nbd is nice for some stuff but lacks fault-tolerance. Of course, you can run RAID, possibly several levels (say, a raid-6 on top of raid-1 or something) on top of nbd devices to trade space for fault-tolerance as much as you want, but you still lack flexibility. The advantage to RAID-over-nbd, on the other hand, is of course that you can do that right now if you want:] (And yes, the nbd server shouldn't be overly hard to run on Windows, one would think; it's rather simple...)
A better solution would work on a bit higher level, though. If a host goes down, it would be desirable to flexibly duplicate its data (from other mirrors and/or parity data) onto others. Possibly such a system could be created on top of nbd as well. Hell, maybe ZFS with an NBD pool could someday hack that, but seems to me they'd need to work out at least bug 4852783 first.
This is not even just a principle thing; aside from possible bug fixes they've done, they have possibly done useful work in porting Wine's DirectX implementation to Windows. (Possibly in the sense that one can't be sure if they've done it in a way that would please the Wine folks enough to integrate.)
This is currently not that big a deal, but as Wine gains DX10 functionality, it'd be quite funky to enable earlier Windows versions to run DX10 games, and therefore reduce the pressure for some people to upgrade to Vista.
multiXnest is designed to run several user X sessions on one computer. It's actually meant for multimonitor setups, but it can certainly do many users on one monitor too; just that it hasn't seemed like a good idea to anyone, and still doesn't to me.
The MS solution does seem to be intending to add some more integration than multiXnest can do, though, what with the moving mouse cursor over to the other session. I'm sure the innovation of being able to move one's cursor over the entire screen will be a patentworthy addition.
Indeed, and a nice thing about cryptophone is that they apparently provide protocol specs and invite others to be compatible (it would have to be reimplemented though).
There has also been talk of encrypted call support (would be nice if compatible with cryptophone, considering the published protocol) in OpenMoko, the open GNU/Linux-based phone OS, though no real work as of yet (hopefully only because the developer sales of the Neo1973 devices haven't properly started on schedule).
It is just a question of software regardless, what with the platform having no relevant restrictions. I suspect encrypted calls will be a reality on the platform before mass-market sales in the fall. I haven't the energy to do something that big myself, though, so just a guess from the sidelines.
This seems good for the upcoming OpenMoko-based ARM smartphone; although the project emphasizes native app development, fact is, there's a lot of mobile Java apps floating around. So once this is ported to OpenEmbedded/Moko, it should boost the platform's usefulness for many users.
So thanks, Sun, for this Christmas present. (Now just waiting for the phone to actually come out...:)
I couldn't seek funding for an idea without ip law because the investors might just steal my idea. Patents and ip law are the ONLY protection I have from being completely screwed over.
Not so. This can be done with suitable NDAs and/or pre-contracts (eg. negotiating a flat high percentage and no money up front for use of the ideas, with the understanding that this is just a stopgap and renegotiating will occur once people know what's on the table). This is of course impractical now, since investors would be taken aback with such agreements, however without patents this would have a high chance of becoming pretty much standard, so smart money would go with it.
Less integration, more overhead, esp. in 3d (though vmware seems to be doing some work in that area).
Plus, it's not all for existing Windows boxes. Some will be happier buying a new Windows-free one if they don't lose a critical app. And there's always resale of your old license, at least in sane jurisdictions...
Sure, HotSpot may be a bit faster than free JVMs, but the free ones do function well enough. Also, free Java compilers are already readily available. For a long time, the main issue has been the maturity of free class libraries (particularly their Swing/AWT implementations), and now Sun says they'll be getting around to releasing that around the end of 2007. Almost smells like timing the release to a date when they think Classpath will have most of it nailed anyway.
And then there's the license bit, but I shan't speculate on that uninformedly.
Slashdot quality moderating at work again, it seems.
Personally, I'm running Linux on amd64, and while I can run x86 apps with relative ease, there's a slight snag in that X.org's DRI drivers still don't handle 32-bit clients on 64-bit kernels.
"... government should be about fostering a dynamic and risk-taking economy, not preserving the certainty of anyone's business models."
The acronym "DMCA" springs to mind. Ah well.
Skype deserves trouble, but perchance not this
on
Net2phone Sues Skype
·
· Score: 1
Skype, what with their continuing attempt to monopolize the (consumer) VOIP market on a proprietary protocol, deserves some shit to happen, but I generally applaud software patent suits only when they're targeted at a pro-swpat lobbyist or someone just misusing the system like hell. I'm not sure Skype/E-Bay qualify at this point, so, bummer. Still, the targeting could be worse, too.
We don't like Burst, but we do like patent problems for the big guys. Lots of problems. More and more, until they get the point that patents are not a Good Idea, and tell their bought and paid for politicians so.
And if outfits like Burst have to make some patent bucks while doing it, so be it. It's a small price to pay, by Apple and Microsoft.
Parity schmarity, if they had features and speed greater than or comparable to the current free DRI solutions, it'd be a plus, even if the Windows drivers would be better.
This would of course be quite well and good. Still, it'd be nice if they also did at least some of the actual coding. (Obviously, I meant free as in speech drivers in the original post.)
Solid support is *much* more impotant to me then politics. I use Linux because it works for me and works well, same reason I use Nvidia cards under Linux.
I find it funny that you immediately followed this up with:
They who would give up an essential liberty for temporary security, deserve neither liberty or security.
The patent license seems to be purely a PR stunt without any real substance whatsoever.
First, they don't even assert that anything is covered by their patents.
Second, they assert a right to start acting up anyway if something that is covered by their patents ends up in the kernel.
Third, even if some patented thing in Linux was covered by this license, its use is only authorized in the Linux kernel as published on kernel.org (not even vendor- or self-patched versions), which is pretty useless.
Fourth, the above condition is incompatible with GPL's clause 7. It follows that if Nokia makes a credible patent claim on something that is in the kernel, then nobody has a valid license to distribute the kernel anymore until the patent issue is sorted out in the usual manner (that is, by getting a GPL-compatible license or working around it).
To sum it up, it seems they're just trying to shine their shield after bashing in some FFII heads here in the EU.
The patent license seems to be purely a PR stunt without any real substance whatsoever.
First, they don't even assert that anything is covered by their patents.
Second, they assert a right to start acting up if something that is covered by their patents ends up in the kernel.
And third, most importantly, if some patented thing were covered by this license, its use is only authorized in the Linux kernel, which is incompatible with GPL's clause 7. It follows that if Nokia makes a credible patent claim on something that is in the kernel, then nobody has a license to distribute the kernel anymore until the patent issue is sorted out in the usual manner.
To sum it up, they're just trying to shine their shield after bashing some FFII heads in here in the EU.
(Sorry for the dup, managed to log in now, and hey, that's slashdot for you.)
Agreed, at least in this case. If/when Hurd matures enough that it's stable and supports most of the stuff that's important to me, I'll seriously consider a switch due to architectural issues. Yes, I'm in the Tanenbaum camp.:)
Meanwhile, I'm not too vexed about running a Linux kernel. It's certainly not bad, and it is free. As I'm not in too much of a hurry, it isn't a big deal for me that the Hurd is taking its time (though I would of course prefer that some of the huge Linux effort would go towards Hurd instead, but it's not my choice to make).
As a side note, having FUSE or similar integrated into the main Linux kernel and proven to be relatively safe would make me a bit happier with Linux, though, since safe usermode filesystems are the first, though not the only practical level issue for me.
And before someone gives the standard "but they can't divulge their precious driver IP" blurb, I'll just note that screw their drivers if they just gave out some specs to drive their cards.
The minute there is a half-decent free driver for a modern 3d card, I'll buy that. For now, my newish AMD64 box sports a 9250. I didn't like buying ATI due to their current policies, but there's precious little options on the market these days; at least with this model they were decently open.
Was your appearance in Outer Limits as the space cadet (or something like that) who ended up not so much saving the day a conscious contrast for Wesley? Was it fun playing havoc with that expected typecast?
Slashdot Mirror
nbd is nice for some stuff but lacks fault-tolerance. Of course, you can run RAID, possibly several levels (say, a raid-6 on top of raid-1 or something) on top of nbd devices to trade space for fault-tolerance as much as you want, but you still lack flexibility. The advantage to RAID-over-nbd, on the other hand, is of course that you can do that right now if you want :] (And yes, the nbd server shouldn't be overly hard to run on Windows, one would think; it's rather simple...)
A better solution would work on a bit higher level, though. If a host goes down, it would be desirable to flexibly duplicate its data (from other mirrors and/or parity data) onto others. Possibly such a system could be created on top of nbd as well. Hell, maybe ZFS with an NBD pool could someday hack that, but seems to me they'd need to work out at least bug 4852783 first.
This is not even just a principle thing; aside from possible bug fixes they've done, they have possibly done useful work in porting Wine's DirectX implementation to Windows. (Possibly in the sense that one can't be sure if they've done it in a way that would please the Wine folks enough to integrate.)
This is currently not that big a deal, but as Wine gains DX10 functionality, it'd be quite funky to enable earlier Windows versions to run DX10 games, and therefore reduce the pressure for some people to upgrade to Vista.
multiXnest is designed to run several user X sessions on one computer. It's actually meant for multimonitor setups, but it can certainly do many users on one monitor too; just that it hasn't seemed like a good idea to anyone, and still doesn't to me.
The MS solution does seem to be intending to add some more integration than multiXnest can do, though, what with the moving mouse cursor over to the other session. I'm sure the innovation of being able to move one's cursor over the entire screen will be a patentworthy addition.
Indeed, and a nice thing about cryptophone is that they apparently provide protocol specs and invite others to be compatible (it would have to be reimplemented though).
There has also been talk of encrypted call support (would be nice if compatible with cryptophone, considering the published protocol) in OpenMoko, the open GNU/Linux-based phone OS, though no real work as of yet (hopefully only because the developer sales of the Neo1973 devices haven't properly started on schedule).
It is just a question of software regardless, what with the platform having no relevant restrictions. I suspect encrypted calls will be a reality on the platform before mass-market sales in the fall. I haven't the energy to do something that big myself, though, so just a guess from the sidelines.
This seems good for the upcoming OpenMoko-based ARM smartphone; although the project emphasizes native app development, fact is, there's a lot of mobile Java apps floating around. So once this is ported to OpenEmbedded/Moko, it should boost the platform's usefulness for many users.
So thanks, Sun, for this Christmas present. (Now just waiting for the phone to actually come out... :)
Not so. This can be done with suitable NDAs and/or pre-contracts (eg. negotiating a flat high percentage and no money up front for use of the ideas, with the understanding that this is just a stopgap and renegotiating will occur once people know what's on the table). This is of course impractical now, since investors would be taken aback with such agreements, however without patents this would have a high chance of becoming pretty much standard, so smart money would go with it.
Less integration, more overhead, esp. in 3d (though vmware seems to be doing some work in that area).
Plus, it's not all for existing Windows boxes. Some will be happier buying a new Windows-free one if they don't lose a critical app. And there's always resale of your old license, at least in sane jurisdictions...
Sure, HotSpot may be a bit faster than free JVMs, but the free ones do function well enough. Also, free Java compilers are already readily available. For a long time, the main issue has been the maturity of free class libraries (particularly their Swing/AWT implementations), and now Sun says they'll be getting around to releasing that around the end of 2007. Almost smells like timing the release to a date when they think Classpath will have most of it nailed anyway.
And then there's the license bit, but I shan't speculate on that uninformedly.
Disregarding your faulty English, FOSS mainline drivers would indeed provide this, being pretty much integrated into the kernel and X.org.
Slashdot quality moderating at work again, it seems.
Personally, I'm running Linux on amd64, and while I can run x86 apps with relative ease, there's a slight snag in that X.org's DRI drivers still don't handle 32-bit clients on 64-bit kernels.
It would be nice if sometimes "Linux support" meant something else than a binary for x86.
The acronym "DMCA" springs to mind. Ah well.
Skype, what with their continuing attempt to monopolize the (consumer) VOIP market on a proprietary protocol, deserves some shit to happen, but I generally applaud software patent suits only when they're targeted at a pro-swpat lobbyist or someone just misusing the system like hell. I'm not sure Skype/E-Bay qualify at this point, so, bummer. Still, the targeting could be worse, too.
We don't like Burst, but we do like patent problems for the big guys. Lots of problems. More and more, until they get the point that patents are not a Good Idea, and tell their bought and paid for politicians so.
And if outfits like Burst have to make some patent bucks while doing it, so be it. It's a small price to pay, by Apple and Microsoft.
Parity schmarity, if they had features and speed greater than or comparable to the current free DRI solutions, it'd be a plus, even if the Windows drivers would be better.
And yeah, free, open source, whatever.
This would of course be quite well and good. Still, it'd be nice if they also did at least some of the actual coding. (Obviously, I meant free as in speech drivers in the original post.)
So, how about Linux drivers? Free ones?
More incentive to use a completely closed and proprietary VOIP solution. This increases its appeal level to less than desirable!
The patent license seems to be purely a PR stunt without any real substance whatsoever.
First, they don't even assert that anything is covered by their patents.
Second, they assert a right to start acting up anyway if something that is covered by their patents ends up in the kernel.
Third, even if some patented thing in Linux was covered by this license, its use is only authorized in the Linux kernel as published on kernel.org (not even vendor- or self-patched versions), which is pretty useless.
Fourth, the above condition is incompatible with GPL's clause 7. It follows that if Nokia makes a credible patent claim on something that is in the kernel, then nobody has a valid license to distribute the kernel anymore until the patent issue is sorted out in the usual manner (that is, by getting a GPL-compatible license or working around it).
To sum it up, it seems they're just trying to shine their shield after bashing in some FFII heads here in the EU.
The patent license seems to be purely a PR stunt without any real substance whatsoever.
First, they don't even assert that anything is covered by their patents.
Second, they assert a right to start acting up if something that is covered by their patents ends up in the kernel.
And third, most importantly, if some patented thing were covered by this license, its use is only authorized in the Linux kernel, which is incompatible with GPL's clause 7. It follows that if Nokia makes a credible patent claim on something that is in the kernel, then nobody has a license to distribute the kernel anymore until the patent issue is sorted out in the usual manner.
To sum it up, they're just trying to shine their shield after bashing some FFII heads in here in the EU.
(Sorry for the dup, managed to log in now, and hey, that's slashdot for you.)
A single-server system on top of a microkernel does not a proper microkernel OS make :)
Agreed, at least in this case. If/when Hurd matures enough that it's stable and supports most of the stuff that's important to me, I'll seriously consider a switch due to architectural issues. Yes, I'm in the Tanenbaum camp. :)
Meanwhile, I'm not too vexed about running a Linux kernel. It's certainly not bad, and it is free. As I'm not in too much of a hurry, it isn't a big deal for me that the Hurd is taking its time (though I would of course prefer that some of the huge Linux effort would go towards Hurd instead, but it's not my choice to make).
As a side note, having FUSE or similar integrated into the main Linux kernel and proven to be relatively safe would make me a bit happier with Linux, though, since safe usermode filesystems are the first, though not the only practical level issue for me.
Hopefully.
Don't see it as likely, though.
And before someone gives the standard "but they can't divulge their precious driver IP" blurb, I'll just note that screw their drivers if they just gave out some specs to drive their cards.
The minute there is a half-decent free driver for a modern 3d card, I'll buy that. For now, my newish AMD64 box sports a 9250. I didn't like buying ATI due to their current policies, but there's precious little options on the market these days; at least with this model they were decently open.
Was your appearance in Outer Limits as the space cadet (or something like that) who ended up not so much saving the day a conscious contrast for Wesley? Was it fun playing havoc with that expected typecast?
;)
(I liked the episode, by the way