Seems to be a bit of selective clipping in the description, the key part that is missing is that the company was operating at a £28.5m loss and still ended up paying some amount of corporate tax. Not sure what the problem is? Is the company supposed to have done something wrong by paying out a large bonus that will be taxed individually anyways? While there are some interesting tidbits in the article about things like profit deferral, none of that seems relevant to the case at hand, so I'm somewhat at a loss as to what the news is supposed to be.
You can still exec()/execve() and wait() without issues.. its just the fork() implementation that has some problems. To get around this, vfork() is supported. Since there's no protection on anything, no real page tables, and there's no sane way to grab page faults (or even ELF relocation), vfork() mostly just blocks to make sure that the parents mm space isn't trampled by the child. I'm not entirely sure if COW works properly either.. I'd guess not, but haven't tried it.
Incase people haven't realized yet, Nintendo's version of the GameCube is utterly useless. Nintendo has proven once again their inability to support standards and instead run off and stick with their own proprietary media. The N64 had the same issue, but if you were lucky enough to pick up a development model, you could've grabbed a base that'll read cd's. Cartridges are best left in systems where low power consumption is a must.. and for things like the GBA, they work great. Anyways.. the point of this post.. Panasonic ended up licensing the GC from Nintendo awhile ago.. and they did their own version of it which supports not only the ultra-lame native media of the GC, but also DVD's. This version of the GC has been floating around japan for awhile, and is already available for import.. (see lik-sang). Anyone wanting a GC that doesn't totally suck should either wait for a US release of the panasonic gc, or just import one.. cost is about the same..
PDA's aren't just an execs toy.. theyre quite popular amongst FAE's (Sun even made their own PDA for their FAE's.. didn't use their own processor naturally, but that's another matter). There's also industrial application, there's a good market for that as well, though the traditional intdustrialized PDA is a lot more durable than what psion is peddling.
This is the same age old problem of incompetent people installing things blindly without first looking at what's actually being changed. If people are going to take that approach, then it's their own fault if there are hidden backdoors and other such things strewn about the code. For those of us that actually write code, or even those who don't, just reading the patch is enough indication to figure out what's going on and whether there's any foul play at work or not. Blaming someone for using backdoor tactics in a system where the source is fully available is no ones fault but your own if you get screwed by it.
Oh yeah, there ya go. Let's make a webserver farm with the idea of fail over and load balancing in mind, and then give it a single point of failure like an NFS mount. Why not just shoot yourself in the foot to begin with and save the cost of the DC's to acheive the same result.
smpeg/mplayer and friends should work just fine on it the way it is.. SDL has been working for awhile on it, but it's a wee bit on the sluggish side due to the slow nature of the framebuffer driver. Though a rewrite of the fb driver is planned soon, which'll utilize store queues and dma access for overall faster performance (possibly also tying into the drm subsystem for cleaner locking and dma handling). The SH4 itself is a fairly speedy processor, MHz is a pretty meaningless concept. It can do mpeg and mp3 decoding and such pretty much effortlessly, though lack of sound support for the AICA in its present incarnation certainly makes mp3s a somewhat meaingless venture.
That's developer ignorance. If you sign over copyright, you give up all your rights. Plain and simple. If the FSF wanted to relicense, they can do so. If Sistina wants to relicense, they can. ReiserFS is developed like this as well, and is constantly relicensed so it can be resold as a commercial product.
That's perfectly GPL compliant. Simply because you do changes to something under the GPL doesn't mean that your changes automatically get pushed into the tree. If you think this kind of requirement is out of line, perhaps you should look at the Free Software Foundation. Things like gcc, glibc, etc, etc. all require that copyrights be signed over _before_ anything will be accepted into the tree.
Since the argument here isn't in regards to free software, debian guidelines are useless for this discussion. As far as violating clause 6 of the OSS definition, that's really a matter of perception. You aren't restricted by using SPLed code in a commercial environment, the only limitation is that you have to pay Sistina royalties in the event that it's used in a place where it's directly involved with generating revenue (which there are hardly any instances of).
What debian decides to package is its own decision. I prefer functional software over politics.
Sistina has always had the policy that in order for your code to be part of the official GFS distribution, you had to sign the copyright over to them in its entirety. This isn't that big of a deal.. if you don't feel like signing over copyright, you aren't forced to, and your stuff can live on through contrib. Sistina isn't really taking anything from the community, as there haven't really been that many people outside of Sistina that contributed code. There was also an issue of people taking GFS, innovating it, and not giving their changes back to Sistina (Ahem Compaq), which is part of the reason that the SPL was started in the first place.
Please actually _look_ at GFS before pretending to have a clue about it. GFS is not distributed as a patch to the linux kernel, there are patches to the linux kernel that GFS requires, and those are GPLed. GFS itself is built outside of the kernel in its own source tree, and has been as far back as I can remember. The SPL is also not a sourced close license, please _READ_ it before thinking, and do us all a favor. Source can be distributed all day long without incident.
Wrong. The 25,000$ was for LVM, not GFS. LVM isn't being relicensed, GFS is. If you actually followed GFS development instead of trying to criticize Sistina without knowing anything about it, you would've realized that the new license was announced quite awhile ago, well before LWCE. There is still discussion going on on the lists regarding the SPL, and they're still open to changes.
For all the people sitting here complaining about how Sistina is violating the GPL and how they're evil and all that sort of nonsense, how many of you have actually used GFS in a way where it directly brings you profit? 2? 3? Get real people, the SPL doesn't prevent you from using it in a commercial environment, it doesn't take away any of your rights, it simply requires you to pay royalties in the event that you're using GFS as a native part of a solution that's being used _directly_ to generate revenue. Most people that are using it are either a) standard commercial environments where GFS is used in the company, but isn't being used directly to generate revenue, b) academic/research usage where a distributed fs becomes necessary to deal with load balancing and overall availability, or c) bored user with some scsi devices and a few machines that feels like playing with something new. None of these kinds of situations are those where Sistina requires royalties.
And for all you people complaining about how Sistina is "taking from the community", have any of you ever so much as looked at / used GFS or read the mailing lists? The majority of community involvement has been testing, which is completely uneffected by the SPL. As far as Sistina's requirement that all copyrights and such be signed over to them, this is nothing new. That requirement has been around just about since the dawn of time, and isn't anything the SPL added. Speaking as one who's written patches against GFS, and use it in a commercial environment, this requirement has been somewhat of a nuisance, especially since my employer doesn't want to give up copyright control. And thus, my code isn't part of the native distribution. This is nothing but a minor inconvenience, and certainly nothing life-threatening.
Sistina makes a great product, requiring royalties from people using GFS to make money is perfectly acceptable, and I'm surprised they didn't do it earlier. I for one welcome the SPL.
Linux is certainly better than wince for a lot of things.. in incredibly tiny spaces or dealing with systems w/o an MMU, it is not. Given the fact I can have a functional wince install in less space than I can get a useless 2.4 image to, is rather annoying. neutrino is kind of cute, but QNX has done a lot of stupid things with it as well *cough* qnx rtp *end cough*. I can't think of too many proprietary commercial RTOS'es that are even worth a damn.. VxWorks, pSOS, and LynxOS being examples of systems better left dead..
Where'd be the fun in that? I for one enjoy seeing lots of ignorant comments about using embedded systems for beowulf clusters, for mass storage, etc, etc. why should people have to be bothered with facts when ignorance is much more easily adopted? No doubt there will be lots of people who say "WinCE sucks! linux rules!", and those people are ignorant. those people have also likely never worked in an embedded environment before, and somehow feel that by owning a PDA they somehow have a clue regarding embedded applications. If slashdot were indeed unbiased, the majority of your ignorant slashdot users would start flaming slashdot for saying that there's something else out there that does a better job than linux in some areas. slashdot is only unbiased as long as it suits the purpose.
You seem to be missing the point of the whole thing. With no current standard in place, one is bound to happen sooner or later. Keep in mind that a good number of these companies that are part of the alliance are already part of the ELC, creating proprietary standards is not anything the alliance wants to do. Your concerns about software development being kept closed source are also unwarranted. The few software companies that are part of the alliance have a large portion of their business structure centered around open source software development and interaction with the open source community. And on top of all this.. what does it really matter? They're free to come up with what "standards" they want. It isn't going to make a difference if no one utilizes the standard.
You have got to be kidding. That's the kind of negligent moronic attitude that can be directly attributed to Columbine. Sheltering your child from things is ludicrous, they'll find out sooner or later, and it's your own fault for not discussing things with them first. On top of all this, you're also claiming that magic is bad. Perhaps you spent too much time criticizing instead of actually _reading_ the books. These books are not meant to be a factual reference, they are labelled as _fantasy_ books for a reason. If your child can't distinguish between fantasy and reality, you're again failing your job as a parent. If you had properly educated the child, these kind of things could be taken as what they were intended, entertainment value. A child is going to be exposed to violence regardless of your futile efforts, sheltering them from the reality of it isn't helping anything. Don't blame an excellent set of books for your inability to educated a child.
In reply to your Dreamcast, MIPS, and PowerPC comments, SH4, MIPS and PPC are all very heavily used in the embedded market, and likely will continue to be for years to come. While I doubt NewOS will ever get anywhere, it's always nice to see new open systems emerge that could be utilized in embedded space. Its ability to run on the Dreamcast (though while not really utilizing most of the hardware) is still fairly impressive considering what a nuisance the majority of the design of the Dreamcast is. *cough* maple bus *end cough*. So, while the Atari and Amiga are most definately dead, the rest of those are very much alive and kicking.
Not once did I mention anything regarding Intel's marketing tactics. If they want to claim that they have a superior product, they're free to do so. I am simply stating that x86 is an ancient, dead, and pathetic attempt at a decent architecture. Making it out to be the second coming is just pathetic. Taking a product that should've been killed off long ago and needlessly dragging it out in a desperate effort to look like innovation, isn't my idea of productivity. If you want to marvel in Intel's ability to beat a dead horse, no one is stopping you. Even AMD realized x86 was a dead end, hence the whole reasoning for licensing the ev6 bus from compaq. When you have to start licensing out features of other systems just to make your own look more worthwhile, you know it's time to start looking for a new product.
Intel didn't break the 1GHz 'barrier', this was already surpassed by countless other architectures before Intel finally got around to it. Sooner or later, Intel will realize they can only get so far by beating a dead horse. The only barriers that Intel has broken are the barriers of incompetence that would prohibit any competent corporation from dragging out a dead product needlessly and then trying to make it look revolutionary.
Slashdot Mirror
Seems to be a bit of selective clipping in the description, the key part that is missing is that the company was operating at a £28.5m loss and still ended up paying some amount of corporate tax. Not sure what the problem is? Is the company supposed to have done something wrong by paying out a large bonus that will be taxed individually anyways? While there are some interesting tidbits in the article about things like profit deferral, none of that seems relevant to the case at hand, so I'm somewhat at a loss as to what the news is supposed to be.
Stop slashdotting my mailserver / irc session box!
You can still exec()/execve() and wait() without issues .. its just the fork() implementation that has some problems. To get around this, vfork() is supported. Since there's no protection on anything, no real page tables, and there's no sane way to grab page faults (or even ELF relocation), vfork() mostly just blocks to make sure that the parents mm space isn't trampled by the child. I'm not entirely sure if COW works properly either .. I'd guess not, but haven't tried it.
Incase people haven't realized yet, Nintendo's version of the GameCube is utterly useless. Nintendo has proven once again their inability to support standards and instead run off and stick with their own proprietary media. The N64 had the same issue, but if you were lucky enough to pick up a development model, you could've grabbed a base that'll read cd's. Cartridges are best left in systems where low power consumption is a must.. and for things like the GBA, they work great. Anyways.. the point of this post.. Panasonic ended up licensing the GC from Nintendo awhile ago.. and they did their own version of it which supports not only the ultra-lame native media of the GC, but also DVD's. This version of the GC has been floating around japan for awhile, and is already available for import.. (see lik-sang). Anyone wanting a GC that doesn't totally suck should either wait for a US release of the panasonic gc, or just import one.. cost is about the same..
PDA's aren't just an execs toy.. theyre quite popular amongst FAE's (Sun even made their own PDA for their FAE's.. didn't use their own processor naturally, but that's another matter). There's also industrial application, there's a good market for that as well, though the traditional intdustrialized PDA is a lot more durable than what psion is peddling.
This is the same age old problem of incompetent people installing things blindly without first looking at what's actually being changed. If people are going to take that approach, then it's their own fault if there are hidden backdoors and other such things strewn about the code. For those of us that actually write code, or even those who don't, just reading the patch is enough indication to figure out what's going on and whether there's any foul play at work or not. Blaming someone for using backdoor tactics in a system where the source is fully available is no ones fault but your own if you get screwed by it.
Oh yeah, there ya go. Let's make a webserver farm with the idea of fail over and load balancing in mind, and then give it a single point of failure like an NFS mount. Why not just shoot yourself in the foot to begin with and save the cost of the DC's to acheive the same result.
The BBA is a hacked up rtl8139, and it's been supported for many many many months. Read first, think later.
smpeg/mplayer and friends should work just fine on it the way it is.. SDL has been working for awhile on it, but it's a wee bit on the sluggish side due to the slow nature of the framebuffer driver. Though a rewrite of the fb driver is planned soon, which'll utilize store queues and dma access for overall faster performance (possibly also tying into the drm subsystem for cleaner locking and dma handling). The SH4 itself is a fairly speedy processor, MHz is a pretty meaningless concept. It can do mpeg and mp3 decoding and such pretty much effortlessly, though lack of sound support for the AICA in its present incarnation certainly makes mp3s a somewhat meaingless venture.
That's developer ignorance. If you sign over copyright, you give up all your rights. Plain and simple. If the FSF wanted to relicense, they can do so. If Sistina wants to relicense, they can. ReiserFS is developed like this as well, and is constantly relicensed so it can be resold as a commercial product.
That's perfectly GPL compliant. Simply because you do changes to something under the GPL doesn't mean that your changes automatically get pushed into the tree. If you think this kind of requirement is out of line, perhaps you should look at the Free Software Foundation. Things like gcc, glibc, etc, etc. all require that copyrights be signed over _before_ anything will be accepted into the tree.
Since the argument here isn't in regards to free software, debian guidelines are useless for this discussion. As far as violating clause 6 of the OSS definition, that's really a matter of perception. You aren't restricted by using SPLed code in a commercial environment, the only limitation is that you have to pay Sistina royalties in the event that it's used in a place where it's directly involved with generating revenue (which there are hardly any instances of).
What debian decides to package is its own decision. I prefer functional software over politics.
Sistina has always had the policy that in order for your code to be part of the official GFS distribution, you had to sign the copyright over to them in its entirety. This isn't that big of a deal.. if you don't feel like signing over copyright, you aren't forced to, and your stuff can live on through contrib. Sistina isn't really taking anything from the community, as there haven't really been that many people outside of Sistina that contributed code. There was also an issue of people taking GFS, innovating it, and not giving their changes back to Sistina (Ahem Compaq), which is part of the reason that the SPL was started in the first place.
You should read it yourself, that's an award, not money. 25,000$ was for LVM, not GFS.
Please actually _look_ at GFS before pretending to have a clue about it. GFS is not distributed as a patch to the linux kernel, there are patches to the linux kernel that GFS requires, and those are GPLed. GFS itself is built outside of the kernel in its own source tree, and has been as far back as I can remember. The SPL is also not a sourced close license, please _READ_ it before thinking, and do us all a favor. Source can be distributed all day long without incident.
Wrong. The 25,000$ was for LVM, not GFS. LVM isn't being relicensed, GFS is. If you actually followed GFS development instead of trying to criticize Sistina without knowing anything about it, you would've realized that the new license was announced quite awhile ago, well before LWCE. There is still discussion going on on the lists regarding the SPL, and they're still open to changes.
For all the people sitting here complaining about how Sistina is violating the GPL and how they're evil and all that sort of nonsense, how many of you have actually used GFS in a way where it directly brings you profit? 2? 3? Get real people, the SPL doesn't prevent you from using it in a commercial environment, it doesn't take away any of your rights, it simply requires you to pay royalties in the event that you're using GFS as a native part of a solution that's being used _directly_ to generate revenue. Most people that are using it are either a) standard commercial environments where GFS is used in the company, but isn't being used directly to generate revenue, b) academic/research usage where a distributed fs becomes necessary to deal with load balancing and overall availability, or c) bored user with some scsi devices and a few machines that feels like playing with something new. None of these kinds of situations are those where Sistina requires royalties.
And for all you people complaining about how Sistina is "taking from the community", have any of you ever so much as looked at / used GFS or read the mailing lists? The majority of community involvement has been testing, which is completely uneffected by the SPL. As far as Sistina's requirement that all copyrights and such be signed over to them, this is nothing new. That requirement has been around just about since the dawn of time, and isn't anything the SPL added. Speaking as one who's written patches against GFS, and use it in a commercial environment, this requirement has been somewhat of a nuisance, especially since my employer doesn't want to give up copyright control. And thus, my code isn't part of the native distribution. This is nothing but a minor inconvenience, and certainly nothing life-threatening.
Sistina makes a great product, requiring royalties from people using GFS to make money is perfectly acceptable, and I'm surprised they didn't do it earlier. I for one welcome the SPL.
Linux is certainly better than wince for a lot of things.. in incredibly tiny spaces or dealing with systems w/o an MMU, it is not. Given the fact I can have a functional wince install in less space than I can get a useless 2.4 image to, is rather annoying. neutrino is kind of cute, but QNX has done a lot of stupid things with it as well *cough* qnx rtp *end cough*. I can't think of too many proprietary commercial RTOS'es that are even worth a damn.. VxWorks, pSOS, and LynxOS being examples of systems better left dead..
Where'd be the fun in that? I for one enjoy seeing lots of ignorant comments about using embedded systems for beowulf clusters, for mass storage, etc, etc. why should people have to be bothered with facts when ignorance is much more easily adopted? No doubt there will be lots of people who say "WinCE sucks! linux rules!", and those people are ignorant. those people have also likely never worked in an embedded environment before, and somehow feel that by owning a PDA they somehow have a clue regarding embedded applications. If slashdot were indeed unbiased, the majority of your ignorant slashdot users would start flaming slashdot for saying that there's something else out there that does a better job than linux in some areas. slashdot is only unbiased as long as it suits the purpose.
You seem to be missing the point of the whole thing. With no current standard in place, one is bound to happen sooner or later. Keep in mind that a good number of these companies that are part of the alliance are already part of the ELC, creating proprietary standards is not anything the alliance wants to do. Your concerns about software development being kept closed source are also unwarranted. The few software companies that are part of the alliance have a large portion of their business structure centered around open source software development and interaction with the open source community. And on top of all this.. what does it really matter? They're free to come up with what "standards" they want. It isn't going to make a difference if no one utilizes the standard.
You have got to be kidding. That's the kind of negligent moronic attitude that can be directly attributed to Columbine. Sheltering your child from things is ludicrous, they'll find out sooner or later, and it's your own fault for not discussing things with them first. On top of all this, you're also claiming that magic is bad. Perhaps you spent too much time criticizing instead of actually _reading_ the books. These books are not meant to be a factual reference, they are labelled as _fantasy_ books for a reason. If your child can't distinguish between fantasy and reality, you're again failing your job as a parent. If you had properly educated the child, these kind of things could be taken as what they were intended, entertainment value. A child is going to be exposed to violence regardless of your futile efforts, sheltering them from the reality of it isn't helping anything. Don't blame an excellent set of books for your inability to educated a child.
In reply to your Dreamcast, MIPS, and PowerPC comments, SH4, MIPS and PPC are all very heavily used in the embedded market, and likely will continue to be for years to come. While I doubt NewOS will ever get anywhere, it's always nice to see new open systems emerge that could be utilized in embedded space. Its ability to run on the Dreamcast (though while not really utilizing most of the hardware) is still fairly impressive considering what a nuisance the majority of the design of the Dreamcast is. *cough* maple bus *end cough*. So, while the Atari and Amiga are most definately dead, the rest of those are very much alive and kicking.
About as funny as people who criticize developers without a second thought yet couldn't write a line of code to save their life...
Not once did I mention anything regarding Intel's marketing tactics. If they want to claim that they have a superior product, they're free to do so. I am simply stating that x86 is an ancient, dead, and pathetic attempt at a decent architecture. Making it out to be the second coming is just pathetic. Taking a product that should've been killed off long ago and needlessly dragging it out in a desperate effort to look like innovation, isn't my idea of productivity. If you want to marvel in Intel's ability to beat a dead horse, no one is stopping you. Even AMD realized x86 was a dead end, hence the whole reasoning for licensing the ev6 bus from compaq. When you have to start licensing out features of other systems just to make your own look more worthwhile, you know it's time to start looking for a new product.
Intel didn't break the 1GHz 'barrier', this was already surpassed by countless other architectures before Intel finally got around to it. Sooner or later, Intel will realize they can only get so far by beating a dead horse. The only barriers that Intel has broken are the barriers of incompetence that would prohibit any competent corporation from dragging out a dead product needlessly and then trying to make it look revolutionary.