These machines need to be on the net. If you could query the price, I'm sure "cokewatch.com" would pop up in no time searching for the cheapest vending units, and sending them to your palm-VII so on demand you can locate the best buy close to you.
All I'm saying is, smarter machines will probably not work out like Coke, the public, or even the well informed slashdotters may think. This sort of technology generally results in odd changes in the way people interact with the systems on a day to day basis.
Re:Just what the hell is that supposed to mean?
on
Which BSD?
·
· Score: 1
I didn't mean to suggest that Linux did not run on non-Intel platforms. I'm sorry if that was the impression I got. What I was referring to was the fact that the people working on Linux/Intel seem to have as one of their goals supporting every last bit of hardware ever produced for an WinTel based system. Most of the BSD variants have no interest in doing this, they support what they consider to be a "good" subset of the hardware (good often meaning well performing hardware that supports free software development), and show little interest in other hardware items.
I give the Linux guys credit for running on as much stuff as they do (in both WinTel land and alternative land), it is quite an accomplishment. Unfortuantely, when you talk about server-land, this is relatively unimportant, as server people tend to buy a select group of top-end components, and don't care if anything else is supported. Some people even consider Linux's broad support a drawback in this area, as the more stuff that is being developed, the more stuff that needs to be tested, and proven solid. Yes, I know things can be disabled, and yes, in the end it probably doesn't make a lot of difference.
My FreeBSD perspective.
on
Which BSD?
·
· Score: 5
Everyone will have a different opinion, and they are all right. I'm going to offer my FreeBSD-slanted opinion as one view.
NetBSD
Coke, original formula. Hard to argue with that. NetBSD has a long and noble history. The NetBSD team does a great job of covering the hardware world. No, not the WinTel hardware world, that's Linux. They cover platforms. By running on so many platforms it is a great platform if you have a lot of different (and/or old:-) sorts of hardware. Unfortunately, it is this platform compatability that slows their progress.
I have nothing bad to say about NetBSD. unfortunately, I have nothing good (feature wise) to say about it when it comes to getting real work done. Anything you buy these days has "better" choices that run on it. I will continue to be a big NetBSD supporter though, as it's the only choices for some of my older machines that still deserve a real operating system.
OpenBSD
I'd tell you about it, but then I would have to kill you.:-) Actually, it's not that bad. OpenBSD is security focused, and so they do go a few extra steps in that direction. About 60% of what they do can be done on NetBSD simply by intelligently securing the box. The other 40% is good security add on work.
Most of the good stuff the OpenBSD folks come up with make it into the other BSD's and Linux shortly afterwards, although not all. I'm not sure on security alone OpenBSD is "better", assuming you have a clueful admin who understands the issues.
IMHO the best thing for the BSD community is if the OpenBSD guys and the NetBSD guys could get together. Unfortunately, the inability to do that is the very reason they are apart.
FreeBSD
The FreeBSD folks want to get real work done. Early on, that resulted in an Intel focus, as that was the only affordable platform available. Now the Alpha is included, and hopefully more soon. When they day is done though they are interested in bang-for-the buck, not on RC5 or quake, but applications like web, ftp, and news. Bread and butter network stuff, rooted deep in the Unix world.
This shows in several places. The VM subsystem they implemented several years back was one of the first of it's kind in the free OS world. The port subsystem is an efficient way to distribute and build tools that may still have compile-time dependancies and configuration without creating a packaging nightmare. The installer is simple, clean, fast, and good for the novice and the expert.
Put simply, FreeBSD makes the admin and the machine the most productive when trying to do Internet application "stuff".
Linux
I'll offer my Linux opinion, to complete my perspective. Linux wants to be everything to everybody. As such, it supports more "options" to everything. There are more device drivers, more supported file systems, and more "applications" than any other free unix. In many cases, this is good, but when it comes to getting real work done, it is questionable at best.
The quality of both some of the "supported" hardware and the drivers are to be questioned, but how are you to know what is good, and what is bad? The releases are more frequent, both to fix bugs, and introduce features. There are often all sorts of new things added you don't need that may affect what you're trying to do.
Summary
Any of them will probably do what you want. All of the BSD's have a very different structure than Linux, not only in code, but in how they are designed, built, and released. They all have core teams, rigid code review and testing procedures, and an emphasis on being correct rather than being first, best, or fastest. For the most part, if there is a feature in a released version, it works, reliably. Linux emulation on FreeBSD works like a dream. If RealPlayer G2 and acroread will run fine under it, anything will. The penality for this stability and reliability is that you're doing to have to pick from the "approved" hardware list, and do without some of the wizbang stuff.
Finally, I have one recommendation. Learn the way each OS wants you to do things. Unix is Unix, unless you're an admin or a programmer. The worst thing anyone switching OS's can do is try to impose one OS's / designers view on another. It's usually a poor fit. Just because one OS does something completely different than another does not automatically make it better or worse, what matters is what you are able to do with it at the end of the day.
I believe this argument comes down to a simple choice. Is your goal to advance software, providing better tools for users, or is your goal to make sure that software innovations must be given away for free? There are of course good reasons to do both, but few reasons for a business to give away software.
Personally, I don't believe all software should be free. Sure, most of it should be free, in particular commodity utility programs, but there is a whole set of software that should not be free. There are a lot of highly specialized software products where the users need to pay the costs of development, because without them doing so _no one would write them_. If we don't have an economic vehicle to encourage people to write these programs, they will never get written.
Free software people like to ignore the minor detail that programmers, when left to their own devices, work on projects they _like_. The free software community is rich with the things programmers like to write, operating systems, tools that help people program, tools that make computers easier to use and more productive. The free software movement is virtually devoid of products that are of little use to most programers (eg office suites) or highly specialized programs (there's a ton of free good CAD programs, right?).
So, to get this software we must incent programmers with money. If we have to pay the programmers, business must be able to make money off the software. If a business is going to make money off the software, they will want certian protections in able to lure investors and other non-programmer types. In these cases, the GPL produces problems.
In this case, Sun should be commended for releasing source for free to the public. That mere fact is a huge win for the software community. If other developers want to write compatable programs, no reverse engineering will be necessary, as they can look at the source for file formats. If you find a bug, you can contribute source back to sun to fix it, which will hopefully speed bug fixes. If you need a custom feature, you can make it happen, and still release diffs against the sun source.
This is lightyears ahead of where we are now, and will be a huge win for users and the software community, yet still allow sun to make a profit. Perhaps in the future this will become commodity software that is no longer an important revenue stream for sun, at which time it can be GPL'ed.
What the free software community needs to do now is support Sun's efforts completely. It's the only way to make other vendors see the light.
Even if we assume that the smallest transister is about to be realized in a silicon product, that doesn't mean there aren't a lot of other ways to go. Current chips only begin to explore the 3-D world, often 10 layers thick. It's easy to imagine a chip that is a true cube, 500 layers thick. Even if transisters don't get smaller, that could pack a lot more punch in a processor. Of course, parallel processing, and neural networks of processors are other world to explore. These can come in traditional (eg SMP) forms, or just more intelligent peripherals. Imagine a video card with a P-III 550 on it, that could offload a lot from the main CPU. Moores law will continue well into the future, even if we reach a limit on transister size. I suspect they will find a way around the transister size limit too, with optical or quantum computers.
Network Computers duplicate ideas that have been pushed over and over again in the networked world. First there was the dumb terminal, then there was the X-Termal. These days there is the web / java terminal. They are all some form of network computer. When you get right down to it there is a "right tool for the job" problem. If there is something that is cost effective and gives you the applications you need, use it, by all means. Linux boxes running X, go for it. PC's runing an X server, go for it. Mac's running native applications, go for it. Mainframes with dumb terminals, go for it. The test of a technology is that the system administrators _AND_ the users are able to easily complete their jobs. If you have a system that is easy to manage, and easy to use, you have a winner. Those who are truely techologically clueful will be able to use any system. Those who are not will need something that makes the particular thing they need a computer to do easy, and each platform has their own strengths in that area.
Slashdot Mirror
These machines need to be on the net. If you could query the price, I'm sure "cokewatch.com" would pop up in no time searching for the cheapest vending units, and sending them to your palm-VII so on demand you can locate the best buy close to you.
All I'm saying is, smarter machines will probably not work out like Coke, the public, or even the well informed slashdotters may think. This sort of technology generally results in odd changes in the way people interact with the systems on a day to day basis.
I didn't mean to suggest that Linux did not run on non-Intel platforms. I'm sorry if that was the impression I got. What I was referring to was the fact that the people working on Linux/Intel seem to have as one of their goals supporting every last bit of hardware ever produced for an WinTel based system. Most of the BSD variants have no interest in doing this, they support what they consider to be a "good" subset of the hardware (good often meaning well performing hardware that supports free software development), and show little interest in other hardware items.
I give the Linux guys credit for running on as much stuff as they do (in both WinTel land and alternative land), it is quite an accomplishment. Unfortuantely, when you talk about server-land, this is relatively unimportant, as server people tend to buy a select group of top-end components, and don't care if anything else is supported. Some people even consider Linux's broad support a drawback in this area, as the more stuff that is being developed, the more stuff that needs to be tested, and proven solid. Yes, I know things can be disabled, and yes, in the end it probably doesn't make a lot of difference.
Everyone will have a different opinion, and they are all right. I'm going to offer my FreeBSD-slanted opinion as one view.
NetBSDCoke, original formula. Hard to argue with that. NetBSD has a long and noble history. The NetBSD team does a great job of covering the hardware world. No, not the WinTel hardware world, that's Linux. They cover platforms. By running on so many platforms it is a great platform if you have a lot of different (and/or old :-) sorts of hardware. Unfortunately, it is this platform compatability that slows their progress.
I have nothing bad to say about NetBSD. unfortunately, I have nothing good (feature wise) to say about it when it comes to getting real work done. Anything you buy these days has "better" choices that run on it. I will continue to be a big NetBSD supporter though, as it's the only choices for some of my older machines that still deserve a real operating system.
OpenBSDI'd tell you about it, but then I would have to kill you. :-) Actually, it's not that bad. OpenBSD is security focused, and so they do go a few extra steps in that direction. About 60% of what they do can be done on NetBSD simply by intelligently securing the box. The other 40% is good security add on work.
Most of the good stuff the OpenBSD folks come up with make it into the other BSD's and Linux shortly afterwards, although not all. I'm not sure on security alone OpenBSD is "better", assuming you have a clueful admin who understands the issues.
IMHO the best thing for the BSD community is if the OpenBSD guys and the NetBSD guys could get together. Unfortunately, the inability to do that is the very reason they are apart.
FreeBSDThe FreeBSD folks want to get real work done. Early on, that resulted in an Intel focus, as that was the only affordable platform available. Now the Alpha is included, and hopefully more soon. When they day is done though they are interested in bang-for-the buck, not on RC5 or quake, but applications like web, ftp, and news. Bread and butter network stuff, rooted deep in the Unix world.
This shows in several places. The VM subsystem they implemented several years back was one of the first of it's kind in the free OS world. The port subsystem is an efficient way to distribute and build tools that may still have compile-time dependancies and configuration without creating a packaging nightmare. The installer is simple, clean, fast, and good for the novice and the expert.
Put simply, FreeBSD makes the admin and the machine the most productive when trying to do Internet application "stuff".
LinuxI'll offer my Linux opinion, to complete my perspective. Linux wants to be everything to everybody. As such, it supports more "options" to everything. There are more device drivers, more supported file systems, and more "applications" than any other free unix. In many cases, this is good, but when it comes to getting real work done, it is questionable at best.
The quality of both some of the "supported" hardware and the drivers are to be questioned, but how are you to know what is good, and what is bad? The releases are more frequent, both to fix bugs, and introduce features. There are often all sorts of new things added you don't need that may affect what you're trying to do.
SummaryAny of them will probably do what you want. All of the BSD's have a very different structure than Linux, not only in code, but in how they are designed, built, and released. They all have core teams, rigid code review and testing procedures, and an emphasis on being correct rather than being first, best, or fastest. For the most part, if there is a feature in a released version, it works, reliably. Linux emulation on FreeBSD works like a dream. If RealPlayer G2 and acroread will run fine under it, anything will. The penality for this stability and reliability is that you're doing to have to pick from the "approved" hardware list, and do without some of the wizbang stuff.
Finally, I have one recommendation. Learn the way each OS wants you to do things. Unix is Unix, unless you're an admin or a programmer. The worst thing anyone switching OS's can do is try to impose one OS's / designers view on another. It's usually a poor fit. Just because one OS does something completely different than another does not automatically make it better or worse, what matters is what you are able to do with it at the end of the day.
Good luck with whatever you try.
I believe this argument comes down to a simple choice. Is your goal to advance software, providing better tools for users, or is your goal to make sure that software innovations must be given away for free? There are of course good reasons to do both, but few reasons for a business to give away software.
Personally, I don't believe all software should be free. Sure, most of it should be free, in particular commodity utility programs, but there is a whole set of software that should not be free. There are a lot of highly specialized software products where the users need to pay the costs of development, because without them doing so _no one would write them_. If we don't have an economic vehicle to encourage people to write these programs, they will never get written.
Free software people like to ignore the minor detail that programmers, when left to their own devices, work on projects they _like_. The free software community is rich with the things programmers like to write, operating systems, tools that help people program, tools that make computers easier to use and more productive. The free software movement is virtually devoid of products that are of little use to most programers (eg office suites) or highly specialized programs (there's a ton of free good CAD programs, right?).
So, to get this software we must incent programmers with money. If we have to pay the programmers, business must be able to make money off the software. If a business is going to make money off the software, they will want certian protections in able to lure investors and other non-programmer types. In these cases, the GPL produces problems.
In this case, Sun should be commended for releasing source for free to the public. That mere fact is a huge win for the software community. If other developers want to write compatable programs, no reverse engineering will be necessary, as they can look at the source for file formats. If you find a bug, you can contribute source back to sun to fix it, which will hopefully speed bug fixes. If you need a custom feature, you can make it happen, and still release diffs against the sun source.
This is lightyears ahead of where we are now, and will be a huge win for users and the software community, yet still allow sun to make a profit. Perhaps in the future this will become commodity software that is no longer an important revenue stream for sun, at which time it can be GPL'ed.
What the free software community needs to do now is support Sun's efforts completely. It's the only way to make other vendors see the light.
Even if we assume that the smallest transister is about to be realized in a silicon product, that doesn't mean there aren't a lot of other ways to go. Current chips only begin to explore the 3-D world, often 10 layers thick. It's easy to imagine a chip that is a true cube, 500 layers thick. Even if transisters don't get smaller, that could pack a lot more punch in a processor. Of course, parallel processing, and neural networks of processors are other world to explore. These can come in traditional (eg SMP) forms, or just more intelligent peripherals. Imagine a video card with a P-III 550 on it, that could offload a lot from the main CPU. Moores law will continue well into the future, even if we reach a limit on transister size. I suspect they will find a way around the transister size limit too, with optical or quantum computers.
Network Computers duplicate ideas that have been pushed over and over again in the networked world. First there was the dumb terminal, then there was the X-Termal. These days there is the web / java terminal. They are all some form of network computer. When you get right down to it there is a "right tool for the job" problem. If there is something that is cost effective and gives you the applications you need, use it, by all means. Linux boxes running X, go for it. PC's runing an X server, go for it. Mac's running native applications, go for it. Mainframes with dumb terminals, go for it. The test of a technology is that the system administrators _AND_ the users are able to easily complete their jobs. If you have a system that is easy to manage, and easy to use, you have a winner. Those who are truely techologically clueful will be able to use any system. Those who are not will need something that makes the particular thing they need a computer to do easy, and each platform has their own strengths in that area.