Stallman has chosen to push the GPL as "as copyright + extra freedoms" rather than "public domain - extra restrictions". When you think of a program with no real owner, public domain comes to mind, but Stallman used copyright so he could add his extra restrictions. The perspective is flexible if you ask me. He's like a lawyer or a politician who can give you one things and sell it as another.
The license can be considered (and used) apart from Stallman. At this point, it is even used by people wh o significantly disagree with Stallman on many points and even use it to serve goals he doesn't agree with. Ask the Ghostscript people. Just because a work is GPLed doesn't mean Stallman or the FSF owns it. Secondly, GPL programs most certainly have "r eal owners". Since this is all about Stallman how do you explain Ghostscript, QT, or reiserfs? All of those are licensed in context dependant ways because they have "real owners". Most people who use the GPL actually do intend "copyright + extra freedo ms". GPL projects that have their code abused by corporations do indeed seem to feel they have ownership.
If you don't care for the GPLed stuff then that's fine. Theres plenty of BSD stuff thats just as good or maybe even better. But you know what, ev en that stuff isn't public domain. BSD coders often get hacked off if their stuff is used without acknowledgement. Yes, Virginia the BSD license has a significant restriction on "total freedom": Don't plagiarize our work. You can use it but have to acknowledge you didn't write it. It has another significant restriction: "You don't get a warranty so if it breaks in half, you get to keep both halves." Of course if that isn't "Free" enough then you can let MS exercise it's freedom to license as it sees fit all over you.
This "public domain - extra restrictions" is a red herring. If you expect to find a full OS and associated utilities in the PD then good luck to you. Even those "totally free" BSD guys disclaim liability and insist on acknowlegement. The moral of story is that there is no such thing as "total freedom". The GPL and even BSD, X, MIT licenses are ways of saying "You're right to throw a punch ends at my nose."Ë
If you don't agree to it then it reverts back to copyrights of the coders. If you don't agree to the GPL then you have to get all the authors persmissions.
Let's be clear about who has to do what. The GPL does not regulate usage; it regulates re-distribution. In regards to the Linux copyright holders, OpenLinux users are in the clear. Since SCO is trying to cloud the waters those users probably shouldn't try to re-distribute SCO's Linux but they are just fine on usage. There's no need for any SCO Linux users to rip out their mail-server say.
SCO on the other hand does need the authors' permission to distribute the code on the new basis they are attempting to create. They have a snowball's chance in hell of getting it.
Stallman claims that the GPL gives you rights and does not take them away. Well, the fact is that the GPL permits you do do certain things and it prohibits you from doing other things.
Let's leave morality and Stallman out of this for a minute and just look at the license. Standard copyright law says you can't do anything without a license or sale of rights from the copyright holder. Well, not quite nothing. Excerpts can be quoted critically and the like but standard copyright law gives no usage or distribution privileges by default.
So far, the GPL hasn't come into this at all. First off, the GPL grants unlimited usage rights. Well they specifically say any usage of the Program is permitted. You can even use Emacs to write "Stallman is a dirty hippy." and post it to Slashdot. It then goes on to grant unlimited distribution provided specific rules are adhered to. At no point has anything been taken from you that you didn't already have. Now you may detest the "specific rules" but that is another argument. The fact remains that the GPL permits far more than no license at all.
And yes, the GPL can be construed as granting rights. I can do any number of things that a software author detests with his work as long as I follow the rules. I can fork the code, publically question its design, benchmark it and so forth. The author can't say crap as long as I follow the rules. The real world works that way. Go yell "Fire!" in a movie theatre if you would a real world demonstration of what rights are and are not. Rights, rules, and prohibitions aren't mutually exclusive except in the minds of certain obtuse philosphers who obviously didn't get out much.
SCO can go ahead and distribute the kernel that way. They can even weasel a judge into blessing it for them. There's only two problems: the 3 million or so lines of code they didn't write and the 300 or so hacked off developers who hold the copyrights on that.
They already had a target pasted to their balls. Now they added flashing neon to it.
What? You mean Gates and Ballmer racked up all those frequent flyer miles for nothing? Oh wait a minute.....I guess they could turn some of them in and get this really neat dual-proc 64 bit Apple machine.......
Apart from well known missing functionality (color pre-press and the like) the only real problem GIMP has is that it isn't Photoshop. Where functionality bullet points do match up (and many of them do) GIMP can get the job done in a reasonable amount user time.
I run into this with stickynote computer users all the time. The application learned in a given category becomes the "One True Way" to do whatever. I have AppleWorks users who hate Office and vice versa. I've seen Gimp processed work that looks just as "high end" and "ultraprofessional" and "magazine quality" as Photoshop. I've messed with both and I hate to break this to you but Photoshop is not simple. A user of either has to be decent artist and at least a competent computer user. Geek friendly? Put my Mom (who can paint) in front of Photoshop and her head would explode.
I like open source, but sometimes I secretly hope for it to fail. Otherwise, I fear, I will be working at MacDonalds, coming home to do my real work for free.
You may work at McDonald's anyway. This economy has been rough on everybody but I'll point something else out. As a user, I've come to heavily distrust vendors who can send BSA goon squads after me and use EULAs and no end of weaseality to pull software from under my feet if I piss them off. If I don't piss them off in particular, they are always looking for ways to raise the rent. I avoid proprietary software vendors because I don't trust them anymore. The parameters of trust are such that to regain mine they might as well go Open Source.
Proprietary software may be a moneymaker but not from me. Read any good EULAs lately? I feel like I'd have to be off my rocker to buy into a deal like that. Its like a record company contract for the masses.
If the RedHat patents are implemented in the kernel then we most certainly can use them in perpetuity as long as they are implemented in GPLed code. The catch is that unlike regular GPLed code, you can't deduce the algorithms and recode it yourself under a different license. It's a two-edged sword. Since RedHat distributed their IP in the form of the GPLed kernel then they are bound by the license which states that any patents must be licensed to anyone who recieves the code. However, the GPL is only thing protecting you from their wrath if you re-implement. RedHat could be bought by MS for all the difference it would make.
BTW, license trolls are apt to jump on this like stink on shit. Most licenses don't have any indemnification against patents whatsoever. The GPL isn't forcing a patent on you; it is giving you protection you otherwise wouldn't have.
Patented IP implemented in other FOSS licenses is more ambiguous. The code itself may be free but since patents which are a different form of IP aren't mentioned in them then the code is legally ambiguous. A good lawyer may be able to argue that if patented IP were distributed (by the patent holder) under the BSD license then permission to use the patent is implied but it won't be cut and dried. Of course, a few good precedents involving other licensed patented IP would help with that a lot.
IF Metallica first apologizes for all the crap they pulled over Napster a few years ago.
They should apologize to the PAN newsreader guys while they're at it. I think PAN was attacked out of spite when Napster wasn't dying fast enough for them.
I wager that it would hurt if all of the sudden Unixware couldn't run Apache, Samba, Netatalk, PHP, Perl,.... etc. Even on proprietary Unices, opensource adds a lot of value. If a large swath of FOSS software suddenly won't run on SCO without a lot of unsupported rehabilitation then that will go a long way towards monkeywrenching their plans. I wouldn't be terribly opposed to doing Solaris the same way if SUN gets any louder trying to take advantage of this situation. A turning point has been reached. Commercial Unix needs FOSS more than FOSS needs it. I don't think it is in the least immature or inappropriate to give them a little reminder.
These guys attacked an open source newsreader project. A couple of years ago, the PAN developers incorporated an mp3 decoder into their NNTP client. For some reason, Metallica felt compelled to sue them over this. Clueless f****. The most priceless courtroom moment came when the plaintiff attorney asked "I wonder how the defendants would like it if someone was giving away their work?" The court seemed a little baffled at the laughter that ensued.
I could almost understand their attacking Napster. But a newsreader? Sorry but:
Of course, the fact that this band was already giving away 99.9% of the music they created doesn't matter to the/. crowd, who wants to yell and scream about how those bastards tried to protect that last little fraction of their own creation. ("How unfair! Only giving away most of their music! We'd better boycott 'em!")
does not suffice to rehabilitate these bungholes. A newsreader? An f****n' newsreader? BTW, the judge was clueless too. The PAN guys had to remove the inline mp3 feature. It still downloads mp3s just fine. Idiots. Lars still needs to be soundly beaten with a clueclub.
Drive letters are another way of saying "multiple root virtual filesystem". The way Windows does it makes it very hard to restructure storage. For instance, try moving Program Files from C: to D:. Windows didn't like it very much did it? The equivalent on a Linux system is moving/usr to another partition or physical disk. All you have to do is move/copy the files and lightly edit a file called fstab. The system won't notice that most of it's software is now on another disk or even another system.
Most Linux users accomplish what drive letters do with a directory structure like this:
/mnt ---->
floppy/ cdrom/ mp3/ junky_hard_drive/
If you like it pointy-clicky both GNOME and KDE will let you assign icons to those directories under/mnt.
Basically since we use a single rooted VFS rather than drive letters, it is much easier to structure or restructure storage. We have a much easier time creating things like thin clients and Live-CDs because the filesystem layer doesn't impose hard coded data locations on us.
I noticed right clicking in Win Explorer in Win2K allowed you to apply attributes to files. Not sure if they were limited to MS Office files.(Sorry, I can't check this anymore;-)
I believe that NTFS supports multiple streams of data per filename much like HFS. However, Windows doesn't make the pervasive use of the capability that Apple did in OS 9. I think one reason is that pretty much anything can go in these multiple streams meaning that particular files are tied even more closely to particular applications. Another reason may be that multiple filestreams complicate transfering files from machine to machine.
You're correct about RPM. It is using a more conventional approach by chucking it's data in a database file. Read about Netatalk's.AppleDouble directories for an approach to foreign datastreams...they can get ugly pretty quick.
Some of this is already present in a more limited form in Unix. For example, I still can't figure out why Windows doesn't have something like the LOCATE database that is set up by default on my FreeBSD box. But WinFS will blow LOCATE away (update is real time, not daily, and it has much more than just the file name).
Win9x did but it was implemented badly. It was continually rebuilding the index database. A common performance tune-up for 9x is to disable Find Fast.
The Linux version can hit speed a little bit during indexing but Linux doesn't continually rebuild the indexes. Most people just have a cron job that rebuilds the index late at night or early in the morning.
If people really wanted it, I suppose modules could be written for find to cache things like id3 tags and graphics metadata. Cache data from stat and file as well and make a nice way to use from KDE and Gnome and a lot of this could be accomplished with major replumbing. It won't be up to the minute but it would be Good Enough for most people....say how about a daemon that can be set to watch particular directories? Find Fast wasn't necessarily a bad idea; it was just badly implemented.
It looks a lot like Reiser4 will be out pretty soon so we'll learn first hand the pros and cons of MS' realtime approach to the problem.
Why would you bother? Macs are such nice computers, why would you want to attempt a witebox version that was pretty much made up of the same parts, but uglier?
Granted, but these machines aren't cheezy "witeboxes". They were reworking the Apple parts into some really snazzy looking low-profile brushed metal cases. These machines would have filled a similar niche to the failed Cube...probably had better QC on them too.
If I write 10,000 lines of proprietary code, and then take a 500 line chunk of GPL'd code, say, and insert it, and then distribute the whole 10,500 as a complete work, all 10,500 lines is now released to the world under the GPL. Anyone who gets a copy has all the rights the GPL grants them; rights I was forced, by the GPL, to grant them, by virtue of using GPL code.
Bzzzzt! Wrong on at least two counts. First off, inserting 500 lines of incompatibly licensed code into your work is pretty damn willful. The GPL didn't force you to do anything. Incidentally the licences on proprietary SDKs can be even more "viral" and "onerous" than the GPL. How come nobody ever bitches about them?
Secondly, the combined work is not GPLed unless you have included a copy of the GPL with the source and you have indicated (preferabably in the files themselves) that the work is GPLed.
The 10,500 line combined work that hasn't been explicitly licensed GPL is illegal to distribute. If you distribute it anyway without complying with the GPL then you are committing copyright violation. Making the whole thing GPL is only one remedy and is probably the result of a settlement. The judge in any litigation involving the code could have any number of legal options. A likely one is an injunction against any further distribution whatsoever and statutory damages. Anyone else in possession of the combined work is no more allowed to distribute it than you were.
I can, in fact, release the work under another more or restrictive license. But I can never retrieve the code I have released under GPL......Otherwise you have no rights to the GPL code under the GPL license.
Correctemundo! He who writes the code chooses the license. If you don't like the license then don't use the code. Not whinging about it would be a nice bonus. Incidentally, that not being able to recall the code is a feature not a bug. One of the reasons Linus gave for putting the GPL on the kernel is that he wanted developers to trust him. That trust is a major reason why I avoid most commercial software. I pay money and they can trump up reasons to revoke my license if I piss them off. That non-recallability is also needed if someone like say...SCO perhaps...wants to renege on the license. Any beatings SCO takes on that account are richly deserved.
Now, taking scenario two further to line up with SCO: let's say I am notified that the code I added was actually not legitimately BSD license, because an employee without the rights to do so 'released' it without permission of the owner (his employer). My code is 'tainted' by virtue of that persons copyright. I cannot release it in any form because it would constitute a derivative work. If I released it already, I have to desist. The people I released it to gained no rights in the code I released because the rights weren't actually mine to give. Now, I can go back and REMOVE that code, however, because I still own the copyright on my own code, and I can then distribute my work again, without the tainted code. In other words, the taint does not irrevocably stain MY code.
The problem here is that SCO needs to rewrite history to get what they want. They distributed Linux for years. The corporate entity (SCO, Caldera, scumbags...whatever) even donated an SMP machine to Alan Cox for SMP kernel work. You can even find technical posts from Caldera/SCO employees on the kernel mailing regarding SMP development. SCO can not credibly pull off the "Wild Cowboy We Didn't Authorize" defense. They participated in Linux development and distribution for years. The scenario you outline is workable apart from that.
That said, what a GPL does do to you, as a copyright holder, *IF* you create a derivative work, is taint that derivative work. Once you incorporate any GPL source (licensed to YOU under the GPL; not including code you create yourself that you license under the GPL), your work can never go back to
Like proprietary vendors also don't "rape and pillage". MS is infamous for it; RAV antivirus is just the latest example. Why don't you just say its a problem with software development in general?
There are a few large-ish companies (and one very large in particular - guess who!) that claim that Open Source in general and Linux in particular aids terrorists by providing them with a reliable and secure tool without intentionally placed backdoors (for law enforcement or otherwise).
Any vendor created backdoor that can be used to catch terrorists can also be used for general espionage. No government in it's right mind would buy an OS from that vendor. That has been a concern with some of these governmental OSS rollouts we've been hearing about. If a "certain large vendor" is playing the terrorism backdoor card then it is cutting it's own throat outside the US.
one has to think their tactic is to get this resolved in a back room, rather than in a court.
The "resolution" starts with SCO dropping it's pants and bending over. Once the dispute has been "resolved" SCO will look something like this. Maybe SCO should just take it's chances with a judge.
SCO has been sent a Cease and Desist by one of the kernel hackers. The hacker is demanding FTP logs and other records so he can determine the degree of infringement against his copyright. He is promising that he will sue unless SCO ceases their claims against other parts of the kernel.
Every contributor to the kernel can do this. I hope those scumbags get hammered for every penny they have.
This has probably already occurred to IBM. They'll get back at them indirectly by calling SCO's bluffing and battering them in court. The first drubbing will occur when IBM mounts their "strong legal defence". The second will come when IBM reaches into their patent portfolio and drops anvils on McBride's head. The third may come when Linux copyright holders file suits of their own. The fourth may come when more Linux business file unfair practices suits. Germany and Poland are only the tips of that iceberg.
I think McBride and friends mean to screw them all by suddenly dumping the stock and running for the hills. I can't see them hanging around for the monster shitstorm that is coming their way.
Slashdot Mirror
Stallman has chosen to push the GPL as "as copyright + extra freedoms" rather than "public domain - extra restrictions". When you think of a program with no real owner, public domain comes to mind, but Stallman used copyright so he could add his extra restrictions. The perspective is flexible if you ask me. He's like a lawyer or a politician who can give you one things and sell it as another.
The license can be considered (and used) apart from Stallman. At this point, it is even used by people wh o significantly disagree with Stallman on many points and even use it to serve goals he doesn't agree with. Ask the Ghostscript people. Just because a work is GPLed doesn't mean Stallman or the FSF owns it. Secondly, GPL programs most certainly have "r eal owners". Since this is all about Stallman how do you explain Ghostscript, QT, or reiserfs? All of those are licensed in context dependant ways because they have "real owners". Most people who use the GPL actually do intend "copyright + extra freedo ms". GPL projects that have their code abused by corporations do indeed seem to feel they have ownership.
If you don't care for the GPLed stuff then that's fine. Theres plenty of BSD stuff thats just as good or maybe even better. But you know what, ev en that stuff isn't public domain. BSD coders often get hacked off if their stuff is used without acknowledgement. Yes, Virginia the BSD license has a significant restriction on "total freedom": Don't plagiarize our work. You can use it but have to acknowledge you didn't write it. It has another significant restriction: "You don't get a warranty so if it breaks in half, you get to keep both halves." Of course if that isn't "Free" enough then you can let MS exercise it's freedom to license as it sees fit all over you.
This "public domain - extra restrictions" is a red herring. If you expect to find a full OS and associated utilities in the PD then good luck to you. Even those "totally free" BSD guys disclaim liability and insist on acknowlegement. The moral of story is that there is no such thing as "total freedom". The GPL and even BSD, X, MIT licenses are ways of saying "You're right to throw a punch ends at my nose."Ë
If you don't agree to it then it reverts back to copyrights of the coders. If you don't agree to the GPL then you have to get all the authors persmissions.
Let's be clear about who has to do what. The GPL does not regulate usage; it regulates re-distribution. In regards to the Linux copyright holders, OpenLinux users are in the clear. Since SCO is trying to cloud the waters those users probably shouldn't try to re-distribute SCO's Linux but they are just fine on usage. There's no need for any SCO Linux users to rip out their mail-server say.
SCO on the other hand does need the authors' permission to distribute the code on the new basis they are attempting to create. They have a snowball's chance in hell of getting it.
Stallman claims that the GPL gives you rights and does not take them away. Well, the fact is that the GPL permits you do do certain things and it prohibits you from doing other things.
Let's leave morality and Stallman out of this for a minute and just look at the license. Standard copyright law says you can't do anything without a license or sale of rights from the copyright holder. Well, not quite nothing. Excerpts can be quoted critically and the like but standard copyright law gives no usage or distribution privileges by default.
So far, the GPL hasn't come into this at all. First off, the GPL grants unlimited usage rights. Well they specifically say any usage of the Program is permitted. You can even use Emacs to write "Stallman is a dirty hippy." and post it to Slashdot. It then goes on to grant unlimited distribution provided specific rules are adhered to. At no point has anything been taken from you that you didn't already have. Now you may detest the "specific rules" but that is another argument. The fact remains that the GPL permits far more than no license at all.
And yes, the GPL can be construed as granting rights. I can do any number of things that a software author detests with his work as long as I follow the rules. I can fork the code, publically question its design, benchmark it and so forth. The author can't say crap as long as I follow the rules. The real world works that way. Go yell "Fire!" in a movie theatre if you would a real world demonstration of what rights are and are not. Rights, rules, and prohibitions aren't mutually exclusive except in the minds of certain obtuse philosphers who obviously didn't get out much.
SCO can go ahead and distribute the kernel that way. They can even weasel a judge into blessing it for them. There's only two problems: the 3 million or so lines of code they didn't write and the 300 or so hacked off developers who hold the copyrights on that.
They already had a target pasted to their balls. Now they added flashing neon to it.
What? You mean Gates and Ballmer racked up all those frequent flyer miles for nothing? Oh wait a minute.....I guess they could turn some of them in and get this really neat dual-proc 64 bit Apple machine.......
Apart from well known missing functionality (color pre-press and the like) the only real problem GIMP has is that it isn't Photoshop. Where functionality bullet points do match up (and many of them do) GIMP can get the job done in a reasonable amount user time.
I run into this with stickynote computer users all the time. The application learned in a given category becomes the "One True Way" to do whatever. I have AppleWorks users who hate Office and vice versa. I've seen Gimp processed work that looks just as "high end" and "ultraprofessional" and "magazine quality" as Photoshop. I've messed with both and I hate to break this to you but Photoshop is not simple. A user of either has to be decent artist and at least a competent computer user. Geek friendly? Put my Mom (who can paint) in front of Photoshop and her head would explode.
Photoshop is Photoshop-Geek friendly.
Or when Linus gets the commercial and media attention Steve Jovs gets. Or when Linus developes a reality distortion field of his own.
. as p
I don't see how that can happen. Linus doesn't possess a reality distortion field; he projects a reality assertion field. Check him out here on SCO:
(fucking lameness filter bitching about this as a link....)
http://www.eweek.com/article2/0,3959,1134098,00
He's understated as ever.
I like open source, but sometimes I secretly hope for it to fail. Otherwise, I fear, I will be working at MacDonalds, coming home to do my real work for free.
You may work at McDonald's anyway. This economy has been rough on everybody but I'll point something else out. As a user, I've come to heavily distrust vendors who can send BSA goon squads after me and use EULAs and no end of weaseality to pull software from under my feet if I piss them off. If I don't piss them off in particular, they are always looking for ways to raise the rent. I avoid proprietary software vendors because I don't trust them anymore. The parameters of trust are such that to regain mine they might as well go Open Source.
Proprietary software may be a moneymaker but not from me. Read any good EULAs lately? I feel like I'd have to be off my rocker to buy into a deal like that. Its like a record company contract for the masses.
To relicense Linux as an SCO-0wned product, you'd have to get the agreement of all the contributors. I doubt that that would happen.
I think they better be more worried about many of the contributers suing them.
Nuff said. Couldn't have said it better myself.
If the RedHat patents are implemented in the kernel then we most certainly can use them in perpetuity as long as they are implemented in GPLed code. The catch is that unlike regular GPLed code, you can't deduce the algorithms and recode it yourself under a different license. It's a two-edged sword. Since RedHat distributed their IP in the form of the GPLed kernel then they are bound by the license which states that any patents must be licensed to anyone who recieves the code. However, the GPL is only thing protecting you from their wrath if you re-implement. RedHat could be bought by MS for all the difference it would make.
BTW, license trolls are apt to jump on this like stink on shit. Most licenses don't have any indemnification against patents whatsoever. The GPL isn't forcing a patent on you; it is giving you protection you otherwise wouldn't have.
Patented IP implemented in other FOSS licenses is more ambiguous. The code itself may be free but since patents which are a different form of IP aren't mentioned in them then the code is legally ambiguous. A good lawyer may be able to argue that if patented IP were distributed (by the patent holder) under the BSD license then permission to use the patent is implied but it won't be cut and dried. Of course, a few good precedents involving other licensed patented IP would help with that a lot.
IF Metallica first apologizes for all the crap they pulled over Napster a few years ago.
They should apologize to the PAN newsreader guys while they're at it. I think PAN was attacked out of spite when Napster wasn't dying fast enough for them.
I wager that it would hurt if all of the sudden Unixware couldn't run Apache, Samba, Netatalk, PHP, Perl, .... etc. Even on proprietary Unices, opensource adds a lot of value. If a large swath of FOSS software suddenly won't run on SCO without a lot of unsupported rehabilitation then that will go a long way towards monkeywrenching their plans. I wouldn't be terribly opposed to doing Solaris the same way if SUN gets any louder trying to take advantage of this situation. A turning point has been reached. Commercial Unix needs FOSS more than FOSS needs it. I don't think it is in the least immature or inappropriate to give them a little reminder.
These guys attacked an open source newsreader project. A couple of years ago, the PAN developers incorporated an mp3 decoder into their NNTP client. For some reason, Metallica felt compelled to sue them over this. Clueless f****. The most priceless courtroom moment came when the plaintiff attorney asked "I wonder how the defendants would like it if someone was giving away their work?" The court seemed a little baffled at the laughter that ensued.
/. crowd, who wants to yell and scream about how those bastards tried to protect that last little fraction of their own creation. ("How unfair! Only giving away most of their music! We'd better boycott 'em!")
I could almost understand their attacking Napster. But a newsreader? Sorry but:
Of course, the fact that this band was already giving away 99.9% of the music they created doesn't matter to the
does not suffice to rehabilitate these bungholes. A newsreader? An f****n' newsreader? BTW, the judge was clueless too. The PAN guys had to remove the inline mp3 feature. It still downloads mp3s just fine. Idiots. Lars still needs to be soundly beaten with a clueclub.
Most Linux users accomplish what drive letters do with a directory structure like this:
floppy/ cdrom/ mp3/ junky_hard_drive/
If you like it pointy-clicky both GNOME and KDE will let you assign icons to those directories under
Basically since we use a single rooted VFS rather than drive letters, it is much easier to structure or restructure storage. We have a much easier time creating things like thin clients and Live-CDs because the filesystem layer doesn't impose hard coded data locations on us.
I noticed right clicking in Win Explorer in Win2K allowed you to apply attributes to files. Not sure if they were limited to MS Office files.(Sorry, I can't check this anymore ;-)
.AppleDouble directories for an approach to foreign datastreams...they can get ugly pretty quick.
I believe that NTFS supports multiple streams of data per filename much like HFS. However, Windows doesn't make the pervasive use of the capability that Apple did in OS 9. I think one reason is that pretty much anything can go in these multiple streams meaning that particular files are tied even more closely to particular applications. Another reason may be that multiple filestreams complicate transfering files from machine to machine.
You're correct about RPM. It is using a more conventional approach by chucking it's data in a database file. Read about Netatalk's
Some of this is already present in a more limited form in Unix. For example, I still can't figure out why Windows doesn't have something like the LOCATE database that is set up by default on my FreeBSD box. But WinFS will blow LOCATE away (update is real time, not daily, and it has much more than just the file name).
Win9x did but it was implemented badly. It was continually rebuilding the index database. A common performance tune-up for 9x is to disable Find Fast.
The Linux version can hit speed a little bit during indexing but Linux doesn't continually rebuild the indexes. Most people just have a cron job that rebuilds the index late at night or early in the morning.
If people really wanted it, I suppose modules could be written for find to cache things like id3 tags and graphics metadata. Cache data from stat and file as well and make a nice way to use from KDE and Gnome and a lot of this could be accomplished with major replumbing. It won't be up to the minute but it would be Good Enough for most people....say how about a daemon that can be set to watch particular directories? Find Fast wasn't necessarily a bad idea; it was just badly implemented.
It looks a lot like Reiser4 will be out pretty soon so we'll learn first hand the pros and cons of MS' realtime approach to the problem.
Why would you bother? Macs are such nice computers, why would you want to attempt a witebox version that was pretty much made up of the same parts, but uglier?
Granted, but these machines aren't cheezy "witeboxes". They were reworking the Apple parts into some really snazzy looking low-profile brushed metal cases. These machines would have filled a similar niche to the failed Cube...probably had better QC on them too.
If I write 10,000 lines of proprietary code, and then take a 500 line chunk of GPL'd code, say, and insert it, and then distribute the whole 10,500 as a complete work, all 10,500 lines is now released to the world under the GPL. Anyone who gets a copy has all the rights the GPL grants them; rights I was forced, by the GPL, to grant them, by virtue of using GPL code.
...Otherwise you have no rights to the GPL code under the GPL license.
Bzzzzt! Wrong on at least two counts. First off, inserting 500 lines of incompatibly licensed code into your work is pretty damn willful. The GPL didn't force you to do anything. Incidentally the licences on proprietary SDKs can be even more "viral" and "onerous" than the GPL. How come nobody ever bitches about them?
Secondly, the combined work is not GPLed unless you have included a copy of the GPL with the source and you have indicated (preferabably in the files themselves) that the work is GPLed.
The 10,500 line combined work that hasn't been explicitly licensed GPL is illegal to distribute. If you distribute it anyway without complying with the GPL then you are committing copyright violation. Making the whole thing GPL is only one remedy and is probably the result of a settlement. The judge in any litigation involving the code could have any number of legal options. A likely one is an injunction against any further distribution whatsoever and statutory damages. Anyone else in possession of the combined work is no more allowed to distribute it than you were.
I can, in fact, release the work under another more or restrictive license. But I can never retrieve the code I have released under GPL...
Correctemundo! He who writes the code chooses the license. If you don't like the license then don't use the code. Not whinging about it would be a nice bonus. Incidentally, that not being able to recall the code is a feature not a bug. One of the reasons Linus gave for putting the GPL on the kernel is that he wanted developers to trust him. That trust is a major reason why I avoid most commercial software. I pay money and they can trump up reasons to revoke my license if I piss them off. That non-recallability is also needed if someone like say...SCO perhaps...wants to renege on the license. Any beatings SCO takes on that account are richly deserved.
Now, taking scenario two further to line up with SCO: let's say I am notified that the code I added was actually not legitimately BSD license, because an employee without the rights to do so 'released' it without permission of the owner (his employer). My code is 'tainted' by virtue of that persons copyright. I cannot release it in any form because it would constitute a derivative work. If I released it already, I have to desist. The people I released it to gained no rights in the code I released because the rights weren't actually mine to give. Now, I can go back and REMOVE that code, however, because I still own the copyright on my own code, and I can then distribute my work again, without the tainted code. In other words, the taint does not irrevocably stain MY code.
The problem here is that SCO needs to rewrite history to get what they want. They distributed Linux for years. The corporate entity (SCO, Caldera, scumbags...whatever) even donated an SMP machine to Alan Cox for SMP kernel work. You can even find technical posts from Caldera/SCO employees on the kernel mailing regarding SMP development. SCO can not credibly pull off the "Wild Cowboy We Didn't Authorize" defense. They participated in Linux development and distribution for years. The scenario you outline is workable apart from that.
That said, what a GPL does do to you, as a copyright holder, *IF* you create a derivative work, is taint that derivative work. Once you incorporate any GPL source (licensed to YOU under the GPL; not including code you create yourself that you license under the GPL), your work can never go back to
Like proprietary vendors also don't "rape and pillage". MS is infamous for it; RAV antivirus is just the latest example. Why don't you just say its a problem with software development in general?
There are a few large-ish companies (and one very large in particular - guess who!) that claim that Open Source in general and Linux in particular aids terrorists by providing them with a reliable and secure tool without intentionally placed backdoors (for law enforcement or otherwise).
Any vendor created backdoor that can be used to catch terrorists can also be used for general espionage. No government in it's right mind would buy an OS from that vendor. That has been a concern with some of these governmental OSS rollouts we've been hearing about. If a "certain large vendor" is playing the terrorism backdoor card then it is cutting it's own throat outside the US.
...the pinky nail held to one side of the mouth and the maniacal laugh.
one has to think their tactic is to get this resolved in a back room, rather than in a court.
The "resolution" starts with SCO dropping it's pants and bending over. Once the dispute has been "resolved" SCO will look something like this. Maybe SCO should just take it's chances with a judge.
http://www.theinquirer.net/?article=10018
SCO has been sent a Cease and Desist by one of the kernel hackers. The hacker is demanding FTP logs and other records so he can determine the degree of infringement against his copyright. He is promising that he will sue unless SCO ceases their claims against other parts of the kernel.
Every contributor to the kernel can do this. I hope those scumbags get hammered for every penny they have.
This has probably already occurred to IBM. They'll get back at them indirectly by calling SCO's bluffing and battering them in court. The first drubbing will occur when IBM mounts their "strong legal defence". The second will come when IBM reaches into their patent portfolio and drops anvils on McBride's head. The third may come when Linux copyright holders file suits of their own. The fourth may come when more Linux business file unfair practices suits. Germany and Poland are only the tips of that iceberg.
I think McBride and friends mean to screw them all by suddenly dumping the stock and running for the hills. I can't see them hanging around for the monster shitstorm that is coming their way.