So you support the mass murder of chinese citizens by their own government?
Before you are so hasty to bash the Chinese, remember that they didn't elect their own government. I think that it's good ( for the Chinese people ) that the government there are adopting linux, even though I don't like their government
The remedy we need is to put the Windows source code, including future versions, in the public domain.
Firstly, this is more or less an act of theft. Secondly, allowing competing versions of windows to enter the market is hardly what the industry wants or needs. I don't think this would help anyone except for Microsoft's competitors. Personally, I think a breakup is a more intelligent solution.
This is absurd. NO-ONE produces products that are completely standards compliant. For example, how many UNIX variants ship a working ANSI C++ compiler ? ie one with decent support for STL, and template functions ? ( even egcs is borderline here ) Have you ever noticed that linux ships several "non standard" GNU utilities ( such as bash in the place of sh. No, bash started as sh does *not* behave exactly like the bourne shell ) ?
So why should MS ? The danger of this is that it would require MS's products to be designed by committees, essentially taking design decisions out of MS's hands.
That wouldn't be an ideal solution, because they could strike deals with each other, or at least continue what they've been doing.
This isn't entirely correct. Each division would be obliged to act in it's own best interests. The OS division would't have any vested interest in giving "MS Apps group" preferential treatment, because the CEO of the "OS division" would no longer be accountable to shareholders in the Applications division. The "OS division" would be doing its shareholders a disservice if they snubbed a competitor to the "apps division".
Or rather than that, they simply need to document all the features used in their file formats.
ONe could argue that they already do. They release their document formats, the only subject of debate is how well they are documented. In any case, I don't think this would help. It's hard to enforce, and it's not clear that they aren't already doing this.
I use it in the simplest sense: to mean an application that a desktop user would use. This includes anything from Pine to Word Perfect to LaTeX. Linux has the advantage that the applications that come with it are all very much up to date. It also has some applications that are unavailable on OpenBSD. Whether or not this is an issue depends on which applications you need/want to use.
As far as desktop environments are concerned, I prefer KDE ( or more to the point, my users do ). Fortunately, there are OpenBSD binaries for KDE now.
The ports collection is definitely better on FreeBSD, though I don't know what the exact number is. I'll fess up and admit that OpenBSD is the only BSD I have root access to.
It seems to me that it's more important for a machine that has many users to be fast
That depends on the number of concurrent users, as well as what those users are doing. If they are all compiling and running emacs, you have a point. However, if the machine is a webserver getting 10000 hits a day or less, then it doesn't need to be very fast. Moreover, there are several applications for which OpenBSD's file system will perform just fine.
As for file system speed, what do you mean?
I guess I can only compare OpenBSD to linux. OpenBSD doesn't cache directory writes, which makes it very slow for recursive operations that require directory writes ( tar xvzf , cp -a , rm -rf ) Unfortunately, I can't do a fair benchmark now because my OpenBSD box ( which used to run linux ) is a Pentium 133 and my linux box is a Pentium II. But I'd suggest you should be able to verify this. I urge you to try benchmarking any or all of tar xvzf, cp -a and rm -rf on a directory containing several files ( thousands ). I bet you a nickel that ext2 wins (-;
OpenBSD is licensed under the BSD license, so porting features to Linux (GPL) without raising licensing issues can be tricky.
Not true. It is the other way around. There is nothing stopping you releasing BSD code under the GPL ( AFAIK ). The GPL is the more restrictive license, and it is bringing GPL'd software into a BSD that proves difficult.
The main difficulty with adding OpenBSD's features to linux is that most distributors have their hands tied by export regulations. So all of OpenBSDs built in cryptography cannot be built in to any distribution that is developed in the USA.
Precisely what features are desirable in a "server"?
Security, reliability, availability of server apps, an OOTB config that is secure, reliable and has the right server apps. Performance is sometimes important, but sometimes ( ie a small server ) it's close to irrelevant.
What features are desirable in a "workstation"?
Desktop applications, desktop applications, and desktop applications. Performance. Of course, security, stability and reliability are still important, but not as critical. The machine is quite possibly behind a firewall which makes security less critical. The machine is not running critical services which makes reliability less critical. Stability is less important than supporting the latest hardware.
Does optimizing for one of these environments pessimize-- or at least compromise--the other situation?
yes, sometimes it can. For example, OpenBSD's default file system is IMO unacceptably slow for a workstation. However, on a critical server, it's reliability could be an advantage. On a workstation, I'd rather use a filesystem with write caching. Another example: if the developers pay more attention to putting in as much software as possible into a distribution, rather than fixing the software that they already have, the result is a nice desktop system that is possibly full of security holes.
Of course, a workstation OS can benefit from enhanced security and reliability all other things being equal. However, in practice, there is often some kind of tradeoff.
Just as you want a solid, sane, robust system for a computer that provides services for an entire department, so too do you wish the same coherence and correctness on my very own computer that you are the principle user of.
All the "correctness" in the world will not help if you can't run the applications that you need to use. For example, OpenBSD's "correctness" isn't much help to java developers.
Certainly, at this point, OpenBSD suffers as a workstation OS for two simple reasons: the fact that the availability of apps is small ( the ports collection is tiny compared to FreeBSD ) and the hardware support is relatively small. On the other hand, while linux leads the way in terms of apps, it lacks OpenBSD's security features.
Cheapbytes give you the option of donating. One of the things I like aboutr Cheapbytes is that they seem genuinely concerned about free software. The reason I purchased from there is because I visit the site a lot. I'll probably get OpenBSD 2.6 from openbsd.org though.
I am a reasonably experienced linux user. I guess I'd call myself intermediate, leaving the term "advanced" for the "real programmers".
I decided some time back that it would be fun to experiment with OpenBSD. I was drawn primarily by it's crypto software. I was installing it on a machine that I tend to use more as a server than anything else. So desktop friendliness was not a major issue.
So first came the install. I ordered my $2- Cheapbytes CD, which came with an installation walk-through. This walk through made it pretty easy. I had a hiccup with my large disk drive ( due to bad bios configuration ) but a post to comp.unix.bsd.openbsd.misc fixed that pretty quickly. The partitioning procedure using the cryptic disklabel tool would have been hell without the walkthrough. However, i just did ( more or less ) what the walk through said, and it went OK.
Which raises another point -- I was surprised to find that the help on Usenet for OpenBSD is on par with usenet linux support. Far from being a bunch of obnoxious RTFM'ers, the OpenBSD crowd are by and large very helpful. I felt honoured that Theo De Raadt himself responded to one of my posts. Regarding support, the "OpenBSD FAQ" is also excellent. It is really more like a users manual than an FAQ. I highly recommend that anyone planning on installing openBSD get a copy of this prior to installation.
Once I had finished the install, I had my openBSD system up and running. I discovered a few things:
First, I was somewhat surprised that the inetd services don't go via TCP wrappers by default. I had to edit inetd.conf to make them do this. I was awfully confused for a little while regarding the fact that my hosts.deny settings ( ALL:ALL ) were not honoured. So I fixed inetd.
What is nice about the default setup is that software such as sudo, skey and kerberos is installed by default. They will be shipping ssh with it in the near future ( 2.6 ), see http://www.openbsd.org/crypto.html#ssh. Until recently, they've had obstructions to shipping this, such as patents. They are actively hacking ssh to remove these obstacles. Crypto is "integrated" into the system. For example, crypt() has built in blowfish encryption ( which is used to encrypt passwords ) See http://www.openbsd.org/crypto.html for more info.
The system also uses shadow passwords out of the box. The ports collection makes it easy to install any other secure software you might want, such as cops, ssh, rsaref, among other things. Just CD to the right directory and type "make install" and openBSD automatically installs the package, *and* looks after any dependencies -- so "make install" always works, even if you don't have some of the required packages to begin with. The ports collection is lean in terms of desktop applications, but contains a good collection of server apps.
However, it's not ideal as a desktop system. The file system is slow ( though very stable ), and the ports collection is somewhat limited compared to FreeBSD and NetBSD. It also trails FreeBSD in hardware support.
Overall, I'd highly recommend it for a user familiar with linux ( in particular, someone not scared of command lines ) who wants to set up a secure server on low end hardware.
I think it's high time that the linux distributors started shipping crypto, even if that means shipping from Canada ( as OpenBSD does ). It would also be nice to see some more sensible default settings on the linux distributions.
OpenBSD is an excellent choice as a solid server OS for low end hardware. However, compared to FreeBSD, it's not really ready for the desktop yet ( FreeBSD has a much larger ports and packages collection. Also, OpenBSD's filesystem, while rock solid is also sloooowwwww ). The nice thing about OpenBSD is you can fire it up and forget about it. You don't need to worry about the "patch of the week" like you do with linux.
The Cheapbytes CD comes with a very clear, step by step guide regarding a sample install of OpenBSD. Also, their "FAQ" ( which is more like a small user's manual ) is excellent. However, I agree that the disklabel is cryptic.
Why is slashdot so slow ? This sucks boulders through straws ( really, the page loads like a boulder getting sucked through a straw ) slashdot is one of the main sites the non-linux users see. And it's dog slow. I can rarely if ever get into the comments. I don't think this is a good advertisement for linux. I mean we all like to chuckle when microsoft.com is down or just dog slow. But right now, the egg is in our face.
And I think Linux and FreeBSD are the wastes of time here. How the hell is Sun gonna make make money from a free OS? Think about it.
The original poster was FUDding Be, and apparently, your response is to post more FUD (about linux).
For the zillionth time, NO, LINUX IS NOT JUST ABOUT FREE BEER. People can and do spend money on linux.
Personally, I think this is mildly good news, but I still feel that Java doesn't live up to it's portability promise. Compare it to python, Tcl and perl. It's not even in the same league. In fact most apps written in C are more portable than java apps, because there are still several platforms which barely run Java or don't run it very well.
Oh yeah... another thing. Even if an entire office was using an office suite other than MS Office, would their clients be happy when they sent documents to that office that couldn't be read?
There are a ton of formats they could use that MSOffice could read just fine, such as RTF and Office 95 ( which just about every office suite can save in ). Of course, there are also standards based alternatives such as pdf.
Java, "platform independent" ??? Well, that's news to me.
One of the big problems with Java is that while it's architecture is platform independent, that doesn't really help much if Java itself doesn't run on many platforms.
And the truth is, java does not run on so many platforms. For example, I can't run Java apps on OpenBSD because there are no decent java implementations for it. Even Linux suffers from second rate java support ( partly because Sun refuse to support linux versions of their software ). IMO, Sun dropped the ball by not getting behind a JDK that would work *well* on any platform. Perhaps "platform independent" is another way of saying "windows or Solaris".
In Australia, ( last I checked ), the classifications were mildly confusing because movies and pictures were classified differently.
As far as movies are concerned, R is softcore. The X rated movies are non-violent hardcore. They are illegal in each state, but can be acquired by mail-order from the capital city ( Canberra . )
As for print media, there are two types of R rated publications. R-category-1 is soft ( for example, it includes most "detailed" nudie pics ), while R-category-2 consists of similar content to the "X-rated" movies.
I am not clear on how online content is classified ( or even if the censors have made any clear rulings ).
Yes, that's one. However, it's not clear that the BSDs really benefited from the fork. There are arguments both ways. Moreover, these projects ( emacs and BSD ) would have still been wildly successful without a fork, so it's not clear that the forks really played a big part in the success of the prjects. I don't think there are that many really good examples.
Slashdot Mirror
Before you are so hasty to bash the Chinese, remember that they didn't elect their own government. I think that it's good ( for the Chinese people ) that the government there are adopting linux, even though I don't like their government
Cheers,
Firstly, this is more or less an act of theft. Secondly, allowing competing versions of windows to enter the market is hardly what the industry wants or needs. I don't think this would help anyone except for Microsoft's competitors. Personally, I think a breakup is a more intelligent solution.
So why should MS ? The danger of this is that it would require MS's products to be designed by committees, essentially taking design decisions out of MS's hands.
This isn't entirely correct. Each division would be obliged to act in it's own best interests. The OS division would't have any vested interest in giving "MS Apps group" preferential treatment, because the CEO of the "OS division" would no longer be accountable to shareholders in the Applications division. The "OS division" would be doing its shareholders a disservice if they snubbed a competitor to the "apps division".
ONe could argue that they already do. They release their document formats, the only subject of debate is how well they are documented. In any case, I don't think this would help. It's hard to enforce, and it's not clear that they aren't already doing this.
Looks like it was probably based on this. It's very similar, but slightly different.
I use it in the simplest sense: to mean an application that a desktop user would use. This includes anything from Pine to Word Perfect to LaTeX. Linux has the advantage that the applications that come with it are all very much up to date. It also has some applications that are unavailable on OpenBSD. Whether or not this is an issue depends on which applications you need/want to use.
As far as desktop environments are concerned, I prefer KDE ( or more to the point, my users do ). Fortunately, there are OpenBSD binaries for KDE now.
The ports collection is definitely better on FreeBSD, though I don't know what the exact number is. I'll fess up and admit that OpenBSD is the only BSD I have root access to.
It seems to me that it's more important for a machine that has many users to be fast
That depends on the number of concurrent users, as well as what those users are doing. If they are all compiling and running emacs, you have a point. However, if the machine is a webserver getting 10000 hits a day or less, then it doesn't need to be very fast. Moreover, there are several applications for which OpenBSD's file system will perform just fine.
As for file system speed, what do you mean?
I guess I can only compare OpenBSD to linux. OpenBSD doesn't cache directory writes, which makes it very slow for recursive operations that require directory writes ( tar xvzf , cp -a , rm -rf ) Unfortunately, I can't do a fair benchmark now because my OpenBSD box ( which used to run linux ) is a Pentium 133 and my linux box is a Pentium II. But I'd suggest you should be able to verify this. I urge you to try benchmarking any or all of tar xvzf, cp -a and rm -rf on a directory containing several files ( thousands ). I bet you a nickel that ext2 wins (-;
Cheers,
Not true. It is the other way around. There is nothing stopping you releasing BSD code under the GPL ( AFAIK ). The GPL is the more restrictive license, and it is bringing GPL'd software into a BSD that proves difficult.
The main difficulty with adding OpenBSD's features to linux is that most distributors have their hands tied by export regulations. So all of OpenBSDs built in cryptography cannot be built in to any distribution that is developed in the USA.
Security, reliability, availability of server apps, an OOTB config that is secure, reliable and has the right server apps. Performance is sometimes important, but sometimes ( ie a small server ) it's close to irrelevant.
What features are desirable in a "workstation"?
Desktop applications, desktop applications, and desktop applications. Performance. Of course, security, stability and reliability are still important, but not as critical. The machine is quite possibly behind a firewall which makes security less critical. The machine is not running critical services which makes reliability less critical. Stability is less important than supporting the latest hardware.
Does optimizing for one of these environments pessimize-- or at least compromise--the other situation?
yes, sometimes it can. For example, OpenBSD's default file system is IMO unacceptably slow for a workstation. However, on a critical server, it's reliability could be an advantage. On a workstation, I'd rather use a filesystem with write caching. Another example: if the developers pay more attention to putting in as much software as possible into a distribution, rather than fixing the software that they already have, the result is a nice desktop system that is possibly full of security holes.
Of course, a workstation OS can benefit from enhanced security and reliability all other things being equal. However, in practice, there is often some kind of tradeoff.
Just as you want a solid, sane, robust system for a computer that provides services for an entire department, so too do you wish the same coherence and correctness on my very own computer that you are the principle user of.
All the "correctness" in the world will not help if you can't run the applications that you need to use. For example, OpenBSD's "correctness" isn't much help to java developers.
Certainly, at this point, OpenBSD suffers as a workstation OS for two simple reasons: the fact that the availability of apps is small ( the ports collection is tiny compared to FreeBSD ) and the hardware support is relatively small. On the other hand, while linux leads the way in terms of apps, it lacks OpenBSD's security features.
Especially when you have the src.rpm packages, and you just need to customise the spec then do an rpm -bb ...
However, OpenBSD is more than just a minimal linux.
Cheapbytes give you the option of donating. One of the things I like aboutr Cheapbytes is that they seem genuinely concerned about free software. The reason I purchased from there is because I visit the site a lot. I'll probably get OpenBSD 2.6 from openbsd.org though.
I decided some time back that it would be fun to experiment with OpenBSD. I was drawn primarily by it's crypto software. I was installing it on a machine that I tend to use more as a server than anything else. So desktop friendliness was not a major issue.
So first came the install. I ordered my $2- Cheapbytes CD, which came with an installation walk-through. This walk through made it pretty easy. I had a hiccup with my large disk drive ( due to bad bios configuration ) but a post to comp.unix.bsd.openbsd.misc fixed that pretty quickly. The partitioning procedure using the cryptic disklabel tool would have been hell without the walkthrough. However, i just did ( more or less ) what the walk through said, and it went OK.
Which raises another point -- I was surprised to find that the help on Usenet for OpenBSD is on par with usenet linux support. Far from being a bunch of obnoxious RTFM'ers, the OpenBSD crowd are by and large very helpful. I felt honoured that Theo De Raadt himself responded to one of my posts. Regarding support, the "OpenBSD FAQ" is also excellent. It is really more like a users manual than an FAQ. I highly recommend that anyone planning on installing openBSD get a copy of this prior to installation.
Once I had finished the install, I had my openBSD system up and running. I discovered a few things:
First, I was somewhat surprised that the inetd services don't go via TCP wrappers by default. I had to edit inetd.conf to make them do this. I was awfully confused for a little while regarding the fact that my hosts.deny settings ( ALL:ALL ) were not honoured. So I fixed inetd.
What is nice about the default setup is that software such as sudo, skey and kerberos is installed by default. They will be shipping ssh with it in the near future ( 2.6 ), see http://www.openbsd.org/crypto.html#ssh. Until recently, they've had obstructions to shipping this, such as patents. They are actively hacking ssh to remove these obstacles. Crypto is "integrated" into the system. For example, crypt() has built in blowfish encryption ( which is used to encrypt passwords ) See http://www.openbsd.org/crypto.html for more info.
The system also uses shadow passwords out of the box. The ports collection makes it easy to install any other secure software you might want, such as cops, ssh, rsaref, among other things. Just CD to the right directory and type "make install" and openBSD automatically installs the package, *and* looks after any dependencies -- so "make install" always works, even if you don't have some of the required packages to begin with. The ports collection is lean in terms of desktop applications, but contains a good collection of server apps.
However, it's not ideal as a desktop system. The file system is slow ( though very stable ), and the ports collection is somewhat limited compared to FreeBSD and NetBSD. It also trails FreeBSD in hardware support.
Overall, I'd highly recommend it for a user familiar with linux ( in particular, someone not scared of command lines ) who wants to set up a secure server on low end hardware.
The install isn't that bad. But you'd better RTFM ... the disklabel tool is hardly easy or intuitive unless you have some docs to walk you through.
OpenBSD is an excellent choice as a solid server OS for low end hardware. However, compared to FreeBSD, it's not really ready for the desktop yet ( FreeBSD has a much larger ports and packages collection. Also, OpenBSD's filesystem, while rock solid is also sloooowwwww ). The nice thing about OpenBSD is you can fire it up and forget about it. You don't need to worry about the "patch of the week" like you do with linux.
The Cheapbytes CD comes with a very clear, step by step guide regarding a sample install of OpenBSD. Also, their "FAQ" ( which is more like a small user's manual ) is excellent. However, I agree that the disklabel is cryptic.
The original poster was FUDding Be, and apparently, your response is to post more FUD (about linux).
For the zillionth time, NO, LINUX IS NOT JUST ABOUT FREE BEER. People can and do spend money on linux.
Personally, I think this is mildly good news, but I still feel that Java doesn't live up to it's portability promise. Compare it to python, Tcl and perl. It's not even in the same league. In fact most apps written in C are more portable than java apps, because there are still several platforms which barely run Java or don't run it very well.
MS Office can read/write formats besides its own native format. There's no reason why they need to exchange in the same format they use internally.
There are a ton of formats they could use that MSOffice could read just fine, such as RTF and Office 95 ( which just about every office suite can save in ). Of course, there are also standards based alternatives such as pdf.
One of the big problems with Java is that while it's architecture is platform independent, that doesn't really help much if Java itself doesn't run on many platforms.
And the truth is, java does not run on so many platforms. For example, I can't run Java apps on OpenBSD because there are no decent java implementations for it. Even Linux suffers from second rate java support ( partly because Sun refuse to support linux versions of their software ). IMO, Sun dropped the ball by not getting behind a JDK that would work *well* on any platform. Perhaps "platform independent" is another way of saying "windows or Solaris".
As far as movies are concerned, R is softcore. The X rated movies are non-violent hardcore. They are illegal in each state, but can be acquired by mail-order from the capital city ( Canberra . )
As for print media, there are two types of R rated publications. R-category-1 is soft ( for example, it includes most "detailed" nudie pics ), while R-category-2 consists of similar content to the "X-rated" movies.
I am not clear on how online content is classified ( or even if the censors have made any clear rulings ).
Cheers,
Yes, that's one. However, it's not clear that the BSDs really benefited from the fork. There are arguments both ways. Moreover, these projects ( emacs and BSD ) would have still been wildly successful without a fork, so it's not clear that the forks really played a big part in the success of the prjects. I don't think there are that many really good examples.