Sun didn't do CDE. They actually resisted it to the bitter end, pushing their OpenView environment over it, until they finally gave in only to have CDE be declared unuseable. Mind you, OpenView was even worse, but that's beside the point. CDE's failings were purely thanks to design by committee.
There are some minor issues with the current implementation of X, that I think alot of people keep confusing with the network transparency thing. Yes, I agree that you do need some kind of user space X server, and that there's probably no better way to talk to it than a UNIX domain pipe. However I really don't think X's driver model should be in user space. At the moment it does use all sorts of acceleration (3d and 2d), but it accesses these features of the card by mmaping/dev/mem. Using this alone lets you set any frame buffers or io ports, but the X server can't sleep on a hardware interrupt. This results in some busy waiting where it wouldn't need to if the graphics driver were in-kernel.
And the real reason apple didn't go with X is because they wanted to use the OpenStep API and that's written to use a display postscript backend. It was easier to change those slightly to use a similar display pdf backend then it would be to rewrite them to use the completely different architecture of X (X is missing things like vector manipulations, resolution intependant objects and generally everything display postscript/pdf does well).
I don't know about the Basques, but the Celts tell stories of invading Ireland and the British Isles and killing off the people that used to be there. And the Gauls were an offshoot of the Celts, and settled in areas previously occupied by the Basques (who were once much more widespread then they are now). Mind you these things happened so long ago that no records are completely accurate and who knows what really happened. Regardless, these peoples don't even claim to be the original ones.
None of the PowerPCs in Apple machines so far are 64 bit. The PowerPC is, however a branch of the POWER architecture of chips that IBM uses in its RS/6000 and eServer line of high end UNIX servers. Of those chips, the most recent (the Power4) is 64 bit. There are rumors of Apple moving to an IBM made 64 bit chip for the next generation of Macs, as apposed to the ever delayed Motorolla G5.
The fact that linux is opensource and can be stripped down *does* make it an ideal embedded os. At least in its stripped down form its smaller than any other embedded os that still has a VM (this doesn't include palm as it was a single tasking os). And as for crusoe, it's only the lowest heat and energy intake for an intel compatible processor. The XScale is the next generation of the ARM processor, a completely different instruction set that just happens to be really easy to design a low power cpu out of.
So you're saying that a company like Nintendo, who owns their own console, and want's to make a great game (Metroid Prime) should make the game for systems they compete with? Really, thats kind of stupid. I can see where you might be upset about halo, because it was already in development for other systems by an independant developer, and MS bought them just to silence those ports. But nintendo themselves (or rather a second party being funded and overseen by nintendo) developed this game specifically to bolster their console. It would be stupid of them to do otherwise. Or do you think that console makers shouldn't also develop games? Thing is, some of the best games out there are developed by console makers because they know the hardware they're working on better than anything else.
And metroid prime is one of the best games I've ever played and well worth a gamecube.
Re:A significantly larger number of systems?
on
Linux Is Cheaper
·
· Score: 2
Its not a complete solution, but take a look at rdesktop http://www.rdesktop.org/ , Its saved my life (or at least significant portions of my free time) many a time.
Re:A significantly larger number of systems?
on
Linux Is Cheaper
·
· Score: 2
Agh, for some reason that was posted anonymous. Meant to be me.
Re:A significantly larger number of systems?
on
Linux Is Cheaper
·
· Score: 5, Insightful
And this is why windows admins don't understand how Linux admins manage that many machines. VNC is a bad hack that you'd only use if you needed cross platform remote display. You do realize that if you really needed to remotely run graphical apps, X programs are inherrently remote displayable. Not only that, but the power of linux administration is the fact that you *never* need to run a graphical program. If I have a farm of 150 web servers, I can make an httpd.conf on one of them and with a 3 line shell script (typed interactively on the command line) scp it to every machine on the network and restart those webservers. Or I could tie it to a cron job, or a script monitoring/var/log/messages for a certain event like a service going down, or have procmail do it when it recieves an email. Yes I know with the right packages and third party tools windows is scriptable as well, but its not designed around scriptability. The simplest way to permanently change the host name of a machine is still to go in with vnc (or terminal services, I have to be fair) and open up the network configuration dialog and change it. Which is easy, I admit. But over a slow network its infuriating. I can ssh in from a modem and edit/etc/sysconfig/network with vi and it'd be just as responsive as if I was on the local network. Again, I see the benefit of both approaches, but never would I personally want to administer a significantly large windows data center when a unix based solution was feasable.
They didn't lie. They never said they were going to support other processor architechtures. The moment they released the details about code morphing, people everywhere (slashdot especially) started speculating on how they *could* support other architectures. After months, speculations turned to rumors and rumors got treated as fact. Someone from Transmeta even said in a press release at one point that the crusoe's instruction set, while adaptable, was designed with emulating x86 in mind. It could emulate other architechtures, but it wouldn't be efficient at all. The code morphing was just there to allow the core to be simpler, smaller, cheaper and lower powered. And he specifically stated that while it might be a possibility to go to other architectures, it wasn't a definite part of their plans.
Re:11: Thou shalt not skywatch in the WA state win
on
Geminid Meteor Shower
·
· Score: 2
As a former (actual) Washingtonian, and current Washington state resident, I take offence at the use of the term Washingtonian. A Washingtonian is someone who lives in Washington DC.
Er something.
Re:I wonder how much of this is quality . . .
on
Critics Pan Nemesis
·
· Score: 2
AGH! I didn't know they made Nightfall (one of my favority Asimov stories) into a movie already, much less a crappy one. I had always held onto the idea that this was one of the few stories that *could* be made into a decent movie and *should* if there is any justice in the world. But they already ruined it!
No the real issue with the wire busses is the fact that none of the usefull ones are wire busses. I live in the U district. I work downtown. I could take the nice quiet electric running 70 or 7, but that would take 45 minutes. The 71 72 and 73 are big smelly and gas powered. But they're express so they get me there in 15 minutes. I eat lunch with a friend in lower queen anne. All the busses that go up first are gas powered. I often go to the best buy in northgate. The 66 and 75 are gas powered. Now where are the electric busses?
Re:How does it compare on windows?
on
Mesa 5.0 Released
·
· Score: 2
Considering the fact that Mesa has a modular rendering system, and the DRI project was made specifically to be a backend for Mesa, your definition of "vanilla" Mesa is kind of limited. The simple fact is that for every video card except nVidia's, the hardware acceleration on linux is done in large part through Mesa, and this will allow existing DRI accelerated cards to get access to OpenGL 1.4 features.
Mesa was originally a software opengl implementation, yes, but now its renderer is modular. In particular, the dri project is built around drivers implementing only extremely low level primitives which are plugged into Mesa which provides the full OpenGL stack to access these drivers. So for hardware acceleration on Linux, every driver uses Mesa except nVidia's which includes its own OpenGL stack.
Mostly because the was unix VMs are designed is much more efficient at multiple process programs than windows is. Windows started doing threads long before smp was all that common. They did it because multi process was slow as hell. But for 90% of tasks it worked just fine in linux. And its not like linux is just now moving to a thread model. Its just making the existing one (which worked well until you scale to many many threads) a bit better. And by better I don't mean similar to windows performance, I mean similar to solaris (which has threading from the gods).
On that note, I'd be curious to see what the speed difference between Carbon and Cocoa apps are (if the dynamic typing of ObjectiveC is as big a burdon as people say it is).
Thing its missing though is all the graphic accelerators that the gba has. Thanks to alot of custom chips, that little thing can do far more than its processor should allow. And while the cpu on this thing is impressive, I don't know if its enough to handle all that.
Actually does anyone know whether this version will have a real AIO implementation? The current version of AIO is implemented completely in userspace (in libc, emulated ontop of pthreads). Will we finally have real asynchronous IO?
Slashdot Mirror
Sun didn't do CDE. They actually resisted it to the bitter end, pushing their OpenView environment over it, until they finally gave in only to have CDE be declared unuseable. Mind you, OpenView was even worse, but that's beside the point. CDE's failings were purely thanks to design by committee.
Actually almost all crays are unix boxes running a version of unix they call UNICOS.
True productivity is having an elisp macro do the work for you.
Mandrake is for newbies. Slackware is for newbies that don't want to stay newbies.
There are some minor issues with the current implementation of X, that I think alot of people keep confusing with the network transparency thing. Yes, I agree that you do need some kind of user space X server, and that there's probably no better way to talk to it than a UNIX domain pipe. However I really don't think X's driver model should be in user space. At the moment it does use all sorts of acceleration (3d and 2d), but it accesses these features of the card by mmaping /dev/mem. Using this alone lets you set any frame buffers or io ports, but the X server can't sleep on a hardware interrupt. This results in some busy waiting where it wouldn't need to if the graphics driver were in-kernel.
And the real reason apple didn't go with X is because they wanted to use the OpenStep API and that's written to use a display postscript backend. It was easier to change those slightly to use a similar display pdf backend then it would be to rewrite them to use the completely different architecture of X (X is missing things like vector manipulations, resolution intependant objects and generally everything display postscript/pdf does well).
Um.... :0 :1 :2
startx --
startx --
startx --
Has always worked for me.
I don't know about the Basques, but the Celts tell stories of invading Ireland and the British Isles and killing off the people that used to be there. And the Gauls were an offshoot of the Celts, and settled in areas previously occupied by the Basques (who were once much more widespread then they are now). Mind you these things happened so long ago that no records are completely accurate and who knows what really happened. Regardless, these peoples don't even claim to be the original ones.
None of the PowerPCs in Apple machines so far are 64 bit. The PowerPC is, however a branch of the POWER architecture of chips that IBM uses in its RS/6000 and eServer line of high end UNIX servers. Of those chips, the most recent (the Power4) is 64 bit. There are rumors of Apple moving to an IBM made 64 bit chip for the next generation of Macs, as apposed to the ever delayed Motorolla G5.
The fact that linux is opensource and can be stripped down *does* make it an ideal embedded os. At least in its stripped down form its smaller than any other embedded os that still has a VM (this doesn't include palm as it was a single tasking os). And as for crusoe, it's only the lowest heat and energy intake for an intel compatible processor. The XScale is the next generation of the ARM processor, a completely different instruction set that just happens to be really easy to design a low power cpu out of.
So you're saying that a company like Nintendo, who owns their own console, and want's to make a great game (Metroid Prime) should make the game for systems they compete with? Really, thats kind of stupid. I can see where you might be upset about halo, because it was already in development for other systems by an independant developer, and MS bought them just to silence those ports. But nintendo themselves (or rather a second party being funded and overseen by nintendo) developed this game specifically to bolster their console. It would be stupid of them to do otherwise. Or do you think that console makers shouldn't also develop games? Thing is, some of the best games out there are developed by console makers because they know the hardware they're working on better than anything else.
And metroid prime is one of the best games I've ever played and well worth a gamecube.
Have you seen the quicktime support in mplayer?
Its not a complete solution, but take a look at rdesktop http://www.rdesktop.org/ , Its saved my life (or at least significant portions of my free time) many a time.
Agh, for some reason that was posted anonymous. Meant to be me.
And this is why windows admins don't understand how Linux admins manage that many machines. VNC is a bad hack that you'd only use if you needed cross platform remote display. You do realize that if you really needed to remotely run graphical apps, X programs are inherrently remote displayable. Not only that, but the power of linux administration is the fact that you *never* need to run a graphical program. If I have a farm of 150 web servers, I can make an httpd.conf on one of them and with a 3 line shell script (typed interactively on the command line) scp it to every machine on the network and restart those webservers. Or I could tie it to a cron job, or a script monitoring /var/log/messages for a certain event like a service going down, or have procmail do it when it recieves an email. Yes I know with the right packages and third party tools windows is scriptable as well, but its not designed around scriptability. The simplest way to permanently change the host name of a machine is still to go in with vnc (or terminal services, I have to be fair) and open up the network configuration dialog and change it. Which is easy, I admit. But over a slow network its infuriating. I can ssh in from a modem and edit /etc/sysconfig/network with vi and it'd be just as responsive as if I was on the local network. Again, I see the benefit of both approaches, but never would I personally want to administer a significantly large windows data center when a unix based solution was feasable.
They didn't lie. They never said they were going to support other processor architechtures. The moment they released the details about code morphing, people everywhere (slashdot especially) started speculating on how they *could* support other architectures. After months, speculations turned to rumors and rumors got treated as fact. Someone from Transmeta even said in a press release at one point that the crusoe's instruction set, while adaptable, was designed with emulating x86 in mind. It could emulate other architechtures, but it wouldn't be efficient at all. The code morphing was just there to allow the core to be simpler, smaller, cheaper and lower powered. And he specifically stated that while it might be a possibility to go to other architectures, it wasn't a definite part of their plans.
As a former (actual) Washingtonian, and current Washington state resident, I take offence at the use of the term Washingtonian. A Washingtonian is someone who lives in Washington DC.
Er something.
AGH! I didn't know they made Nightfall (one of my favority Asimov stories) into a movie already, much less a crappy one. I had always held onto the idea that this was one of the few stories that *could* be made into a decent movie and *should* if there is any justice in the world. But they already ruined it!
No the real issue with the wire busses is the fact that none of the usefull ones are wire busses. I live in the U district. I work downtown. I could take the nice quiet electric running 70 or 7, but that would take 45 minutes. The 71 72 and 73 are big smelly and gas powered. But they're express so they get me there in 15 minutes. I eat lunch with a friend in lower queen anne. All the busses that go up first are gas powered. I often go to the best buy in northgate. The 66 and 75 are gas powered. Now where are the electric busses?
Considering the fact that Mesa has a modular rendering system, and the DRI project was made specifically to be a backend for Mesa, your definition of "vanilla" Mesa is kind of limited. The simple fact is that for every video card except nVidia's, the hardware acceleration on linux is done in large part through Mesa, and this will allow existing DRI accelerated cards to get access to OpenGL 1.4 features.
Mesa was originally a software opengl implementation, yes, but now its renderer is modular. In particular, the dri project is built around drivers implementing only extremely low level primitives which are plugged into Mesa which provides the full OpenGL stack to access these drivers. So for hardware acceleration on Linux, every driver uses Mesa except nVidia's which includes its own OpenGL stack.
Mostly because the was unix VMs are designed is much more efficient at multiple process programs than windows is. Windows started doing threads long before smp was all that common. They did it because multi process was slow as hell. But for 90% of tasks it worked just fine in linux. And its not like linux is just now moving to a thread model. Its just making the existing one (which worked well until you scale to many many threads) a bit better. And by better I don't mean similar to windows performance, I mean similar to solaris (which has threading from the gods).
On that note, I'd be curious to see what the speed difference between Carbon and Cocoa apps are (if the dynamic typing of ObjectiveC is as big a burdon as people say it is).
Not to mention C++, ObjectiveC, Perl, Java, C# and any other language that's borrowed syntax from C (ie. most languages made since the 80s).
Thing its missing though is all the graphic accelerators that the gba has. Thanks to alot of custom chips, that little thing can do far more than its processor should allow. And while the cpu on this thing is impressive, I don't know if its enough to handle all that.
Actually does anyone know whether this version will have a real AIO implementation? The current version of AIO is implemented completely in userspace (in libc, emulated ontop of pthreads). Will we finally have real asynchronous IO?