I would avoid any PCS service. The high frequencies of PCS (1900MHz) don't penetrate walls as well as cellular (800MHz). Verizon is mostly cellular except for a few states.
Also, I've heard from some network engineers who claim that the cellular carriers have better tower placement, in bigger cities, because they came well before PCS.
Holy s&^*t, ARM has conditional branching?? That sounds way cooler than the x86 unconditional branches.
Now that I've got the sarcasm out of me, a processor would be useless if there was no conditional branching. I believe you are mistaking "conditional branching" for "predicated execution" my friend.
Because if you offer 35k/year to someone who knows this stuff, whether CCIE or not, they're gonna laugh all the way to their interview with your competitor.
You see, existing large-scale routers require someone to really know the stuff inside and out. Just like how a UNIX admin really has to know his/her stuff to be a competent sysadmin.
Now they are often being replaced by an MSCE who can run a Win2k installation CD.
The key is removing complexity and make it appear cheap to the customer. This is how Microsoft sells their server products, not because they are more capable. Installing *any* server used to be very complex. Right now, configuring any Cisco router is complex. Someday, a product will be available that is easy to set up and is capable of doing the same job as that Cisco router. Just like how Win2k is capable of performing many (but not all) of the jobs once reserved for complex Unix installations.
The day will come when a competitor of Cisco develops a router that is easily configured by a Sysadmin who isn't a CCIE that knows the Cisco IOS terminal-driven interface like the back of their hand.
When that day comes, Cisco will feel the pressure and follow suit. Why pay a CCIE $80,000 to do your configuration when you can pay somebody $35,000? Same reason why many companies choose Windows instead of Linux. I'm not saying its right or even cost effective, but its a lot cheaper (up front) to hire someone to set up a Windows server than a Linux server. Especially for simple things like file and printer sharing.
There will always be a place for a CCIE expert, however their choice of places may become fewer.
Oh please. I didn't know whether to respond or mod you down on this one. I choose the former.
Of course Stingray (Windows for cellphones) uses an underlying RTOS. I interviewed with that group. We talked about it. All of the fancy UI/Windows stuff is in a low-priority task.
Just like with the Palm OS. The "Palm OS" doesn't actually run the PDA. It runs on top of a small RTOS kernel that handles interrupts, hardware drivers, and other real-time things that have little to do with the UI. In fact, I wouldn't be surprised if the same Palm OS runs on different RTOS kernels. In fact, I can almost guarantee you that the Qualcomm Palm thingy they had a few years ago likely did not use the same RTOS kernel as my Palm Vx.
And you comment about Cisco Certified Internet Engineers and Command-line interfaces is a joke, right? A CCIE doesn't make more money because he or she can use a terminal instead of dragging icons. They make more money because internetworking is more of a niche and is arguably more complex than setting up Windows. It also costs much more to become CCIE than MCSE. Trust me, Cisco's next generation routers will be configured remotely through a graphical interface. I guess when that happens, CCIE's will be making just as much as a MCSE, huh?
Get yourself a big electrolytic capacitor from radioshack and apply some voltage (20v should do) in opposite polarity. Make sure you cover your ears when you do;)
Actually Windows 3.1 is closer to an RTOS than you think. In most RTOS's, a task can starve any other task running at the same priority (or lower). Same as the cooperative multitasking model in Windows 3.1.
EPIC has some similarities to VLIW, but it is not VLIW. VLIW, in the purist sense, is where an instruction word specifies the behavior of every functional unit. If one unit can't be used during this instruction, then the instruction must contain no-op bits for that unit.
EPIC specifies instruction bundles, which are essentially groups of instructions that have no inter-dependencies. However the execution engine still has the ability to dispatch parts of the instruction bundle to any function unit it wants to as long as dependencies between instruction bundles are maintained.
This is a big difference. The problem with VLIW is that once you define an instruction set, you are essentially hand-cuffed to a specific micro-architecture.
Intel breaks those cuffs with EPIC. The micro-architecture of Itanium compared with Itanium2 is vastly different.
Yes, I used the same approach when I maintained my LFS system. Except I used a tool to manage the symlinks (I forgot the name of what I used though GNU has a similar tool). It still got tedious after awhile....at least in my opinion. Also symlinks do have a very small effect in speed and usually take up 4k per symlink depending on your choice of file system.
LFS is a great way to learn Linux. It truly helped take me to the next level of my personal understanding of how things work.
However I would never recommend it for a production system. Even using it for a personal workstation takes loads of time to manage. One doesn't appreciate package management until they have installed a LFS system!!! Of course one could always use RPM/APT/DEB after doing a LFS installation...
Because developers can create proprietary applications without buying a QT license. GTK is LGPL.
This really is important. My university is developing applications using GTK even though the students preferred QT. However finding $3000 per developer is not easy in a bureaucratically run institution.
The auto industry is required to make parts available for 10 years past the model year. Makes sense. Why not apply the same rule to software security fixes? Sure would do a lot to motivate better design.
Because software isn't really regulated. Think about it...can you build your own "open-source" automobile and operate it on public highways without it being approved by the Department of Transportation? I'm not sure, but I'm guessing you can't.
If the software industry is forced to make "security fixes" available for 10 years after initial release, then there will have to be some kind of authority that approves software packages (which of course would cost money) such that a company is legally responsible. Then there would be even less incentive for businesses to use open-source packages because their closed-source competitors have to legally provide 10 years worth of security fixes.
Thats comparing Apples to Oranges. Mozilla in Linux is especially a dog. Just resize windows quickly for a demonstration of this. Actually, I don't think anything is snappy in Linux except for maybe Opera. I use Linux all day at work on a 2ghz P4 w/ 1gb RAM. Therefore, Mozilla just never swaps for me at work. At home, I use Win2k. Having the UI/Windowing code in the kernel really does make a difference as far as overall responsiveness;)
I think my problems with Moz at home relate to my frequent game playing. Mozilla just gets swapped out to disk because its got so much resident. Meanwhile, IE is integrated into the OS and its pages are pinned in memory so that it can't get swapped. I'll admit that its not fair however Opera seems to manage just fine. I recently opened an IE window, an Opera window, and a Mozilla window. I then started Medal of Honor::Allied Assault and played for a few minutes. I quit, and opened all three windows (such that they were all visible). IE popped up nearly instantaneously. Opera was next. Mozilla was chugging on the disk for about 10 seconds before its contents were displayed.
By the way, vim has the option to integrate a python interpreter. I don't know about you, but I'd prefer a Python engine to a Lisp engine any day of the week:)
Man, I love tabbed browsing and pop-up add blocking but can't they make the UI a little faster and the memory usage lower?
I try to use Mozilla but I'm always drawn back to IE because its just snappier. I think that Microsoft pins the IE pages also. Even when I keep Mozilla resident, my system swaps like no tomorrow when using Mozilla on a PIII 866Mhz system w/ 384mb RAM
Its the same experience I have with emacs. I keep trying but always succumb to vi. vi is just more responsive.
I take exception to your statement that "many claim." It makes you look like a troll. If you wrote "some claim" I would accept that. Regardless, I disagree. From a technology perspective Linux device drivers can do anything that "that other OS" kernel can. The big difference is that linux drivers don't usually come precompiled. There is no good reason for this other then history.
I guess I hear this too much by being in the academic world (Computer Science). However I do know that many linux device drivers, such as the OpenAFS client file system module, rely on knowing the exact offsets of fields in the task_struct. There is no run-time method of retrieving these offsets as far as I know. So our administrators are forced to re-compile the openafs module everytime they apply a patch, or run a different kernel on a system that has changed the layout of task_struct. When maintaining systems that may serve different purposes, I can see why this would be an administration headache. We use a distributed filesystem where most binaries, such as device drivers, are maintained on a central server instead of being on each local box.
Constrast this scenario which Solaris, which is claimed to have a great device driver model by these people I speak of. I can use the same device driver binary I used in Solaris 2.6 as I use in Solaris 8.0.
My question is why custom kernels are needed anyways? Except for embedded applications, such as Tivo, why should the common user have to build a custom kernel to get certain hardware support? Is the Linux device driver model really flawed as many claim?
Certainly its nice for development, or experimental patches such as low-latency patches. However it often seems necessary to build a kernel to get certain modules or hardware functionality.
I'm reading a Computer Networks textbook right now. The author frequently points out that the success of the TCP/IP stack can be largely attributed to the BSD license because companies had access to a well-engineered network staff for free. Otherwise, OSI protocols might have been chosen as they were the "hot" research/development topic of the 80's.
Slashdot Mirror
I would avoid any PCS service. The high frequencies of PCS (1900MHz) don't penetrate walls as well as cellular (800MHz). Verizon is mostly cellular except for a few states.
Also, I've heard from some network engineers who claim that the cellular carriers have better tower placement, in bigger cities, because they came well before PCS.
Conditonal branching... It was a real joy.
Holy s&^*t, ARM has conditional branching?? That sounds way cooler than the x86 unconditional branches.
Now that I've got the sarcasm out of me, a processor would be useless if there was no conditional branching. I believe you are mistaking "conditional branching" for "predicated execution" my friend.
Because if you offer 35k/year to someone who knows this stuff, whether CCIE or not, they're gonna laugh all the way to their interview with your competitor.
You see, existing large-scale routers require someone to really know the stuff inside and out. Just like how a UNIX admin really has to know his/her stuff to be a competent sysadmin.
Now they are often being replaced by an MSCE who can run a Win2k installation CD.
The key is removing complexity and make it appear cheap to the customer. This is how Microsoft sells their server products, not because they are more capable. Installing *any* server used to be very complex. Right now, configuring any Cisco router is complex. Someday, a product will be available that is easy to set up and is capable of doing the same job as that Cisco router. Just like how Win2k is capable of performing many (but not all) of the jobs once reserved for complex Unix installations.
The day will come when a competitor of Cisco develops a router that is easily configured by a Sysadmin who isn't a CCIE that knows the Cisco IOS terminal-driven interface like the back of their hand.
When that day comes, Cisco will feel the pressure and follow suit. Why pay a CCIE $80,000 to do your configuration when you can pay somebody $35,000? Same reason why many companies choose Windows instead of Linux. I'm not saying its right or even cost effective, but its a lot cheaper (up front) to hire someone to set up a Windows server than a Linux server. Especially for simple things like file and printer sharing.
There will always be a place for a CCIE expert, however their choice of places may become fewer.
Oh please. I didn't know whether to respond or mod you down on this one. I choose the former.
Of course Stingray (Windows for cellphones) uses an underlying RTOS. I interviewed with that group. We talked about it. All of the fancy UI/Windows stuff is in a low-priority task.
Just like with the Palm OS. The "Palm OS" doesn't actually run the PDA. It runs on top of a small RTOS kernel that handles interrupts, hardware drivers, and other real-time things that have little to do with the UI. In fact, I wouldn't be surprised if the same Palm OS runs on different RTOS kernels. In fact, I can almost guarantee you that the Qualcomm Palm thingy they had a few years ago likely did not use the same RTOS kernel as my Palm Vx.
And you comment about Cisco Certified Internet Engineers and Command-line interfaces is a joke, right? A CCIE doesn't make more money because he or she can use a terminal instead of dragging icons. They make more money because internetworking is more of a niche and is arguably more complex than setting up Windows. It also costs much more to become CCIE than MCSE. Trust me, Cisco's next generation routers will be configured remotely through a graphical interface. I guess when that happens, CCIE's will be making just as much as a MCSE, huh?
I stopped reading this post as soon as you said "Pagemaker was a desktop publishing app that basically put Adobe on the map".
Sorry, but Adobe Pagemaker was formally known as Aldus Pagemaker.
Postscript is what put Adobe on the map. Then they developed Photoshop and the rest is history.
Get yourself a big electrolytic capacitor from radioshack and apply some voltage (20v should do) in opposite polarity. Make sure you cover your ears when you do ;)
Actually Windows 3.1 is closer to an RTOS than you think. In most RTOS's, a task can starve any other task running at the same priority (or lower). Same as the cooperative multitasking model in Windows 3.1.
Like Dell or not, they make some of the quietest Wintel machines there are.
There is a saying in computer science that any CS problem can be solved by adding another level of indirection.
There are "pointers" all over architectures. In fact, directory-based cache coherence protocols simply use an array of pointers to actual nodes.
This is no different than tacking 32-bit mode on top of a segmented 16-bit architecture, or the bizarre MMX/fp register sharing nonsense.
Hmmm...thats funny that AMD is doing something so similar with Hammer
If they aren't interested in performance, then why do they achieve pretty damn good SPEC numbers??
EPIC has some similarities to VLIW, but it is not VLIW. VLIW, in the purist sense, is where an instruction word specifies the behavior of every functional unit. If one unit can't be used during this instruction, then the instruction must contain no-op bits for that unit.
EPIC specifies instruction bundles, which are essentially groups of instructions that have no inter-dependencies. However the execution engine still has the ability to dispatch parts of the instruction bundle to any function unit it wants to as long as dependencies between instruction bundles are maintained.
This is a big difference. The problem with VLIW is that once you define an instruction set, you are essentially hand-cuffed to a specific micro-architecture.
Intel breaks those cuffs with EPIC. The micro-architecture of Itanium compared with Itanium2 is vastly different.
Yes, I used the same approach when I maintained my LFS system. Except I used a tool to manage the symlinks (I forgot the name of what I used though GNU has a similar tool). It still got tedious after awhile....at least in my opinion. Also symlinks do have a very small effect in speed and usually take up 4k per symlink depending on your choice of file system.
LFS is a great way to learn Linux. It truly helped take me to the next level of my personal understanding of how things work.
However I would never recommend it for a production system. Even using it for a personal workstation takes loads of time to manage. One doesn't appreciate package management until they have installed a LFS system!!! Of course one could always use RPM/APT/DEB after doing a LFS installation...
Cracking down on this will drive Microsofts target audience away. Perhaps they've shot themselves in the foot with this.
I bet that 90%, or more, of the people of own an Xbox don't even know what a "mod chip" is.
Because developers can create proprietary applications without buying a QT license. GTK is LGPL.
This really is important. My university is developing applications using GTK even though the students preferred QT. However finding $3000 per developer is not easy in a bureaucratically run institution.
I dislike Q3 as well, and any game that feels like Q3.
Best multiplayer FPS game I play right now is Medal of Honor:Allied Assault.
The demo is just horrible. Great graphics, horrible gameplay. Don't waste your time downloading this.
If anyone else thinks otherwise, I'd really like to hear about it.
The auto industry is required to make parts available for 10 years past the model year. Makes sense.
Why not apply the same rule to software security fixes? Sure would do a lot to motivate better design.
Because software isn't really regulated. Think about it...can you build your own "open-source" automobile and operate it on public highways without it being approved by the Department of Transportation? I'm not sure, but I'm guessing you can't.
If the software industry is forced to make "security fixes" available for 10 years after initial release, then there will have to be some kind of authority that approves software packages (which of course would cost money) such that a company is legally responsible. Then there would be even less incentive for businesses to use open-source packages because their closed-source competitors have to legally provide 10 years worth of security fixes.
Thats comparing Apples to Oranges. Mozilla in Linux is especially a dog. Just resize windows quickly for a demonstration of this. Actually, I don't think anything is snappy in Linux except for maybe Opera. I use Linux all day at work on a 2ghz P4 w/ 1gb RAM. Therefore, Mozilla just never swaps for me at work. At home, I use Win2k. Having the UI/Windowing code in the kernel really does make a difference as far as overall responsiveness ;)
:)
I think my problems with Moz at home relate to my frequent game playing. Mozilla just gets swapped out to disk because its got so much resident. Meanwhile, IE is integrated into the OS and its pages are pinned in memory so that it can't get swapped. I'll admit that its not fair however Opera seems to manage just fine. I recently opened an IE window, an Opera window, and a Mozilla window. I then started Medal of Honor::Allied Assault and played for a few minutes. I quit, and opened all three windows (such that they were all visible). IE popped up nearly instantaneously. Opera was next. Mozilla was chugging on the disk for about 10 seconds before its contents were displayed.
By the way, vim has the option to integrate a python interpreter. I don't know about you, but I'd prefer a Python engine to a Lisp engine any day of the week
Man, I love tabbed browsing and pop-up add blocking but can't they make the UI a little faster and the memory usage lower?
I try to use Mozilla but I'm always drawn back to IE because its just snappier. I think that Microsoft pins the IE pages also. Even when I keep Mozilla resident, my system swaps like no tomorrow when using Mozilla on a PIII 866Mhz system w/ 384mb RAM
Its the same experience I have with emacs. I keep trying but always succumb to vi. vi is just more responsive.
I take exception to your statement that "many claim." It makes you look like a troll. If you wrote "some claim" I would accept that. Regardless, I disagree. From a technology perspective Linux device drivers can do anything that "that other OS" kernel can. The big difference is that linux drivers don't usually come precompiled. There is no good reason for this other then history.
I guess I hear this too much by being in the academic world (Computer Science). However I do know that many linux device drivers, such as the OpenAFS client file system module, rely on knowing the exact offsets of fields in the task_struct. There is no run-time method of retrieving these offsets as far as I know. So our administrators are forced to re-compile the openafs module everytime they apply a patch, or run a different kernel on a system that has changed the layout of task_struct. When maintaining systems that may serve different purposes, I can see why this would be an administration headache. We use a distributed filesystem where most binaries, such as device drivers, are maintained on a central server instead of being on each local box.
Constrast this scenario which Solaris, which is claimed to have a great device driver model by these people I speak of. I can use the same device driver binary I used in Solaris 2.6 as I use in Solaris 8.0.
Regards...I wasn't trying to be a troll.
My question is why custom kernels are needed anyways? Except for embedded applications, such as Tivo, why should the common user have to build a custom kernel to get certain hardware support? Is the Linux device driver model really flawed as many claim?
Certainly its nice for development, or experimental patches such as low-latency patches. However it often seems necessary to build a kernel to get certain modules or hardware functionality.
Any comments on the Linux device driver model?
I'm reading a Computer Networks textbook right now. The author frequently points out that the success of the TCP/IP stack can be largely attributed to the BSD license because companies had access to a well-engineered network staff for free. Otherwise, OSI protocols might have been chosen as they were the "hot" research/development topic of the 80's.