Um, no. I've had a Dell since 1998, and it's in perfect condition. And Dell's aren't built without any R&D. There aren't as many PCs (including the newest G4 Towers) that have cooling solutions as good as a Dimension's.
No, lmbench doesn't measure the fixed system call overhead, but system call performance. Meaning it takes twice as long to do an mmap() because slower code is servicing that call. It takes twice as long to do a context switch, twice as long to open a file, IPC is half as fast, etc. And contrary to what you say, most applications (particularly I/O bound applications) spend a great deal of their time inside system calls.
Um, I don't think Apple even *makes* a machine similar to a Dell. Dell's mid-range machines start with a 3GHz P4. Apple doesn't even have a processor that fast. Just for kicks, I tried to configure a $2500 machine, which I consider mid-range to high-end.
$2506 bought me a (single) 1GHz G4, 256MB of RAM, 60GB disk, GeForce 4MX, 17" LCD, and Apple "Pro" speakers.
$2558 (admittedly after rebate, but I bought an Inspiron last year and my rebate came quickly) bought me a 3 GHz P4, 512MB of RAM, 120GB disk, Radeon 9800 Pro, 18.1" LCD, and Altec Lansing 4.1 speakers.
Both had 3 year warrenties, though Dell's came standard and Apple's cost $200.
These Dell's are great machines. Pretty much everyone in our dorm has one, and I've owned one since 1998. They are whisper quiet. I have one running 5 feet behind me and I just had to turn around to check if the power light was on. They're easy to open, easy to work inside of, and the power supplies will last forever (the one in my 1998 Dimension has been going for years in dusty basements).
Hmm. I've never heard this before. The slowness of OS X has zippo to do with the configuration mechanism and everything to do with Mach and Quartz. Quartz is largely done in software (even in QE, where OpenGL is only used for the final compositing step) and Mach is just plain slow. In lmbench numbers (measure of the speed of basic kernel primitives like IPC, mmap(), etc) OS X 10.1 ws shown to be about half as fast as NetBSD on the same machine. It's probably improved since then, but even 10.3 is (optimistically) probably 25% slower than NetBSD on lmbench.
Sorry to break it to you, but there isn't some magical hardware in the average Apple comp that you can't get in a Dell or IBM. They use pretty much the same components as anyone else. In many cases the components are crappier (Apple "Pro" speakers my ass, give me a pair of Klipsches any day!) If you take a look at the cost of the machine, you'll see it goes to three main places - CPU - Motherboard - Apple markup The last one is a little high, but okay for the level of case engineering, QA, and tech support Apple supposedly has. The first two (much larger) ones are an unavoidable effect of having a non-standard hardware platform.
Wow. This reminds me of the GCC FUD that is sometimes spread. You used to occasionally hear "GCC requires the output to be GPL" or something like that. That was preposterous, and this seems just as stupid.
There is a difference between "should" and "does." In theory, we'd all check for "malloc()" returning zero (or "operator new" throwing an exception) but we don't. The VM system allows programs to pretty much ignore memory limits, and like it or not, a lot of programs (maybe most programs) don't rigorously check to see if allocations actually succeed.
It would be difficult to use RAM to cache disk writes, since dirty areas would have to be flushed when a process wanted more memory. >>>>>>>>>> Um, it already does this. Linux will merrily cache pretty much everything in the filesystem until something actually needs the memory. If I'm sitting there just downloading something in the background, and reading something on the net, a big several hundred meg download will lead to very little disk activity, with the occasional write being done to update the filesystem journal as the filesize increases. Then, if I start up my compiler, the system's memory demand will suddenly increase, and I'll hear the drive writing out that entire file in one go.
Shit! This has got to be the most humiliating thing I've ever posted! Not only was my original post apparently a dupe, but it was caused by my inability to spell "perpetual" properly. I thought this whole "online community" thing was supposed to get rid of the "humiliation in front of strangers" thing but nooooo!
I haven't even read the thread yet. But I assume there are the obligatory vibrator jokes. I've done a search, and this one doesn't appear to be in the database.
Cool use for this technology: prepetual motion vibrator!
Actually, if you hadn't been listening to the hype, Doom 3 is supposed to be a big departure from the Quake series, in that it focuses on the single-player experience. All the emphasis on the moody lighting and all that is to re-create the genuinely creepy feeling of the first Doom games.
The problem with research like this is that solid evidence -- the fact that women have weaker spatial skills -- leads to random speculation -- that this weakness is somehow inherent/evolutionary. To tell the truth, we are still suffering from a society that has been dominated for thousands of years by men. A lot of western tradition, particularly in Britain and the US, makes the situation even worse than it is elsewhere (relatively, when you take into account that non-western countries are far less industrialized). To this day, I see huge social disparities everywhere I look. It's better in some places, worse in others, but nowhere is it very good. Now, it very well may be that these social factors cannot explain everything and that girls are inherently more limited in certain skills, and more proficient in others, than guys. However, until we analyze exactly what effects society still has on females (or better yet remove those social factors entirely) we cannot reliably speculate on the origin of perceived differences.
PS> The words "male" and "female" are traditionally not applied to human beings. It would be like saying "two people mated" rather than "they had sex."
I do. There is a significant quality loss in the translation from MP3 to ATRAC (the only thing the NetMD players play internally), and the OpenMG software absolutely sucks rocks. The kicker is that it doesn't work in Linux!
You think that's bad. At Georgia Tech, the morons used your SSN as your student ID! The damn thing was on every test I took for an entire semester. The minute I saw the crazy scheme I realized it was a major security risk. And lo and behold, a few months later, they wise up and decide to switch to a randomly generated 9 digit number.
Ha ha! See, someone needs Postscript versions for "serious printing"! This is a reply to that loser who said that Postscript fonts were dead, and TrueType had "won."
What sucks about X is not poor rendering (current mechanisms can give extremely high quality output, easily on a par with OS X for everything, and Windows for everything except maybe italics) but spotty availability. Not all distributions have it, and not all applications support it. However, if you run a straight Gnome-2 or KDE 3.x setup on a modern distro (Gentoo, Mandrake 9, RedHat 9, SuSE 8.2) and have a decent set of fonts (like Vera, the MS Core Fonts, or some commercial Adobe or Monotype fons) then your font quality should be top-notch.
The other explanation that was posted was a bit inaccurate. I've used FreeType to render in one of my own projects, so I think I can explain things a little better
Fontconfig -- X-independent library that handles font management. This includes finding font files on the hard drive, mapping them to Xft font names, and matching non-existant font requests to similar available fonts. Xft2 -- Xft2 is the actual client API. It's primary job is to handle rendering requests. It auto-detects if XRender is available and uses that to render, or else it falls back to the core X protocol. Xft2 also abstracts many of FreeType's services, such as access to font metrics information. FreeType -- This handles the actual rasterization. TrueType/Postscript fonts are stored as vector graphics inside font files. FreeType itself doesn't do any font management (you can't give it a font name for example) or anything like that. The FreeType API consists of functions to open a given font file, and functions to rasterize a given letter (glyph) to a user-provided memory buffer. It also has an API to cache glyphs, so if you rasterize the same letters at the same sizes over and over, you can just use the bitmaps without rasterizing them a gain.
So the basic thing is that FontConfig actually finds the font files for a given name ("Times New Roman -- 12 pt -- Bold,") Xft2 gets the request to render a string of text with that font, it uses FreeType2 to render each individual character to a buffer, and then uses XRender and it's own algorithms align those bitmaps properly and draw strings of them onto the window.
why the US would spend $100+ billion to control Iraqi oil revenues that are a twentieth of that annually... surely one could get a higher return elsewhere? >>>>>>>>> Um, because most of that money comes right back to the US? In the global economy, it's not about how much you spend, but where that money goes. There are a few different catagories of costs to the war:
1) Money that is up in smoke. This covers bombs, spent fuel, destroyed buildings, etc. 2) Money used to pay troops. This is a large part of the deployment cost, and again, it comes right back to the US. 3) Money used to rebuild Iraq. Most of this money comes right back to us. It's the same principle behind how USAID (our foreign AID department) works. 70% of US foreign AID comes right back to US contractors and subcontractors. Guess which companies will get dibs on rebuilding Iraq's infrastructure? Not the Iraqi companies that need the money, but the American companies that don't.
When all these factors are taken into account, the *actual* cost, in terms of money that flows out of the US into other nations, is much, much less.
Now, what are the potential payoffs?
1) US companies getting markets in Iraq. Long term, this is the big one. 2) More secure oil source. It doesn't matter how much oil costs now. It matters how much it costs 50 years from now. Having the world's second largest reserves of oil in a nation indebted to the US is going to look a whole lot more favorable decades from now, when the oil starts to run out and the other OPEC countries get antsy. Even small oil crises (like in the 70's and 80's) can have a huge impact on the economy, and a big one would just be disastrous. 3) Rise in the economy due to increased consumer certainty.
Now, when you take the potential profits into account, the tens of billions of dollars that the war would actually cost seems quite a reasonable bargain.
Of course it would be stupid to say profit is the only motivation. Governments rarely act because of a single motivation. It is however a major one, and I would argue it is *the* major one.
Besides, given the position the US has put itself in, it should expect people to question their motives. If the US really wanted to put itself beyond accusation (and still felt it needed to go to war), it should just put up the money upfront and remove itself from the possibility of profiting from the rebuilding effort. This *would* cost nearly $100 billion in actual currency. Yes, this is asking a lot. No, you can't expect a country to just put aside it's own interests like that. I think the primary problem that most people have is not that the US isn't doing this, but that the US isn't doing this but (through all the talk of liberation, etc) but *acting* like they are doing this.
Ugh. I still refuse to watch most Sci-Fi because of the pervasive leather-clad women. I have nothing against hot women. It's just that it's so cliche'ed and stupid it's not even funny.
But current 2D primitives aren't very useful for a whole lot of things. They do lines, fills and blits. In today's interfaces, that's becoming very limiting. For example, who wants non-antialiased lines anymore? The minute you hit a gradient (and there's a *lot* of those) you're back in software. OpenGL can support most of the SVG imaging model, including filters and gradients, in hardware. The SVGL project has seen up to 100x performance improvements just in preliminary code. The whole point is moving beyond today's relatively poor interfaces to richer more detailed (and resolution independent!) ones. OpenGL is just the enabling mechanism for that.
Quote from the original grand-parent post (brakets mine):
"America provided Iraq assistance back in the 80s against Iran, sure, but to make the leap that they [directly sold] chemical/biological/nuclear weapons is fantasy." Clearly, it's not fantasy, but real and documented fact.
Quote from my parent post, agreeing with the above post:
"I think the worst that you can accuse the US of is condoning Saddam's possession of bio-chem weapons back when the US thought Saddam was on its side against the Islamic radicals." The US didn't just condone Saddam's possession of these weapons, it supplied him with many of them. Maybe not most, but that's not what we're talking about.
Beyond that, I would remind you that we had no business supporting Saddam against Iran in the first place. The revolution in Iran was an internal revolution by the people, mostly the students. Now I don't know how young people would be stupid enough to support a conservative government like that, and I hope the current generation has some sense and tries to change it, but I'm not exactly in a position to second guess their decision.
That's exactly what Dr. Frankenstein said when he created his monster:) Not that I agree with that viewpoint, but it's kinda funny how you fell into the classic trap. Personally, I think we are going to have to put up with some monsters once in a while to advance human civilization.
Slashdot Mirror
Um, there are dozens of dialogues in WinXP that look exactly like that (notably the "MyComputer" property pages).
Um, no. I've had a Dell since 1998, and it's in perfect condition. And Dell's aren't built without any R&D. There aren't as many PCs (including the newest G4 Towers) that have cooling solutions as good as a Dimension's.
No, lmbench doesn't measure the fixed system call overhead, but system call performance. Meaning it takes twice as long to do an mmap() because slower code is servicing that call. It takes twice as long to do a context switch, twice as long to open a file, IPC is half as fast, etc. And contrary to what you say, most applications (particularly I/O bound applications) spend a great deal of their time inside system calls.
Um, I don't think Apple even *makes* a machine similar to a Dell. Dell's mid-range machines start with a 3GHz P4. Apple doesn't even have a processor that fast. Just for kicks, I tried to configure a $2500 machine, which I consider mid-range to high-end.
$2506 bought me a (single) 1GHz G4, 256MB of RAM, 60GB disk, GeForce 4MX, 17" LCD, and Apple "Pro" speakers.
$2558 (admittedly after rebate, but I bought an Inspiron last year and my rebate came quickly) bought me a 3 GHz P4, 512MB of RAM, 120GB disk, Radeon 9800 Pro, 18.1" LCD, and Altec Lansing 4.1 speakers.
Both had 3 year warrenties, though Dell's came standard and Apple's cost $200.
These Dell's are great machines. Pretty much everyone in our dorm has one, and I've owned one since 1998. They are whisper quiet. I have one running 5 feet behind me and I just had to turn around to check if the power light was on. They're easy to open, easy to work inside of, and the power supplies will last forever (the one in my 1998 Dimension has been going for years in dusty basements).
Hmm. I've never heard this before. The slowness of OS X has zippo to do with the configuration mechanism and everything to do with Mach and Quartz. Quartz is largely done in software (even in QE, where OpenGL is only used for the final compositing step) and Mach is just plain slow. In lmbench numbers (measure of the speed of basic kernel primitives like IPC, mmap(), etc) OS X 10.1 ws shown to be about half as fast as NetBSD on the same machine. It's probably improved since then, but even 10.3 is (optimistically) probably 25% slower than NetBSD on lmbench.
Sorry to break it to you, but there isn't some magical hardware in the average Apple comp that you can't get in a Dell or IBM. They use pretty much the same components as anyone else. In many cases the components are crappier (Apple "Pro" speakers my ass, give me a pair of Klipsches any day!) If you take a look at the cost of the machine, you'll see it goes to three main places
- CPU
- Motherboard
- Apple markup
The last one is a little high, but okay for the level of case engineering, QA, and tech support Apple supposedly has. The first two (much larger) ones are an unavoidable effect of having a non-standard hardware platform.
Who do you think makes Apple's computers? And I'm more worried about Apple hardware competing with cheap American Dells than Taiwanese no-name clones.
Wow. This reminds me of the GCC FUD that is sometimes spread. You used to occasionally hear "GCC requires the output to be GPL" or something like that. That was preposterous, and this seems just as stupid.
There is a difference between "should" and "does." In theory, we'd all check for "malloc()" returning zero (or "operator new" throwing an exception) but we don't. The VM system allows programs to pretty much ignore memory limits, and like it or not, a lot of programs (maybe most programs) don't rigorously check to see if allocations actually succeed.
It would be difficult to use RAM to cache disk writes, since dirty areas would have to be flushed when a process wanted more memory.
>>>>>>>>>>
Um, it already does this. Linux will merrily cache pretty much everything in the filesystem until something actually needs the memory. If I'm sitting there just downloading something in the background, and reading something on the net, a big several hundred meg download will lead to very little disk activity, with the occasional write being done to update the filesystem journal as the filesize increases. Then, if I start up my compiler, the system's memory demand will suddenly increase, and I'll hear the drive writing out that entire file in one go.
Shit! This has got to be the most humiliating thing I've ever posted! Not only was my original post apparently a dupe, but it was caused by my inability to spell "perpetual" properly. I thought this whole "online community" thing was supposed to get rid of the "humiliation in front of strangers" thing but nooooo!
I haven't even read the thread yet. But I assume there are the obligatory vibrator jokes. I've done a search, and this one doesn't appear to be in the database.
Cool use for this technology: prepetual motion vibrator!
Actually, if you hadn't been listening to the hype, Doom 3 is supposed to be a big departure from the Quake series, in that it focuses on the single-player experience. All the emphasis on the moody lighting and all that is to re-create the genuinely creepy feeling of the first Doom games.
The problem with research like this is that solid evidence -- the fact that women have weaker spatial skills -- leads to random speculation -- that this weakness is somehow inherent/evolutionary. To tell the truth, we are still suffering from a society that has been dominated for thousands of years by men. A lot of western tradition, particularly in Britain and the US, makes the situation even worse than it is elsewhere (relatively, when you take into account that non-western countries are far less industrialized). To this day, I see huge social disparities everywhere I look. It's better in some places, worse in others, but nowhere is it very good. Now, it very well may be that these social factors cannot explain everything and that girls are inherently more limited in certain skills, and more proficient in others, than guys. However, until we analyze exactly what effects society still has on females (or better yet remove those social factors entirely) we cannot reliably speculate on the origin of perceived differences.
PS> The words "male" and "female" are traditionally not applied to human beings. It would be like saying "two people mated" rather than "they had sex."
I do. There is a significant quality loss in the translation from MP3 to ATRAC (the only thing the NetMD players play internally), and the OpenMG software absolutely sucks rocks. The kicker is that it doesn't work in Linux!
You think that's bad. At Georgia Tech, the morons used your SSN as your student ID! The damn thing was on every test I took for an entire semester. The minute I saw the crazy scheme I realized it was a major security risk. And lo and behold, a few months later, they wise up and decide to switch to a randomly generated 9 digit number.
Ha ha! See, someone needs Postscript versions for "serious printing"! This is a reply to that loser who said that Postscript fonts were dead, and TrueType had "won."
What sucks about X is not poor rendering (current mechanisms can give extremely high quality output, easily on a par with OS X for everything, and Windows for everything except maybe italics) but spotty availability. Not all distributions have it, and not all applications support it. However, if you run a straight Gnome-2 or KDE 3.x setup on a modern distro (Gentoo, Mandrake 9, RedHat 9, SuSE 8.2) and have a decent set of fonts (like Vera, the MS Core Fonts, or some commercial Adobe or Monotype fons) then your font quality should be top-notch.
The other explanation that was posted was a bit inaccurate. I've used FreeType to render in one of my own projects, so I think I can explain things a little better
Fontconfig -- X-independent library that handles font management. This includes finding font files on the hard drive, mapping them to Xft font names, and matching non-existant font requests to similar available fonts.
Xft2 -- Xft2 is the actual client API. It's primary job is to handle rendering requests. It auto-detects if XRender is available and uses that to render, or else it falls back to the core X protocol. Xft2 also abstracts many of FreeType's services, such as access to font metrics information.
FreeType -- This handles the actual rasterization. TrueType/Postscript fonts are stored as vector graphics inside font files. FreeType itself doesn't do any font management (you can't give it a font name for example) or anything like that. The FreeType API consists of functions to open a given font file, and functions to rasterize a given letter (glyph) to a user-provided memory buffer. It also has an API to cache glyphs, so if you rasterize the same letters at the same sizes over and over, you can just use the bitmaps without rasterizing them a gain.
So the basic thing is that FontConfig actually finds the font files for a given name ("Times New Roman -- 12 pt -- Bold,") Xft2 gets the request to render a string of text with that font, it uses FreeType2 to render each individual character to a buffer, and then uses XRender and it's own algorithms align those bitmaps properly and draw strings of them onto the window.
why the US would spend $100+ billion to control Iraqi oil revenues that are a twentieth of that annually... surely one could get a higher return elsewhere?
>>>>>>>>>
Um, because most of that money comes right back to the US? In the global economy, it's not about how much you spend, but where that money goes. There are a few different catagories of costs to the war:
1) Money that is up in smoke. This covers bombs, spent fuel, destroyed buildings, etc.
2) Money used to pay troops. This is a large part of the deployment cost, and again, it comes right back to the US.
3) Money used to rebuild Iraq. Most of this money comes right back to us. It's the same principle behind how USAID (our foreign AID department) works. 70% of US foreign AID comes right back to US contractors and subcontractors. Guess which companies will get dibs on rebuilding Iraq's infrastructure? Not the Iraqi companies that need the money, but the American companies that don't.
When all these factors are taken into account, the *actual* cost, in terms of money that flows out of the US into other nations, is much, much less.
Now, what are the potential payoffs?
1) US companies getting markets in Iraq. Long term, this is the big one.
2) More secure oil source. It doesn't matter how much oil costs now. It matters how much it costs 50 years from now. Having the world's second largest reserves of oil in a nation indebted to the US is going to look a whole lot more favorable decades from now, when the oil starts to run out and the other OPEC countries get antsy. Even small oil crises (like in the 70's and 80's) can have a huge impact on the economy, and a big one would just be disastrous.
3) Rise in the economy due to increased consumer certainty.
Now, when you take the potential profits into account, the tens of billions of dollars that the war would actually cost seems quite a reasonable bargain.
Of course it would be stupid to say profit is the only motivation. Governments rarely act because of a single motivation. It is however a major one, and I would argue it is *the* major one.
Besides, given the position the US has put itself in, it should expect people to question their motives. If the US really wanted to put itself beyond accusation (and still felt it needed to go to war), it should just put up the money upfront and remove itself from the possibility of profiting from the rebuilding effort. This *would* cost nearly $100 billion in actual currency. Yes, this is asking a lot. No, you can't expect a country to just put aside it's own interests like that. I think the primary problem that most people have is not that the US isn't doing this, but that the US isn't doing this but (through all the talk of liberation, etc) but *acting* like they are doing this.
Ugh. I still refuse to watch most Sci-Fi because of the pervasive leather-clad women. I have nothing against hot women. It's just that it's so cliche'ed and stupid it's not even funny.
But current 2D primitives aren't very useful for a whole lot of things. They do lines, fills and blits. In today's interfaces, that's becoming very limiting. For example, who wants non-antialiased lines anymore? The minute you hit a gradient (and there's a *lot* of those) you're back in software. OpenGL can support most of the SVG imaging model, including filters and gradients, in hardware. The SVGL project has seen up to 100x performance improvements just in preliminary code. The whole point is moving beyond today's relatively poor interfaces to richer more detailed (and resolution independent!) ones. OpenGL is just the enabling mechanism for that.
God willing.
Quote from the original grand-parent post (brakets mine):
"America provided Iraq assistance back in the 80s against Iran, sure, but to make the leap that they [directly sold] chemical/biological/nuclear weapons is fantasy."
Clearly, it's not fantasy, but real and documented fact.
Quote from my parent post, agreeing with the above post:
"I think the worst that you can accuse the US of is condoning Saddam's possession of bio-chem weapons back when the US thought Saddam was on its side against the Islamic radicals."
The US didn't just condone Saddam's possession of these weapons, it supplied him with many of them. Maybe not most, but that's not what we're talking about.
Beyond that, I would remind you that we had no business supporting Saddam against Iran in the first place. The revolution in Iran was an internal revolution by the people, mostly the students. Now I don't know how young people would be stupid enough to support a conservative government like that, and I hope the current generation has some sense and tries to change it, but I'm not exactly in a position to second guess their decision.
That's exactly what Dr. Frankenstein said when he created his monster:) Not that I agree with that viewpoint, but it's kinda funny how you fell into the classic trap. Personally, I think we are going to have to put up with some monsters once in a while to advance human civilization.