love to insinuate that all operating systems play on the same level field. (So the only way to make sure you're safe is to BUY OUR [poorly written] PRODUCT!) Now, because a couple people came up with a proof-of-concept exploit, they get their chance to say "See, Mac OS X isn't any more secure than Windows is!"
But of course, we know that there's hardly any truth behind those statements. Sure, all software systems have holes, and flaws, and bugs. But depending on how the software is written, the threats from those vulnerabilities can be mitigated and even made altogether impotent. I'm not an OS X user, but I do exclusively use operating systems that live in the same family tree, the modern BSD Unixen. And I can state from experience that these systems are inherently more secure than some other commercial operating systems, simply because of the way they are written. (For example, the level of source auditing found in OpenBSD, etc. etc.)
It takes a lot of work to be a "geek" in any real sense of the word. Learning the ins and outs of math/science/technology/whatever requires time and commitment, which most people can't (or won't) afford. So why go through all the hassle when you can just pretend you're "down with it" and join the crowd. It's just like any other cultural trend: take (more modern) rap music for example. Not everybody can rap like P. Diddy, but if you wear his clothes and represent yourself as if you were part of his culture, you too can become cool.
I would assume (given the content of the articles presented here) that most in the Slashdot community support the Open Source / Free Software movement. This movement is founded in the idea that the more people working on and watching over a project, the better that project will turn out. How is it any FOSS project different from the Internet? With one country controlling the root servers, with one organization under the Department of Commerce controlling the naming system, how can the Internet possibly be anything other than a manifestation of that singular country?
One might argue against me, saying that all great FOSS projects have a benevolent dictator, say Linus for example. But Linus can only keep up with his project because he has a group of people to whom he trusts absolutely the responsibility to commit changes, &c. These people are his peers, just as the United States should have peers to aid in harnessing the power of the Internet. The power of control should be distributed.
I'm sure the Senator is a freedom-loving man, by his statements; what kind of freedom is controlled by one land alone? That sounds an awful lot like the kind of rule that those who founded this country tried to escape from. It is the principle of both our country, and the idea of democracy itself, to balance power and evenly distribute rights and capabilities. Remember the three branches of government? Checks and balances? What about "all men created equal", with the rights to "life, liberty, and the pursuit of happiness"?
So, why should the United States be the only country to control the Internet?
The same can be said about Windows vs. Linux.
You're absolutely right; but this thread wasn't about Windows vs. Linux, it was about Linux vs. OpenBSD. And sometimes, you need Linux for functionality more than you need OpenBSD for security.
OpenBSD is great, but a lot of programs run only on Linux. Granted, a lot of these are closed down; but if you need to run the Sun JVM, or Borland's JBuilder (two programs I have to use for a computer science class), you need Linux. Even under FreeBSD's Linux Binary Compatibility, they don't work very good, if at all.
I heard that Acclaim went bankrupt shortly after releasing Burnout 2, and the Burnout distribution rights were purchased by EA. Why EA doesn't release the Burnout series for GameCube, I don't know (they offer plenty of other games for it, so it may be a Criterion thing). The first two were so awesome, it's sad that single-system Nintendo owners like myself haven't gotten a chance to play the latest and greatest.
It sounds heartless, but 'jargon' is absolutely necessary for people who do anything beyond the basics with computers. How else would we communicate precise information? And the truth is, most computer language comes from other contexts, where the phrases are even more established.
So if comprehending the difference between kilo-, mega-, and gigabytes is too difficult (even though it's exactly the same for any other measurement dimension, like length or weight!), they should just surf over to Wikipedia and type it in. Professionals can't spend their whole day educating clueless users; they'll have to do it themselves.
I'd honestly expect more from someone who worked at NASA and the DoD. Somehow, the author manages to confuse Unix with UNIX; two very different things. Unix is an idea; UNIX is a product that has earned the name from OpenGroup. Unix is Linux and FreeBSD and even Solaris; UNIX is just one of those things sold by SCO and Sun Microsystems.
It's true that Windows beat UNIX, though not because of technical superiority. Rather, it was the result of ruthless business strategy. The UNIX vendors fought each other (instead of banding together to fight the common enemy that could ruin them all), while Microsoft busily secured the PC platform. No wonder Microsoft rules the home office / small business world these days.
But Windows did not beat Unix. The Unix idea lives on in Linux and BSD, as well as Solaris. And these operating systems are not fighting each other. The 'infighting' among the BSD groups that the author refers to doesn't exist. The BSD groups have always been willing to help in each other's projects. They exist separately because they have different focus areas, not because they hate each other. Sun is opening more of its programs every day to other open source programmers (sure, not Java yet, but let's hold out hope). Linux is willing to share anything and everything with other FOSS projects. How else could you run Linux programs under FreeBSD?
This brings us to another of the author's reasons that UNIX failed - program incompatibility. For UNIX, he's somewhat right. Closed-down programs were un-portable in even the smallest variances between systems. But that's one of the reasons the open source community came into existence all those years ago, to share code so it could be used on more than just your own box. That's why the idea of the portable C compiler sprung forth like light from darkness. That's why the first Unix was written in C on the PDP-11. Portability! And that is exactly why this argument, which only barely applies to UNIX, doesn't apply at all to the modern, open source Unixen, like Linux and FreeBSD.
In short, this article is full of holes. Unix is not UNIX. It didn't take Microsoft to help many UNIX vendors throw themselves over the cliff. The BSDs don't hate each other. C is the language of Unix; C is that language of portability.
Slashdot Mirror
One of the biggest problems (if not the biggest problem) here is in the filesystem. FreeBSD at least has had this fixed for several years since UFS2 uses 96-bit time structures: 64 of those for seconds and 32 for nanoseconds. The kernel, on the other hand, and from a cursory glance, seems to have support ready (i.e. functions to convert and utilize 64-bit time_t values) but not in place by default [e.g. sys/kern/subr_clock.c has a section to return EINVAL when the year > 2037]. Please correct if I'm wrong.
Distributing a bunch of free software for use on any platform, both free and non-free, hampers the adoption of free software by providing less incentive to use a free system to access said software.
The concurrent, distributable, error-tolerant, functional, hot-swappable programming language.
All different, relatively parallel types of hardware have their own properties that make them unique and interesting. I always thought it'd be cool to have an IBM PPC or Sun SPARC box, just to see if there was any difference in the processors and their capabilities. Seems like more companies (probably out of cheap-ness and brand-identification) are moving to Intel x86 procs, which is sad. At least companies like Nintendo and Sony are still showing some love for IBM processors.
(Oh, and the much-hyped Ruby on Rails: 38!)
Let us not scoff at humble beginnings, assuming they even are such. As would seem particularly apt for Ruby: "A journey of a thousand miles must begin with but a single step."
I hope you were not too offended by my gentle mockery of your post. If you were, I apologise.
:) Despite the fact that I've been watching Slashdot for a couple years, I've got about three total posts; the environment is somewhat different from what I'm used to. Oh, and I apologize for mis-spelling your username; that's a mistake I often try not make.
:)
Eh, just a little bit. But that's OK; it's bound to happen when two people obviously have such strong feelings about something. Plus I wasn't expressing my thoughts accurately in the beginning, so I understand. But it's interesting to have caused something of a stir.
Anyways, I'm only a freelance web programmer by trade, scripting mostly in PHP with some Perl thrown in when necessary. But to restate the conditions you posed:
What is the alternative to Java? If you want to use something that has support from many sources, is portable, really high performance, has a cross-platform GUI, has a safe memory model, and is free or inexpensive?
Most of my work I do for small businesses who don't have any very specific requirements (in terms of structure) - they just want a website that will do what they want it to do. And for these types of jobs I can figure out a way to coordinate PHP and MySQL to get what they want.
But I guess that doesn't quite answer your question, so as far as my own personal programming goes, I use whatever seems to fit best with the project. If there's a lot of text processing to be done, it's Perl. Most projects where it's going to be highly modularised (a somewhat complex object-oriented design) I'll do in Ruby. [Actually, Ruby uses a lot of the ideas found in Smalltalk, but I find the syntax to be much easier.] If it's got to interact with the OS at a low level, C or C++ is my choice. I also use Python when there's a builtin library that can really take care of what I need, and also for curses programming.
Performance isn't a hot issue for me with the sort of small/medium projects I do, but you can extend these languages in C (just like using JNI) if you've got a severe code bottleneck. And as far as the GUI goes, all of these languages support GTK2, which looks pretty and runs in a lot of different places. And if GTK get's too complicated, there's always just plain Tk.
I guess the biggest problem I have with Java is the centricity of Sun's implementation; it's the industry standard, as they say. [Or at least, so it seems to me, but I admittedly don't know much about IBM's efforts.] But Sun didn't support my workstation (FreeBSD) until about a month ago, and they still don't natively support my other systems (NetBSD and OpenBSD). [As if it weren't obvious; I'm a follower of the BSD "copycenter" (as opposed to both copyright and copyleft) ideology.] There are, as you said, alternative "clean-room" implementations available, but in my experiences they don't work that well, and they aren't up to date with the industry standard. Maybe it's just my platform, or maybe they've improved in the last few months and I haven't realised it yet.
I really do appreciate the flexibility and power of the Sun's Java and classpath, but it's unfortunately not under a liberal enough license. So until a relatively comparable open source (GPL or better) version comes around for my systems, I'll just substitute with a more focused approach.
Just to make mention on your last comment:
I use Java because Smalltalk was too slow and too fragmented between different incompatible versions, and because C++ a maintenance nightmare (I find Java's strack traces on exceptions to be a huge benefit).
If you liked some of the ideas behind Smalltalk (like full object-orientation and the "do:" blocks) you might want to give Ruby a try. I really like it; it's very clean and easy to write, and it scales very well (i.e. doesn't get insanely complicated when you write a big project). I agree that C++ is really hard to maintain; or at least, i
After some more careful thought, I concede, Decaf, that you are correct on nearly all points - at least, at this moment in time. For my own part, however, as long as Sun's relationship with Java remains as it is now, I will continue to use Java only when all other more open options are exhausted. And I will also continue to work towards the ultimate success of open source, no matter what others may say, because I believe in its cause and potential. To me, it is more important to do what I think is the Right Thing, than to do what is most convenient.
I would lay claim that Perl and PHP power more online transactions than Java does. Look at the world's most popular forum system, PHPBB. Yes, that runs on PHP. And people don't run LAMJ servers, they run LAMP servers: the P stands for PHP or Perl or Python, all of them open languages. I didn't say that Java is now unpopular in all domains, that is false. But I think that it will degrade because it will not have the ability to adapt like other languages can. If Sun goes down, methinks it would be all over for Java.
And if you want to chide somebody for wanting to overcome the competition, fine. But don't forget that the origin of all open projects is the desire to build a better product, and it's only because we want to be better that we can achieve that. Wanting something is the root cause for it happening. That's not a guarantee, but it's as close as we can get.
Closed products can only compete with open projects if they are significantly superior in quality and available on a sufficient number of platforms. While I think Java fills the latter requirement, it does not the former; it is at least on the same level as equivalent products, and perhaps lower than some others. No amount of marketing can change this: if Java is not sufficiently opened, it will remain on the path to obscurity. Without new ideas being able to add to the product, it will decay.
Why is it that Sun never offers anything for BSD? The FreeBSD community only weeks ago finally got access to official Java binaries, after years of tweaking and weedling from the FreeBSD Foundation. Could it be that Sun has forgotten their heritage, that the only reason they exist is because of BSD Unix? Such arrogance towards the BSD community from a company founded by one of the original BSD developers is astonishing and shameful.
People don't use kernels, people use applications. In other words, the OS is the applications. If the distro gives you the options of KDE or GNOME for desktop management, one important option is missing: None. You know, there are plenty of decent window managers out there that get a lot of use: FVWM2 is my fave, but Fluxbox, BlackBox, WindowMaker... the list goes on and on. And plenty of people use mv/cp/ls rather than Nautilus. The point is, a lot of us still use and work on projects that don't require a boatload of resources.
As far as switching the project to NetBSD goes, I think that'd be a great idea. But one of the things that would make NetBSD so good to fill this position is the fact that you have to choose exactly what to put on it. Don't want GNOME or KDE? Fine, don't pkg_add them; try FVWM2 instead. Unfortunately, an increasing number of Linux distros make that choice for you, and just assume what you want.
Of course, this isn't really about Linux itself, but about the many Linux distributions that have bloat problems, notably Fedora and Suse. These are the guys who put everything, including the kitchen sink, into their basic installs. In other words, the Emacs of Linux distros. Whatever happened to keeping the number of required packages to a minimum? On the other hand, Ubuntu and Slackware manage to install a great OS with small footprints.
BUT... since we're talking about the exotic hardware here, why not give NetBSD a try; it's very small and has great support for all kinds of strange hardware. And it's more free than Linux! It would be very difficult to go wrong with a BSD.
With everything that's been going on lately, it sounds like the American government really wants to take Google down in the war of public opinion. The gov't just keeps trying to make them look worse and worse. And since the American courts typically just allow the gov't to do whatever it wants, they're winning.
Yes, if there's anyone who can stand to make a quick buck off security worries, it's the so-called "security software" businesses themselves. Of course, they all love to insinuate that all operating systems play on the same level field. (So the only way to make sure you're safe is to BUY OUR [poorly written] PRODUCT!) Now, because a couple people came up with a proof-of-concept exploit, they get their chance to say "See, Mac OS X isn't any more secure than Windows is!" But of course, we know that there's hardly any truth behind those statements. Sure, all software systems have holes, and flaws, and bugs. But depending on how the software is written, the threats from those vulnerabilities can be mitigated and even made altogether impotent. I'm not an OS X user, but I do exclusively use operating systems that live in the same family tree, the modern BSD Unixen. And I can state from experience that these systems are inherently more secure than some other commercial operating systems, simply because of the way they are written. (For example, the level of source auditing found in OpenBSD, etc. etc.)
Advanced Programming In The Unix Environment, 2nd Ed: http://www.awprofessional.com/title/0201433079
It takes a lot of work to be a "geek" in any real sense of the word. Learning the ins and outs of math/science/technology/whatever requires time and commitment, which most people can't (or won't) afford. So why go through all the hassle when you can just pretend you're "down with it" and join the crowd. It's just like any other cultural trend: take (more modern) rap music for example. Not everybody can rap like P. Diddy, but if you wear his clothes and represent yourself as if you were part of his culture, you too can become cool.
I would assume (given the content of the articles presented here) that most in the Slashdot community support the Open Source / Free Software movement. This movement is founded in the idea that the more people working on and watching over a project, the better that project will turn out. How is it any FOSS project different from the Internet? With one country controlling the root servers, with one organization under the Department of Commerce controlling the naming system, how can the Internet possibly be anything other than a manifestation of that singular country? One might argue against me, saying that all great FOSS projects have a benevolent dictator, say Linus for example. But Linus can only keep up with his project because he has a group of people to whom he trusts absolutely the responsibility to commit changes, &c. These people are his peers, just as the United States should have peers to aid in harnessing the power of the Internet. The power of control should be distributed. I'm sure the Senator is a freedom-loving man, by his statements; what kind of freedom is controlled by one land alone? That sounds an awful lot like the kind of rule that those who founded this country tried to escape from. It is the principle of both our country, and the idea of democracy itself, to balance power and evenly distribute rights and capabilities. Remember the three branches of government? Checks and balances? What about "all men created equal", with the rights to "life, liberty, and the pursuit of happiness"? So, why should the United States be the only country to control the Internet?
The same can be said about Windows vs. Linux. You're absolutely right; but this thread wasn't about Windows vs. Linux, it was about Linux vs. OpenBSD. And sometimes, you need Linux for functionality more than you need OpenBSD for security.
OpenBSD is great, but a lot of programs run only on Linux. Granted, a lot of these are closed down; but if you need to run the Sun JVM, or Borland's JBuilder (two programs I have to use for a computer science class), you need Linux. Even under FreeBSD's Linux Binary Compatibility, they don't work very good, if at all.
I heard that Acclaim went bankrupt shortly after releasing Burnout 2, and the Burnout distribution rights were purchased by EA. Why EA doesn't release the Burnout series for GameCube, I don't know (they offer plenty of other games for it, so it may be a Criterion thing). The first two were so awesome, it's sad that single-system Nintendo owners like myself haven't gotten a chance to play the latest and greatest.
It sounds heartless, but 'jargon' is absolutely necessary for people who do anything beyond the basics with computers. How else would we communicate precise information? And the truth is, most computer language comes from other contexts, where the phrases are even more established. So if comprehending the difference between kilo-, mega-, and gigabytes is too difficult (even though it's exactly the same for any other measurement dimension, like length or weight!), they should just surf over to Wikipedia and type it in. Professionals can't spend their whole day educating clueless users; they'll have to do it themselves.
I'd honestly expect more from someone who worked at NASA and the DoD. Somehow, the author manages to confuse Unix with UNIX; two very different things. Unix is an idea; UNIX is a product that has earned the name from OpenGroup. Unix is Linux and FreeBSD and even Solaris; UNIX is just one of those things sold by SCO and Sun Microsystems.
It's true that Windows beat UNIX, though not because of technical superiority. Rather, it was the result of ruthless business strategy. The UNIX vendors fought each other (instead of banding together to fight the common enemy that could ruin them all), while Microsoft busily secured the PC platform. No wonder Microsoft rules the home office / small business world these days.
But Windows did not beat Unix. The Unix idea lives on in Linux and BSD, as well as Solaris. And these operating systems are not fighting each other. The 'infighting' among the BSD groups that the author refers to doesn't exist. The BSD groups have always been willing to help in each other's projects. They exist separately because they have different focus areas, not because they hate each other. Sun is opening more of its programs every day to other open source programmers (sure, not Java yet, but let's hold out hope). Linux is willing to share anything and everything with other FOSS projects. How else could you run Linux programs under FreeBSD?
This brings us to another of the author's reasons that UNIX failed - program incompatibility. For UNIX, he's somewhat right. Closed-down programs were un-portable in even the smallest variances between systems. But that's one of the reasons the open source community came into existence all those years ago, to share code so it could be used on more than just your own box. That's why the idea of the portable C compiler sprung forth like light from darkness. That's why the first Unix was written in C on the PDP-11. Portability! And that is exactly why this argument, which only barely applies to UNIX, doesn't apply at all to the modern, open source Unixen, like Linux and FreeBSD.
In short, this article is full of holes. Unix is not UNIX. It didn't take Microsoft to help many UNIX vendors throw themselves over the cliff. The BSDs don't hate each other. C is the language of Unix; C is that language of portability.
Thank you, Ken Thompson.