And equally importantly this shows the impossibility of eliminating government. The East India companies became de facto governments, even if they were not de jure governments. But rather than being representative governments that had the interests of their citizens in mind, they were imperialistic institutions that exploited the people under their "care" ruthlessly. There's every reason to think that big businesses would do exactly the same thing today if we were to relax control over them.
A settlement isn't something that can be appealed, as far as I know; and even if it could be, I'd bet certainly not by an uninvolved party.
As somebody else pointed out, SCO is USL's successor in interest to the SysV code, so they actually are an interested party in the BSD case. Beyond that, they wouldn't actually be attacking the settlement, per se. Instead they'd be attacking the behavior of the other settling party by claiming that they weren't living up to their side of the bargain.
This might actually be a vaguely clever move. The settlement in the BSD case is under seal, so an ordinary person won't actually know exactly what its terms are, and the other side isn't allowed to disclose the exact terms. That makes it a perfect FUD target. SCO is free to imply all sorts of things about the settlement so long as they don't say anthing specific. They can even say that they'd love to specify exactly what the BSDers are doing wrong but they're forbidden to by the secrecy around the settlement. Nobody can rebut them without actually spelling out the terms of the settlement, which would be illegal. It might even have been effective if SCO hadn't shown itself to be all bark and no bite by now.
It seems to me that this is just more evidence that computer systems will wind up looking like biology. First we had viruses and similar infectious things. Now people are trying to create the machine equivalent of an immune system. The problem is that in the process they're likely to rediscover all of the problems that our immune system causes as well as the benefits.
This particular case is quite similar to allergies in the natural immune system. It's an overly aggressive response to an essentially harmless signal. The big problem is that virus and worm scanners are going to be succeptible to the computer equivalent of autoimmune disease; they'll start thinking that essential system files have been corrupted and try to wipe out something really important. I just hope they never develop the computerized equivalent of leukemia.
Lawyer: do you believe all software should be given away for free?
RMS: If by "free" you mean, everyone should be able to use, modify, and redistribute software without fear of breaking the law, then yes.
That's not what RMS would say, though. RMS has always tried very hard to separate the issue of software freedom from the issue of price, and I can't imagine that he'd pass up the opportunity to dismantle the whole zero price argument. He'd say something like:
RMS: No. That's a common misperception, but the Free Software Foundation encourages people to sell software for as much as they wish or can. The point of Free Software is not one of price but of freedom- the freedom to use, modify, and redistribute the software. If you have those freedoms, the software is Free no matter what price you pay for it.
It's surprising that they didn't include Bruce Perens and/or ESR in their list.
There are much better reasons for talking to Linus and RMS than there are for talking to ESR or Bruce Perens. Linus obviously knows more than anyone else in the world about what code is in the Linux kernel and how it got there. Since that's the single most important factual issue in the case, it shouldn't be at all surprising that SCO wants to ask him questions. Similarly, RMS is an expert on the GPL, which is clearly relevant to IBM's claims of GPL infringement by SCO. Both ESR and Bruce Perens are vocal advocates of Open Source Software, but they lack the depth of case specific knowledge that makes Linus and RMS such good targets for the subpoena.
I strongly suspect that part of the reason that Stallman has such strong views about precision of terminology is precisely because he's been working on the legal side of things for so long. The GPL is as much about hacking the legal system as Linux (err, GNU/Linux) is about hacking computer systems. To create a hack as elegant as the GPL, it's necessary to be pretty well versed in the medium you're hacking, and it seems that a fair bit has worn off on RMS.
Nope. If you want to complain that it's no longer a "justice system" you'd have to prove that it used to be one. Most of the complaints about the legal system- that it gives too much advantage to people with deep pockets, that it lets people sue with little or no cause, etc.- have always been true. At the same time, the current system is clearly better in some respects; a good example is the ability of racial minorities to get a fair trial.
I think some of the strongest arguments that IBM has are (1) the existance of the side memo that lets them off the hook, if that memo exists (big companies misplace legal documents -- it's entirely possible that SCO could have bought the contract and never been given a copy of such a side agreement),
SCO gets no help here. The side letter was given by SCO as an Exhibit C to their original complaint against IBM. Ammendment X was given as Exhibit D. SCO clearly knew about the modifications to their licensing agreement with IBM since they launched the suit. They are, contrary to all reason, trying to repudiate those modifications to the agreement, but they certainly can't claim lack of knowledge or fabrication on the part of IBM.
It looks at least a tiny bit more refined than that. When I run egrep -ril 'smp|jfs|rcu' * | wc -l on the kernel source in my Fedora Core 1 system, for instance, I get a total of 2437 files, while SCO only mentioned a few hundred. That also misses some of the files that SCO did mention, like arch/i386/kernel/mca.c- not surprising considering that particular file originated back in the mid 90s and deals mostly with some crufty old hardware. It mentions IBM, though, since MCA was originally an IBM design, and that's probably why it was chosen.
While I don't consider myself a vigilante, I do use both the SPEWS and Spamcop blacklists for my home server.
And the fact that it's your home server is exactly the difference between being a user and being a vigilante. It's fine to say "I trust SPEWS and Spamcop to blacklist only people I don't want to receive mail from" when you apply it to your own mail. If the blacklisters make a mistake (or decide to change their policies in unfortunate ways without telling you) you're only hurting yourself, and you've made your decision with knowledge of that risk. It's not OK to make that decision for somebody else, at least without telling them what you're doing and giving them a chance to opt out. If you do apply a blacklist to them without their consent, you risk blocking mail that they want without telling them.
Bayesian filtering is very good at removing the annoyance of having your mailbox fill up with spam. Unfortunately, those are just the symptom of a deeper problem, and one that Bayesian filtering does little to solve. Spam causes a number of infrastructure problems because it creates large volumes of traffic and takes up enormous amounts of space on ISPs' mail servers. Filtering means that you don't see the problem, but it doesn't help the ISPs one bit. And- the real problem- it doesn't do much to discourage spammers from sending out their messages, because the people who care enough about spam to set up filters are smart enough not to buy anything from a spammer in the first place.
If you really want to shut the spammers down, you need to kill their profit margin. You can do that two ways. One is to take away their customers. ISP-level filtering (which prevents spam from getting to gullible recipients) can help there, as can education programs. The other is to increase their costs. Forcing them to jump from service provider to service provider, implementing some kind of pay-per-mail system, or letting users sue spammers can all work on that front.
Yeah, because blacklisting has been so effective thus far, we just need to do more of it. Yeah, right. Blacklisting is basically playing a game of whack-a-mole; it makes things a bit less convenient for spammers, but doesn't seem to be doing them serious harm. OTOH, crippling the email of innocent bystanders who happen to share IP blocks with spammers seems a rather steep price to pay for something that does very little to stop spam.
Spam is a tough problem, and it's going to take more than just vigilante action to deal with it. What's needed is a two pronged approach. One prong is legal and is being followed fairly well; pass laws that make spamming illegal. The other prong, which is still under development, is to make technical changes to email so that spammers can't hide their addresses. Neither one will succeed alone- laws can't help as long as spammers can hide, and making spammers stand still won't help if there's no legal recourse against them- but the combination of the two should help a lot.
Bleah. Straight to the wax, man, straight to the wax! If there's no need for mixing- like a symphony performance- you can record directly from the microphones to the Vinyl Master. If it's done right it can sound really fabulous, though I suspect that a lot of that is that the sound never gets mangled in the mixing process, rather than that the vinyl is particularly great. A straight to high-quality digital recording would probably sound just as good.
When ever I need something....if I know the name of the program, just do emerge and it grabs the code, takes care of all the dependencies, grabs any other progs it needs...voila, in minutes, I have the program compiled for my system, and with the flags I want for it.
Unfortunately it's only "within minutes" if the program you're interested in is comparatively small. Just try to emerge the latest version of Mozilla, or KDE, or GNOME, or something substantial like that and see if it takes minutes. More likely it will take hours. If you're talking about installing a whole system from scratch, I can easily download and install all 3 Fedora disks faster than I could get Gentoo to the point of having X compiled.
I'm quite happy with the new kernel version, since it supports DMA mode on my new SATA drive, while the kernel in RH9 would occasionally hang if you tried to turn DMA on. Not a huge thing, unless you happen to have an affected drive, in which case it makes a world of difference. There are some other kernel changes that should make life nicer for laptop users.
Some other random changes that I notice in their release notes:
Improved ACPI support
Include new Bitstream Vera fonts
Added D-BUS
Added DVD+RW tools
Moved from GNOME 2.2 to 2.4
They've also removed a number of packages that are either no longer necessary (e.g. QT2, LPRng) or have licensing issues (e.g. pine, some aspell packages).
One other big change is that they're moving from using ASCII to UTF8 wherever they can. That's one of those things that's kind of annoying while the switch is taking place (they mention that this could cause problems with telnet and ssh, since they don't explicitly negotiate the encoding) but will presumably make everything easier in the long run.
I really don't understand the general Linux user's need to get the latest and greatest at all times. Most of us ran windows for years, and simply waited for Microsoft to say "Ok, it's ready, here you go" every two or more years before upgrading.
I can think of two possible reasons for this. One is that part of the reason that people moved to Linux in the first place is because they were so tired of the "wait for the vendor to put out his next version" process. Experiences like Windows ME are bound to do that to people. One of the cool things about Free Software is that you can upgrade when enough new features have come out that you think its worthwhile, not when your vendor decides for you. A system like Debian, where it can be a long time before the cool new stuff percolates down to stable, can be frustrating for people who have moved specifically to get faster access to that cool new stuff.
Another important aspect is the speed with which that cool new stuff actually becomes available. When you look at it, the rate of development in the Windows world is a lot slower than in Linux. Admittedly, some of that is because the Windows software is more mature and there's less room for improvement, but no matter the reason for it, the pace of change is clearly much faster. That means that there's more motivation to update. Two years ago in Windows, I was running the same operating system (Win2000) and almost the same applications software (Office 97->2000) that I am today. In Linux, though, I've moved from Gnome 1.4 to Gnome 2.4, which is a huge upgrade both in the desktop environment and the applications that go with it. That rapid pace of development makes it much more reasonable to care about upgrading frequently.
It shouldn't be a bandwidth problem. Groklaw is now hosted by Ibiblio (AKA Metalab AKA Sunsite), which has more bandwidth than god. If they're having problems it's probably their servers lacking sufficient horsepower to run a dynamic site when hit this hard, rather than their pipe being too thin.
Thanks, but I want a distribution, not just a core.
Fedora Core is rather poorly named. It covers roughly the same territory as the standard Redhat did, so if that was adequate for you then Fedora Core should be, too. It's also designed to be more friendly to external repositories (which makes sense for a community project) so there should be sources for most things that aren't on the 3 CDs of the standard install.
And, as people have repeatedly pointed out, there's actually more protection in open source than there would be in typical proprietary software. After all, with open source (meaning only that the code is inspectable, not necessarily that it's hackable) anyone who wants to check it for violations may do so. If the code isn't available, finding violators is that much more difficult. You really have to wonder if the ability to hide how things are done isn't just fooling people into thinking that closed source is better for protecting IP.
SCO isn't saying that IBM has to tell them what code was contributed just who did it how and when.
But that isn't what they're asking for. If you look at their question, they're asking for everything that anyone at IBM ever released without confidentiality agreements. They don't even restrict the request to things that happened since IBM licensed SysV, or to people associated with Linux, or anything. That's a grossly overbroad request, and there's no reason that IBM should have to answer it. The process really should go something like:
SCO tells IBM which code in Linux it thinks has been included in violation of its contracts.
IBM tells SCO who, if anyone, at IBM was responsible for releasing that particular code.
Both sides start looking in detail about how the code was release, how it relates to SysV, etc. SCO is allowed to take depositions from IBM engineers who contributed the code in question.
That's the reasonable way for things to proceed. Instead, it sounds very much as though SCO hasn't actually done much of anything to identify what from IBM has been contributed to the kernel, or they would have made stronger claims than they have. Instead, it seems they're waiting for IBM to give them a list of everything that IBM has ever contributed, at which point they'll claim that everything on the list is improper. That's not the way that things are supposed to work; the plaintiff has to make specific claims of wrongdoing before the defendant must respond.
Slashdot Mirror
And equally importantly this shows the impossibility of eliminating government. The East India companies became de facto governments, even if they were not de jure governments. But rather than being representative governments that had the interests of their citizens in mind, they were imperialistic institutions that exploited the people under their "care" ruthlessly. There's every reason to think that big businesses would do exactly the same thing today if we were to relax control over them.
As somebody else pointed out, SCO is USL's successor in interest to the SysV code, so they actually are an interested party in the BSD case. Beyond that, they wouldn't actually be attacking the settlement, per se. Instead they'd be attacking the behavior of the other settling party by claiming that they weren't living up to their side of the bargain.
This might actually be a vaguely clever move. The settlement in the BSD case is under seal, so an ordinary person won't actually know exactly what its terms are, and the other side isn't allowed to disclose the exact terms. That makes it a perfect FUD target. SCO is free to imply all sorts of things about the settlement so long as they don't say anthing specific. They can even say that they'd love to specify exactly what the BSDers are doing wrong but they're forbidden to by the secrecy around the settlement. Nobody can rebut them without actually spelling out the terms of the settlement, which would be illegal. It might even have been effective if SCO hadn't shown itself to be all bark and no bite by now.
It seems to me that this is just more evidence that computer systems will wind up looking like biology. First we had viruses and similar infectious things. Now people are trying to create the machine equivalent of an immune system. The problem is that in the process they're likely to rediscover all of the problems that our immune system causes as well as the benefits.
This particular case is quite similar to allergies in the natural immune system. It's an overly aggressive response to an essentially harmless signal. The big problem is that virus and worm scanners are going to be succeptible to the computer equivalent of autoimmune disease; they'll start thinking that essential system files have been corrupted and try to wipe out something really important. I just hope they never develop the computerized equivalent of leukemia.
Given that this is /., I'd think that the answer would be obvious:
That's not what RMS would say, though. RMS has always tried very hard to separate the issue of software freedom from the issue of price, and I can't imagine that he'd pass up the opportunity to dismantle the whole zero price argument. He'd say something like:
There are much better reasons for talking to Linus and RMS than there are for talking to ESR or Bruce Perens. Linus obviously knows more than anyone else in the world about what code is in the Linux kernel and how it got there. Since that's the single most important factual issue in the case, it shouldn't be at all surprising that SCO wants to ask him questions. Similarly, RMS is an expert on the GPL, which is clearly relevant to IBM's claims of GPL infringement by SCO. Both ESR and Bruce Perens are vocal advocates of Open Source Software, but they lack the depth of case specific knowledge that makes Linus and RMS such good targets for the subpoena.
I strongly suspect that part of the reason that Stallman has such strong views about precision of terminology is precisely because he's been working on the legal side of things for so long. The GPL is as much about hacking the legal system as Linux (err, GNU/Linux) is about hacking computer systems. To create a hack as elegant as the GPL, it's necessary to be pretty well versed in the medium you're hacking, and it seems that a fair bit has worn off on RMS.
Nope. If you want to complain that it's no longer a "justice system" you'd have to prove that it used to be one. Most of the complaints about the legal system- that it gives too much advantage to people with deep pockets, that it lets people sue with little or no cause, etc.- have always been true. At the same time, the current system is clearly better in some respects; a good example is the ability of racial minorities to get a fair trial.
SCO gets no help here. The side letter was given by SCO as an Exhibit C to their original complaint against IBM. Ammendment X was given as Exhibit D. SCO clearly knew about the modifications to their licensing agreement with IBM since they launched the suit. They are, contrary to all reason, trying to repudiate those modifications to the agreement, but they certainly can't claim lack of knowledge or fabrication on the part of IBM.
It may be gratis (i.e. Free Beer), but it won't be libre (i.e. Free Speech) until they let users modify and share the source.
It looks at least a tiny bit more refined than that. When I run egrep -ril 'smp|jfs|rcu' * | wc -l on the kernel source in my Fedora Core 1 system, for instance, I get a total of 2437 files, while SCO only mentioned a few hundred. That also misses some of the files that SCO did mention, like arch/i386/kernel/mca.c- not surprising considering that particular file originated back in the mid 90s and deals mostly with some crufty old hardware. It mentions IBM, though, since MCA was originally an IBM design, and that's probably why it was chosen.
And the fact that it's your home server is exactly the difference between being a user and being a vigilante. It's fine to say "I trust SPEWS and Spamcop to blacklist only people I don't want to receive mail from" when you apply it to your own mail. If the blacklisters make a mistake (or decide to change their policies in unfortunate ways without telling you) you're only hurting yourself, and you've made your decision with knowledge of that risk. It's not OK to make that decision for somebody else, at least without telling them what you're doing and giving them a chance to opt out. If you do apply a blacklist to them without their consent, you risk blocking mail that they want without telling them.
Bayesian filtering is very good at removing the annoyance of having your mailbox fill up with spam. Unfortunately, those are just the symptom of a deeper problem, and one that Bayesian filtering does little to solve. Spam causes a number of infrastructure problems because it creates large volumes of traffic and takes up enormous amounts of space on ISPs' mail servers. Filtering means that you don't see the problem, but it doesn't help the ISPs one bit. And- the real problem- it doesn't do much to discourage spammers from sending out their messages, because the people who care enough about spam to set up filters are smart enough not to buy anything from a spammer in the first place.
If you really want to shut the spammers down, you need to kill their profit margin. You can do that two ways. One is to take away their customers. ISP-level filtering (which prevents spam from getting to gullible recipients) can help there, as can education programs. The other is to increase their costs. Forcing them to jump from service provider to service provider, implementing some kind of pay-per-mail system, or letting users sue spammers can all work on that front.
Yeah, because blacklisting has been so effective thus far, we just need to do more of it. Yeah, right. Blacklisting is basically playing a game of whack-a-mole; it makes things a bit less convenient for spammers, but doesn't seem to be doing them serious harm. OTOH, crippling the email of innocent bystanders who happen to share IP blocks with spammers seems a rather steep price to pay for something that does very little to stop spam.
Spam is a tough problem, and it's going to take more than just vigilante action to deal with it. What's needed is a two pronged approach. One prong is legal and is being followed fairly well; pass laws that make spamming illegal. The other prong, which is still under development, is to make technical changes to email so that spammers can't hide their addresses. Neither one will succeed alone- laws can't help as long as spammers can hide, and making spammers stand still won't help if there's no legal recourse against them- but the combination of the two should help a lot.
Don't you know anything? An apple a day keeps the doctor away.
Bleah. Straight to the wax, man, straight to the wax! If there's no need for mixing- like a symphony performance- you can record directly from the microphones to the Vinyl Master. If it's done right it can sound really fabulous, though I suspect that a lot of that is that the sound never gets mangled in the mixing process, rather than that the vinyl is particularly great. A straight to high-quality digital recording would probably sound just as good.
Unfortunately it's only "within minutes" if the program you're interested in is comparatively small. Just try to emerge the latest version of Mozilla, or KDE, or GNOME, or something substantial like that and see if it takes minutes. More likely it will take hours. If you're talking about installing a whole system from scratch, I can easily download and install all 3 Fedora disks faster than I could get Gentoo to the point of having X compiled.
I'm quite happy with the new kernel version, since it supports DMA mode on my new SATA drive, while the kernel in RH9 would occasionally hang if you tried to turn DMA on. Not a huge thing, unless you happen to have an affected drive, in which case it makes a world of difference. There are some other kernel changes that should make life nicer for laptop users.
Some other random changes that I notice in their release notes:
They've also removed a number of packages that are either no longer necessary (e.g. QT2, LPRng) or have licensing issues (e.g. pine, some aspell packages).
One other big change is that they're moving from using ASCII to UTF8 wherever they can. That's one of those things that's kind of annoying while the switch is taking place (they mention that this could cause problems with telnet and ssh, since they don't explicitly negotiate the encoding) but will presumably make everything easier in the long run.
I can think of two possible reasons for this. One is that part of the reason that people moved to Linux in the first place is because they were so tired of the "wait for the vendor to put out his next version" process. Experiences like Windows ME are bound to do that to people. One of the cool things about Free Software is that you can upgrade when enough new features have come out that you think its worthwhile, not when your vendor decides for you. A system like Debian, where it can be a long time before the cool new stuff percolates down to stable, can be frustrating for people who have moved specifically to get faster access to that cool new stuff.
Another important aspect is the speed with which that cool new stuff actually becomes available. When you look at it, the rate of development in the Windows world is a lot slower than in Linux. Admittedly, some of that is because the Windows software is more mature and there's less room for improvement, but no matter the reason for it, the pace of change is clearly much faster. That means that there's more motivation to update. Two years ago in Windows, I was running the same operating system (Win2000) and almost the same applications software (Office 97->2000) that I am today. In Linux, though, I've moved from Gnome 1.4 to Gnome 2.4, which is a huge upgrade both in the desktop environment and the applications that go with it. That rapid pace of development makes it much more reasonable to care about upgrading frequently.
That's interesting, but Idealab is not a Silicon Valley company. It's based in Pasadena, which means that it missed Silicon Valley by about 350 miles.
I think that it would be closer to:
Clippy: It looks like you have been wounded. Would you like me to help you:
It shouldn't be a bandwidth problem. Groklaw is now hosted by Ibiblio (AKA Metalab AKA Sunsite), which has more bandwidth than god. If they're having problems it's probably their servers lacking sufficient horsepower to run a dynamic site when hit this hard, rather than their pipe being too thin.
Fedora Core is rather poorly named. It covers roughly the same territory as the standard Redhat did, so if that was adequate for you then Fedora Core should be, too. It's also designed to be more friendly to external repositories (which makes sense for a community project) so there should be sources for most things that aren't on the 3 CDs of the standard install.
And, as people have repeatedly pointed out, there's actually more protection in open source than there would be in typical proprietary software. After all, with open source (meaning only that the code is inspectable, not necessarily that it's hackable) anyone who wants to check it for violations may do so. If the code isn't available, finding violators is that much more difficult. You really have to wonder if the ability to hide how things are done isn't just fooling people into thinking that closed source is better for protecting IP.
But that isn't what they're asking for. If you look at their question, they're asking for everything that anyone at IBM ever released without confidentiality agreements. They don't even restrict the request to things that happened since IBM licensed SysV, or to people associated with Linux, or anything. That's a grossly overbroad request, and there's no reason that IBM should have to answer it. The process really should go something like:
That's the reasonable way for things to proceed. Instead, it sounds very much as though SCO hasn't actually done much of anything to identify what from IBM has been contributed to the kernel, or they would have made stronger claims than they have. Instead, it seems they're waiting for IBM to give them a list of everything that IBM has ever contributed, at which point they'll claim that everything on the list is improper. That's not the way that things are supposed to work; the plaintiff has to make specific claims of wrongdoing before the defendant must respond.