What would be REALLY interesting is if the kernel source had a script triggered by something like "make update" that downloaded and installed the diff.
Oops... don't know exactly how that happenned. I was saying "It doesn't matter if the RAMDAC is good or not, 'cause bits is bits." or something else obvious like that. Frankly, even on CRT monitors, I believe the DAC would belong on the Monitor's end, as it could be optimized for the display characteristics of the monitor. It rather irks me that DVI isn't more common on even damn LCD monitors! I want to see some DVI-only models, as they'd likely be cheaper.
That used to be true, but the gap is closing. Most GeForce FX cards have pretty fantastic RAMDACs. Yeah, the Parhelia does look a hair better (but only on your primary monitor, if you're using more than one with it). NVIDIA beats Matrox in price, performance, and driver quality.
Besides, I'm not going to be using an analog output for too long... DVI kills the whole "2d quality" argument; the color values are passed digitally via a TMDS transmitter. Doesn't matter if
I've never seen backup software that didn't suck. I used to work for a company that will remain nameless that makes backup software. They were developing a new GUI for the product, and my specialty for 7 years was HCI (Human-Computer Interaction), so of course they put me on the team that maintains low-level device interfaces on UNIX platforms.
Anyway, I was the only one on the team that wasn't there when the software was written. Every damn function had at least one parameter that was a poorly named and undocumented void *. In fact, there was NO documentation of a single function parameter in the entire 100million+ lines of code that was our product source tree.
In addition to this, we had no Internet access on our desks. So when I wanted to look up docs for the hardware or OS I was working on so I could interface with the appropriate IOCTL's and such, I had to wait in line at the lone Internet terminal that was shared by 40+ developers, break it up into 10MB chunks if it was big, and email it to myself through our internal corporate mail system and pray it didn't get caught by the virus/spam filters.
5 months of this job nearly made me cry. Thank God I found another in such short order.
50 million stoners in the US & we can't vote out the War on (some) Drugs? WTF?
Remember -- we don't win on the popular vote. And many stoners aren't of voting age. The people that vote the most are senior citizens.
I'm speaking as a non-stoner who thinks it's silly to waste tax-dollars fighting drugs. I say legalize crack and heroin -- let the morons kill themselves. And require all health-care providers to nullify coverage of anyone who uses dangerous narcotics. The problem will go away withing 2 generations.
Yes, post anonymously when telling the unvarnished truth about Apple. There's an agenda here at Slashdot... in case you hadn't noticed.
I think that more to the point the poster is a Dell employee and doesn't want to get in trouble at work. People around here are way too quick to yell "Help! Help! I'm being opressed!"
Whats not to like about that? Why do I need Linux?
The future. The Netware kernel is aging, and cost of continuing hardware support is high. By using Linux, Novell gets a wider range of hardware support largely for free. They also get to capitalize on other open source software, like Samba, rather than implementing thier own CIFS layer.
If it ain't broke, don't fix it, by any means! However, in 5-10 years when you outgrow your current setup, you'll be happy that Novell switched to the Linux kernel for Netware. They can now focus on innovating more in the userland stuff, and take comfort in the fact that almost all new hardware they'd want to use will be supported by the community.
Wrong -- it's available under two liscences: GPL and QPL. If you accept it under the GPL, you CANNOT use it in closed source projects -- direct violation of the liscence. Under the QPL you can, but to get the QPL Liscence, you have to pay for it.
Re:When you write a kernel the world can use...
on
Linus on Linux in 1994
·
· Score: 2, Insightful
Yes, because we all know that the executives that would care about such things delve through pages of source code, and are familiar with tools like grep.
I'm sorry, but I believe you're wrong. Part of what makes the kernel so great is Linus's personal tastes. Lots of people write code -- many of which write different implementations of the same thing. Linus ultimately still determines:
1) What code goes into the official kernel 2) What direction it takes
Stuff that Linus thinks sucks or isn't ready doesn't go in. Yes, Linus deferrs much of this to his "liutennants", and yes, there are other forks of the kernel tree, but most of them try to stay in line with Linus' tree. There are no true splits that have any momentum behind them, like in the BSD's. Every other popular tree is the Linus tree +/- some patches. And Linus, even if only as a figurehead, keeps this together because of the amount of respect the community has for him. He didn't write every line of code, but he defined, and still continues to define, what Linux is.
Not familiar with the 3Ware controller. Honestly, I'd stick with 2.4.25 for now. I've been using 2.6 on the desktop since 2.6.1, and am happy with it, but have had a few problems (nForce onboard ethernet driver isn't very mature, and has cause one or two kernel panics).
Depends on your hardware. If you're using SMP, it will help; the scheduler is a bit better. Also has better support for the P4/Xeon's Hyperthreading. Overall lower latency operation as well. udev is a nice upgrade from devfs. No need to use proprietary sound or ethernet drivers on an nForce platform. If you're dealing with LOTS of traffic, it will perform better under stress.
No one is forcing any of these programmers to write code for free. No one is forcing me to GPL the stuff I write outside of work. If someone makes money off my code, more power to them. I know for a fact that some Linux-based smart-phones that sell for hundreds of dollars in Asia contain some code I've written. I think that's great.
And IBM, HP, Red Hat, SuSE, and Trolltech, among others, DO hire open source developers. IBM has donated millions to the community. These companies do give back to us. Maybe they don't give every damn contributer a paycheck, but they donated RCU, JFS, and a number of other things.
And you know what, I don't have a problem not getting paid for service trips to third world countries, or the 20+ hours a week I spend doing youth ministry.
No one is forcing programmers to contribute to the open source community. We do so to scratch an itch, to learn, to give back to a community that has given us so much, and because some of us think it's the right thing to do. If Longhorn is costing Microsoft that much, they're doing something wrong. Imagine if they opened their code, and had a community of millions helping them with code audits. Frankly, I want good, reliable software that I can trust. I can't trust proprietary software; I don't know what it's doing. Internet Explorer could, in theory, be capturing my credit card number in web forms and sending them to Bill Gates. Although I suspect that this is not true, there's no way for me to find out, because running a fucking network sniffer and trying to crack any encryption used is likely against the EULA.
Sorry, I disagree. Red Hat sells RHEL for a lot of money, and lots of companies are willing to pay. I don't follow Stallman -- I think he's too extreme, but I have a strong preference to free software. Why? Because I can use it any way I want. I can modify it. Yeah, end users don't give a damn about that, but it's nice to know that if they change their mind, they can. Also, if they want to hire me to modify free software for their needs, they can. I've done so. I can't do that with proprietary software.
Do I think all software should be free? I'd like so, but if people don't want to join the free software club, it's their choice. I largely don't care. That doesn't mean I have to use it, and outside of work (I have to use Flash and Windows for lots of stuff), I don't.
Gates is not fanatical about all software being 'non-free'. Proprietary yes. Microsoft produces plenty of software that runs on Windows and OSX that's (surprise, surprise) actually free.
Wrong definition of "free". When Stallman talks of "free", it has NOTHING to do with price. It's "free" as in "free speech" rather than "free beer".
My guess is that the initial rev will, in fact, be slower on 64-bit. Microsofts compilers are new to 64-bitness, and a reasonable amount of memory bandwidth will be wasted on larger ints. On the other hand, in 64-bit mode on the Opteron, there are twice as many GPR's, so it could wind up being faster. My bets are still on slower largely due to immaturity of the platform.
Belkin makes a model, but it ain't cheap. MSRP is $325, and it comes with no cables. The cables are $80 each. So if you want to connect 4 PC's, that's a whopping $645.
Yes, you are supposed to use a type of active splitter/hub type device to split SATA into multiple channels.
Personally, I have no need for external drives. SATA fixes clutter now inside my case, so I like it. Plus, there are affordable 10,000RPM SATA drives available, and I like that.
Saying "everybody" hates GIMP's interface is a rather foolish statement. Photoshop on the Mac doesn't use MDI and Photoshop on Windows does. It does for legacy reasons.
And when people talk about MDI, they mean "windows within windows", not "an interface that allows you to view multiple documents in one window". Under the current accepted definition of MDI, tabbed browsing doesn't fit the bill. Tabbed browsing makes sense for a couple of reasons:
- you're talking about viewing different clumps of similar types of data, rather than editing - the tabs are always visible, unless you have way, way too many open - The metaphor is actually a bookmark metaphor
Tabbed browsing, in many ways, is inefficient as well. Honestly, the best approach to this I've seen was BeOS, where one could stack multiple windows on top of each other, and then treat the titlebars as tabs. This way, normal switching between windows hotkeys still work, rather than having to memorize a new set. They behaved both like windows AND tabs.
As for MS-Windows-style MDI, with current toolkits, there is no point. There are better, easier ways of doing it (like my suggestion in another post -- have all tool palletes raised when the document gains focus). And also, in the GIMP 2.0 betas, the UI is notably improved, particularly with the new dockable tool palletes. I can still make room for my terminal view and my document, but have the palletes dock at the side of the screen.
And I do accept Photoshop does fine its way. I just think, IMHO, that GIMP happens to be better in this respect.
SATA has a lot of advantages... the cables can be made cheaply, the standard form-factors of drives are designed to slide in/out of a backplane (which no one has made yet) and it has room for growth. The next gen version will support 300MB/s and splitting a channel into 15 other channels. There's also talk of an external version of the standard. Plus, it's largely backward compatible with ATA.
As for SAS (Serial attached SCSI), it has some nice features as well. First and foremost, it's backward compatible with SCSI and compatible with SATA. That means you can take an SATA drive and plug it into a SAS controller just fine. The inverse is not true; SAS drives won't work on SATA. This means if someone wants to make a common SAS backplane for a PC case, one can just slide in any SAS or SATA drive in any combination. Very cool, as far as I'm concerned.
The throughput would increase with striping, but not latency. Still, a single 7200RPM full-size drive would be faster than two striped 5400RPM 2.5" drives, and far less expensive.
As far as reliability, yes, actually being in a laptop adds to that. But a hard drive has moving parts, and making parts larger generally means they'll wear better. In laptop drives they use smaller bearings, smaller read-heads, thinner spindle, and thinner platter. Something is going to give over time, and over significantly less time than on 3.5" drives. And frankly, given the price and capacity difference, I'd much rather see one or two slots for 3.5" drives than a dozen slots for laptop drives. 3.5" drives currently top out at 300GB @7200RPM or 74GB @10,000RPM. 2.5" drives top out at 80GB @5400RPM and 60GB @7200RPM. There are still a number of laptop drives on the market running at 4000RPM.
Slashdot Mirror
The reason is that they are higher-end models. Low-end TV's don't come without speakers, because no one would buy them.
If, however, all graphics cards for the next 5 years came with DVI outs and NOT VGA, the DVI-only flat-panels would be the norm, and thus cheaper.
What would be REALLY interesting is if the kernel source had a script triggered by something like "make update" that downloaded and installed the diff.
Oops... don't know exactly how that happenned. I was saying "It doesn't matter if the RAMDAC is good or not, 'cause bits is bits." or something else obvious like that. Frankly, even on CRT monitors, I believe the DAC would belong on the Monitor's end, as it could be optimized for the display characteristics of the monitor. It rather irks me that DVI isn't more common on even damn LCD monitors! I want to see some DVI-only models, as they'd likely be cheaper.
That used to be true, but the gap is closing. Most GeForce FX cards have pretty fantastic RAMDACs. Yeah, the Parhelia does look a hair better (but only on your primary monitor, if you're using more than one with it). NVIDIA beats Matrox in price, performance, and driver quality.
Besides, I'm not going to be using an analog output for too long... DVI kills the whole "2d quality" argument; the color values are passed digitally via a TMDS transmitter. Doesn't matter if
I've never seen backup software that didn't suck. I used to work for a company that will remain nameless that makes backup software. They were developing a new GUI for the product, and my specialty for 7 years was HCI (Human-Computer Interaction), so of course they put me on the team that maintains low-level device interfaces on UNIX platforms.
Anyway, I was the only one on the team that wasn't there when the software was written. Every damn function had at least one parameter that was a poorly named and undocumented void *. In fact, there was NO documentation of a single function parameter in the entire 100million+ lines of code that was our product source tree.
In addition to this, we had no Internet access on our desks. So when I wanted to look up docs for the hardware or OS I was working on so I could interface with the appropriate IOCTL's and such, I had to wait in line at the lone Internet terminal that was shared by 40+ developers, break it up into 10MB chunks if it was big, and email it to myself through our internal corporate mail system and pray it didn't get caught by the virus/spam filters.
5 months of this job nearly made me cry. Thank God I found another in such short order.
50 million stoners in the US & we can't vote out the War on (some) Drugs? WTF?
Remember -- we don't win on the popular vote. And many stoners aren't of voting age. The people that vote the most are senior citizens.
I'm speaking as a non-stoner who thinks it's silly to waste tax-dollars fighting drugs. I say legalize crack and heroin -- let the morons kill themselves. And require all health-care providers to nullify coverage of anyone who uses dangerous narcotics. The problem will go away withing 2 generations.
- Yes, post anonymously when telling the unvarnished truth about Apple. There's an agenda here at Slashdot... in case you hadn't noticed.
I think that more to the point the poster is a Dell employee and doesn't want to get in trouble at work. People around here are way too quick to yell "Help! Help! I'm being opressed!"Whats not to like about that? Why do I need Linux?
The future. The Netware kernel is aging, and cost of continuing hardware support is high. By using Linux, Novell gets a wider range of hardware support largely for free. They also get to capitalize on other open source software, like Samba, rather than implementing thier own CIFS layer.
If it ain't broke, don't fix it, by any means! However, in 5-10 years when you outgrow your current setup, you'll be happy that Novell switched to the Linux kernel for Netware. They can now focus on innovating more in the userland stuff, and take comfort in the fact that almost all new hardware they'd want to use will be supported by the community.
Wrong -- it's available under two liscences: GPL and QPL. If you accept it under the GPL, you CANNOT use it in closed source projects -- direct violation of the liscence. Under the QPL you can, but to get the QPL Liscence, you have to pay for it.
Yes, because we all know that the executives that would care about such things delve through pages of source code, and are familiar with tools like grep.
I'm sorry, but I believe you're wrong. Part of what makes the kernel so great is Linus's personal tastes. Lots of people write code -- many of which write different implementations of the same thing. Linus ultimately still determines:
1) What code goes into the official kernel
2) What direction it takes
Stuff that Linus thinks sucks or isn't ready doesn't go in. Yes, Linus deferrs much of this to his "liutennants", and yes, there are other forks of the kernel tree, but most of them try to stay in line with Linus' tree. There are no true splits that have any momentum behind them, like in the BSD's. Every other popular tree is the Linus tree +/- some patches. And Linus, even if only as a figurehead, keeps this together because of the amount of respect the community has for him. He didn't write every line of code, but he defined, and still continues to define, what Linux is.
Not familiar with the 3Ware controller. Honestly, I'd stick with 2.4.25 for now. I've been using 2.6 on the desktop since 2.6.1, and am happy with it, but have had a few problems (nForce onboard ethernet driver isn't very mature, and has cause one or two kernel panics).
Depends on your hardware. If you're using SMP, it will help; the scheduler is a bit better. Also has better support for the P4/Xeon's Hyperthreading. Overall lower latency operation as well. udev is a nice upgrade from devfs. No need to use proprietary sound or ethernet drivers on an nForce platform. If you're dealing with LOTS of traffic, it will perform better under stress.
What hardware do you have?
No one is forcing any of these programmers to write code for free. No one is forcing me to GPL the stuff I write outside of work. If someone makes money off my code, more power to them. I know for a fact that some Linux-based smart-phones that sell for hundreds of dollars in Asia contain some code I've written. I think that's great.
And IBM, HP, Red Hat, SuSE, and Trolltech, among others, DO hire open source developers. IBM has donated millions to the community. These companies do give back to us. Maybe they don't give every damn contributer a paycheck, but they donated RCU, JFS, and a number of other things.
And you know what, I don't have a problem not getting paid for service trips to third world countries, or the 20+ hours a week I spend doing youth ministry.
No one is forcing programmers to contribute to the open source community. We do so to scratch an itch, to learn, to give back to a community that has given us so much, and because some of us think it's the right thing to do. If Longhorn is costing Microsoft that much, they're doing something wrong. Imagine if they opened their code, and had a community of millions helping them with code audits. Frankly, I want good, reliable software that I can trust. I can't trust proprietary software; I don't know what it's doing. Internet Explorer could, in theory, be capturing my credit card number in web forms and sending them to Bill Gates. Although I suspect that this is not true, there's no way for me to find out, because running a fucking network sniffer and trying to crack any encryption used is likely against the EULA.
I was responding to commentary about Gates vs. Stallman. So I'd say Stallman's definition is fairly relevant.
Sorry, I disagree. Red Hat sells RHEL for a lot of money, and lots of companies are willing to pay. I don't follow Stallman -- I think he's too extreme, but I have a strong preference to free software. Why? Because I can use it any way I want. I can modify it. Yeah, end users don't give a damn about that, but it's nice to know that if they change their mind, they can. Also, if they want to hire me to modify free software for their needs, they can. I've done so. I can't do that with proprietary software.
Do I think all software should be free? I'd like so, but if people don't want to join the free software club, it's their choice. I largely don't care. That doesn't mean I have to use it, and outside of work (I have to use Flash and Windows for lots of stuff), I don't.
Gates is not fanatical about all software being 'non-free'. Proprietary yes. Microsoft produces plenty of software that runs on Windows and OSX that's (surprise, surprise) actually free.
Wrong definition of "free". When Stallman talks of "free", it has NOTHING to do with price. It's "free" as in "free speech" rather than "free beer".
Seriously, external serial modems are ALWAYS Linux compatible, easier to deal with, and usually of high quality.
Script-fu -- incredibly useful for automating content generation. Very clean, easy, and powerful.
I'm not a pro, but I use GIMP because I find it simpler and less daunting than Photoshop, and still almost as powerful.
My guess is that the initial rev will, in fact, be slower on 64-bit. Microsofts compilers are new to 64-bitness, and a reasonable amount of memory bandwidth will be wasted on larger ints. On the other hand, in 64-bit mode on the Opteron, there are twice as many GPR's, so it could wind up being faster. My bets are still on slower largely due to immaturity of the platform.
Belkin makes a model, but it ain't cheap. MSRP is $325, and it comes with no cables. The cables are $80 each. So if you want to connect 4 PC's, that's a whopping $645.
Yes, you are supposed to use a type of active splitter/hub type device to split SATA into multiple channels.
Personally, I have no need for external drives. SATA fixes clutter now inside my case, so I like it. Plus, there are affordable 10,000RPM SATA drives available, and I like that.
Saying "everybody" hates GIMP's interface is a rather foolish statement. Photoshop on the Mac doesn't use MDI and Photoshop on Windows does. It does for legacy reasons.
And when people talk about MDI, they mean "windows within windows", not "an interface that allows you to view multiple documents in one window". Under the current accepted definition of MDI, tabbed browsing doesn't fit the bill. Tabbed browsing makes sense for a couple of reasons:
- you're talking about viewing different clumps of similar types of data, rather than editing
- the tabs are always visible, unless you have way, way too many open
- The metaphor is actually a bookmark metaphor
Tabbed browsing, in many ways, is inefficient as well. Honestly, the best approach to this I've seen was BeOS, where one could stack multiple windows on top of each other, and then treat the titlebars as tabs. This way, normal switching between windows hotkeys still work, rather than having to memorize a new set. They behaved both like windows AND tabs.
As for MS-Windows-style MDI, with current toolkits, there is no point. There are better, easier ways of doing it (like my suggestion in another post -- have all tool palletes raised when the document gains focus). And also, in the GIMP 2.0 betas, the UI is notably improved, particularly with the new dockable tool palletes. I can still make room for my terminal view and my document, but have the palletes dock at the side of the screen.
And I do accept Photoshop does fine its way. I just think, IMHO, that GIMP happens to be better in this respect.
SATA has a lot of advantages... the cables can be made cheaply, the standard form-factors of drives are designed to slide in/out of a backplane (which no one has made yet) and it has room for growth. The next gen version will support 300MB/s and splitting a channel into 15 other channels. There's also talk of an external version of the standard. Plus, it's largely backward compatible with ATA.
As for SAS (Serial attached SCSI), it has some nice features as well. First and foremost, it's backward compatible with SCSI and compatible with SATA. That means you can take an SATA drive and plug it into a SAS controller just fine. The inverse is not true; SAS drives won't work on SATA. This means if someone wants to make a common SAS backplane for a PC case, one can just slide in any SAS or SATA drive in any combination. Very cool, as far as I'm concerned.
The throughput would increase with striping, but not latency. Still, a single 7200RPM full-size drive would be faster than two striped 5400RPM 2.5" drives, and far less expensive.
As far as reliability, yes, actually being in a laptop adds to that. But a hard drive has moving parts, and making parts larger generally means they'll wear better. In laptop drives they use smaller bearings, smaller read-heads, thinner spindle, and thinner platter. Something is going to give over time, and over significantly less time than on 3.5" drives. And frankly, given the price and capacity difference, I'd much rather see one or two slots for 3.5" drives than a dozen slots for laptop drives. 3.5" drives currently top out at 300GB @7200RPM or 74GB @10,000RPM. 2.5" drives top out at 80GB @5400RPM and 60GB @7200RPM. There are still a number of laptop drives on the market running at 4000RPM.