When OS X 10.3 (Panther) discontinued support for the Beige G3s, the line was six years old and had been discontinued for four. It makes sense that Apple chose to abandon support for a line of machines that wouldn't have run the new OS acceptably anyway.
Linux can get away with supporting ancient hardware because, well, because they don't actually have to support it. Nobody calls up the GNOME foundation complaining that 2.16 crawls on their PIII-450 with 256MB of RAM. In comparison, Apple actually has to live up to the specifications they outline on the box.
Air-quote all you want, but Castlevania is a classic series of games. It first came out on multiple platforms (most popularly the NES) 20 years ago, and since that time has appeared on a number of different Nintendo, Sega, Sony, and arcade platforms.
At the end of 2006, the DS had 36m units installed, the PSP had 25m units. So the DS has the lead, but the PSP's installed base is still pretty big, slightly larger than the XBox's or Gamecube's.
As for seeing one --- I see them all the time on the subway in Atlanta. I have yet to see a DS here, though, outside of the Georgia Tech campus.
The online TV revolution is still a ways away, while HDTV will be pushing major market share (over 1/3) in only the next few years. The number of people who are set up to take advantage of video-download services is minscule. Only a fraction of people have TVs near their internet connection, and virtually nobody is set up with a media-center PC. Not to mention the fact that although broadband penetration is up to 74% now, only a tiny fraction of that is fast enough to reasonably use TV download services. I've got an 8mbit connection, for which I pay Comcast a dear $70/month. For me, TV or movie download is acceptably fast (5-15 minutes), but far from video-on-demand. For most of the internet-using population, who are using connections that are a couple of megabits at most, the wait times of up to an hour for a single show are unacceptably long. And of course, unless you're using Usenet or Bittorrent, there is almost no content. iTunes' TV selection is feeble (and the quality isn't that great either).
No, over the lifecycle of Blu-Ray anyway, getting an HDTV to watch the "big game" in high-res is going to be much more popular than dealing with the trails and travails of online video.
Despite all the hype surrounding HD-this and High Resolution-that, there hasn't been a major push by consumers to move to the new High Definition televisions.
This is really inaccurate. In the United States, HD market penetration has gone from almost nothing five years ago to pushing 20-20% now. Last year, HDTVs accounted for something like 40% of TV sales, and the majority of sales by revenue. That's "major push" if you ask me.
Don't let the "DVD" in HD-DVD fool you into thinking that it's an open format and Blu-Ray isn't. Both are proprietary but openly-licensable formats. In this regard, they're equivalent to each other, and to DVD and CD, not to mention SD, CompactFlash, MicroSD, etc. Incidentally, neither MemoryStick, SA-CD, MiniDisc, or UMD are Sony-proprietary formats. All of these are openly-licensable formats, just like DVD or CD. The only difference is in how many companies are backing the standard. Interestingly enough, more (and bigger) companies are backing Blu-Ray than HD-DVD. It is different from MemoryStick, MiniDisc, or UMD in this regard.
No sane Mac user buys a.0 release of an Apple OS. Most are wary of being Rev A of any new Apple hardware.
The reason Mac users stay with Apple isn't because Apple never makes mistakes. It's because they try to offer a good user experience, they fix things pretty quickly when something breaks, and they have a reasonable progress schedule, timely but not overly aggressive. Each release of OS X is an incremental step. They have enough new features to get excited about, but aren't jarring breaks from previous releases. And of course, each new release is faster, and after the first maintainence release more stable, than the previous one.
A lot of this has to do with Apple's development methodology. It's very evolutionary, and very incremental. When Leopard comes out, there will have been six major OS X releases in six years (10.0 in Spring '01 to 10.5 in Spring '07). In-between major releases, Apple does a minor release on average about every month or every other month. Each major release of OS X improves the key subsystems in substantial, but easy-to-digest ways. They introduce a few new APIs and deprecate some old ones, all while giving developers a chance to incrementally keep up with changes to the OS.
Let's use Carbon as an example. Carbon descends from the Quicktime for Windows codebase, and in 10.0 looked a lot like the OS 9 toolbox APIs. In 10.0, Apple introduced Quartz, but Carbon still used Quickdraw for most things. Apple also introduced Carbon Event Manager, replacing the old polling-based event API in Carbon. In 10.2, Apple added the HIView framework, based on composited Quartz drawing. In 10.3, Apple moved most of the standard widgets and standard menu handling code in Carbon over to HIView. In 10.4, Apple introduced the resolution independence APIs, and officially deprecated Quickdraw. Now, 10.5 will get rid of the remnants of Quickdraw in Carbon (eg: adding new methods for creating windows that take Quartz rects instead of Quickdraw rects), and fully support resolution independence. So in 10.5, crusty old Carbon is actually a throughly modern widget framework, supporting resolution independence, anti-aliased drawing, composited widgets, and an object-oriented substrate that serves as the basis for higher-level Cocoa APIs. The best part is that for developers who kept up with the little updates over the years, using the new Carbon features in 10.5 will be a piece of cake. By the end of 2007, the majority of actively-developed OS X applications will be utilizing these features. In contrast, Avalon is a complete break from Win32. Porting applications to Avalon is going to take a substantial amount of time, and it'll be years before the majority of common apps are based on the new APIs.
Both Dylan and Lisp are excellent procedural languages. Lisp makes a better imperative language than C. Yeha, the foldr and cdrs are there and whatnot, but you don't have to use them. Just start out using "dolist" (basically a foreach), and progress to "map" or "fold" where it makes sense.
Pre-PC, the market was for kids for the most part.
It was the original PlayStation that did that. Back in 1994, when it came out, the PC gaming market was not nearly big enough to have that kind of impact.
PC's brought 3D to the table.
It was arcades that brought 3D to the table, and even 3Dfx first became successful because of their involvement in making arcade hardware. The Voodoo 1 was the first popular 3D chipset for PCs, and it came out in October 1996, a full year and nine months after the PS1 brought arcade-style 3D gaming into the home. Even the much-delayed Nintendo 64, which was fully competitive with the Voodoo 1 courtesy of its SGI-designed graphics chip, came out several months earlier. And of course the N64 had Mario 64, which was the first game to fully utilize the 3D environment in ways most FPSs still don't allow you to do.
Being able to overclock lower-rated chips and being able to increase the clock on your top line are two completely different things. With conventional cooling, the Core2 isn't seeing much above 3.5 GHz, and its already shipping at 2.93 GHz. It'll go up another 10-20% as the process matures, but Intel isn't holding back in any sense.
This is actually the nature of Intel's P-M and Core designs. Unlike previous designs, they're aimed at a frequency target, then tuned for power consumption. Circuits that could run faster than the given frequency target are redesigned to save power, thus reducing the frequency headroom. So in general you're not going to see huge clock-speed increases on the same process. You saw the same thing with Dothan at 90nm and Yonah at 65nm. Neither gained much more than 130 MHz during their life times.
The TDP and voltage levels are part of the platform specification. Intel can't just up them without requiring motherboards, cooling units, etc, to be upgraded to handle the new spec. They might get away with it for some consumer level stuff, but not in the server market where Clovertown and Barcelona are competing. The server folks are going to want some substantial lead-time to rejigger everything to meet higher TDP and Vcore specs.
All Intel has to do is turn up the clock the day before Barcelona ships. We already know that the Core 2 Duo chips are very overclockable, and getting another 40% -- or even 50%+ out of them -- shouldn't be a problem.
The performance a chip can get with overclocking is way higher than what the manufacturer can deliver in final products. They have to be highly reliable at their specified clockspeed with (relatively) poor cooling, and while meeting the given voltage and thermal dissipation specifications. I've seen the Core 2 over-clock to 3.5 GHz (with conventional cooling) online, but how many of those are doing it at the stock Vcore while staying within the 65 watt TDP?
Yes, there were four cores. Two cores on one chip, two cores on the other. They key words from the titles you quoted are "platform" and "multi-socket", neither of which imply a single die with four cores.
Of course there were two sets of reasoning behind the expectation, and it is the difference between them that has led to this sorry affairs.
And by "we", I meant "them, not me". I could've told you this was coming back when the invasion was first announced. Why? I'm Bengali. I know first-hand that freedom has jack-shit to do with peacefulness, and that democracy is just a system of government, not a solution to all the political and social problems that are present in a country.
This is something that George Bush and many Americans fail to understand. There was a really choice quote in his State of the Union address this week:
"What every terrorist fears most is human freedom -- societies where men and women make their own choices, answer to their own conscience, and live by their hopes instead of their resentments. Free people are not drawn to violent and malignant ideologies -- and most will choose a better way when they are given a chance."
Bullshit. Studies have actually shown that democracies are no less warlike than non-democracies, they just don't fight each other. Iran is the most democratic government in the middle east, save for Israel. Its government was put into place by a popular revolution, and its infrastructure for elections is very established (read up on the political situation in Iran, and the backlash against Ahmadinejad's party in various elections). Of all the Islamic countries in that region, it's the one that looks most likely to look like a free democracy in the next few generations, without our intervention. However, that doesn't stop them from seeing the United States as a bitter enemy! And of course we can't forget Great Britain, the oldest existing free society, and their history of colonial rule and historical ideas of racial and cultural superiority.
Ultimately, it's not democracy that makes Westerners, well, Westernized. It's a couple of thousand years of shared culture, history, and civilization. Democracy is an expression of the underlying mindset of the West, not the underpinnings of that mindset. It took the West a long time to get its societies to the point where they could support democracy. It took France a 150 years to build a supportable Republic, and France had one of the longest traditions of liberal Enlightenment thinking in Europe! How could anybody be stupid enough to think that we could have pushed Iraq to do it in a few years, much less a few decades?
Seriously. How often do you encounter people in science who oppose the government being in science? It's exceedingly rare, for the simple reason that scientists realize that science is expensive and risky, and private industry often can't stomach it. There are even economic theories that show why government spending in research is necessary, based on the concept of "public goods", which benefit everyone but which are hard to get specific payment for.
But my question is, do manufacturers expect consumers to be able to understand all of this mess? What ever happened to plugging a game system into a TV?
When was this ever true? Even the original NES had people dealing with RF-output versus composite, and back then customers also had to deal with TVs that took mono-in, and figuring out that you could just connect the RCA jack for one stereo channel and it would work.
The irony here is that at least with Hussain there, we didn't have to worry about these things. The interests of the United States were better served with a low-level dictator in place than the current unpredictable and uncontrollable situation. Of course, we did not expect the dictator to be replaced by general chaos, but it seems that we did not realize that Hussain was the thing plugging up the dike.
You're missing the point. If the process technology had progressed as expected, the a fast P4 wouldn't have needed huge amounts of power. Look at Power6. It's about 130-watts at 5 GHz. Which is very good power dissipation considering that it has a ton of hardware (really wide busses, huge caches, massive SMP fabric) that Core 2 doesn't.
The point I'm trying to get across is that the P4's design isn't inherently bad, for a desktop/workstation chip. The problem was that it was designed for process technology that turned out to have very different power usage characteristics than were projected.
More than that. The fact that the Sputnik's were so huge relative to the American satellites revealed that their launch capability (read: warhead delivery technology) was far ahead of ours.
It's about performance, not MHz. Let's use SPECint as the metric. SPECint_rate scales almost perfectly with both clockspeed and core count. A P4 gets about 6.5 SPECint_rate/GHz/core, while a Core 2 gets about 11.5 SPECint_rate/GHz/core. So an 8 GHz P4 would get a score of 51.68, while a 3.4 GHz Core 2 would get 78.2.
The P4's single-core results would be substantially higher than the Core 2's single-core results, though. Interestingly, it points to what the P4 was originally designed to do: achieve high performance through high clockspeed. If process technology had met Intel's original projections, we'd have 6+ GHz P4s by now that would have been competitive with current Core 2 chips.
Of course the government printed those extra $3. There's a whole bunch of machinery in the Treasury and the Fed dedicated to keeping the money supply in line with increases in real GDP.
Actually, money is not zero sum at all over any finite time scale. If I invest $100 today, I can count on having about $103 a year from now (US GDP growth rate is 3.2%). Did those $3 comes from somebody else's pocket? No! The entire economy gained 3% worth of value over that time period.
Slashdot Mirror
When OS X 10.3 (Panther) discontinued support for the Beige G3s, the line was six years old and had been discontinued for four. It makes sense that Apple chose to abandon support for a line of machines that wouldn't have run the new OS acceptably anyway.
Linux can get away with supporting ancient hardware because, well, because they don't actually have to support it. Nobody calls up the GNOME foundation complaining that 2.16 crawls on their PIII-450 with 256MB of RAM. In comparison, Apple actually has to live up to the specifications they outline on the box.
Air-quote all you want, but Castlevania is a classic series of games. It first came out on multiple platforms (most popularly the NES) 20 years ago, and since that time has appeared on a number of different Nintendo, Sega, Sony, and arcade platforms.
At the end of 2006, the DS had 36m units installed, the PSP had 25m units. So the DS has the lead, but the PSP's installed base is still pretty big, slightly larger than the XBox's or Gamecube's.
As for seeing one --- I see them all the time on the subway in Atlanta. I have yet to see a DS here, though, outside of the Georgia Tech campus.
The online TV revolution is still a ways away, while HDTV will be pushing major market share (over 1/3) in only the next few years. The number of people who are set up to take advantage of video-download services is minscule. Only a fraction of people have TVs near their internet connection, and virtually nobody is set up with a media-center PC. Not to mention the fact that although broadband penetration is up to 74% now, only a tiny fraction of that is fast enough to reasonably use TV download services. I've got an 8mbit connection, for which I pay Comcast a dear $70/month. For me, TV or movie download is acceptably fast (5-15 minutes), but far from video-on-demand. For most of the internet-using population, who are using connections that are a couple of megabits at most, the wait times of up to an hour for a single show are unacceptably long. And of course, unless you're using Usenet or Bittorrent, there is almost no content. iTunes' TV selection is feeble (and the quality isn't that great either).
No, over the lifecycle of Blu-Ray anyway, getting an HDTV to watch the "big game" in high-res is going to be much more popular than dealing with the trails and travails of online video.
Despite all the hype surrounding HD-this and High Resolution-that, there hasn't been a major push by consumers to move to the new High Definition televisions.
This is really inaccurate. In the United States, HD market penetration has gone from almost nothing five years ago to pushing 20-20% now. Last year, HDTVs accounted for something like 40% of TV sales, and the majority of sales by revenue. That's "major push" if you ask me.
Don't let the "DVD" in HD-DVD fool you into thinking that it's an open format and Blu-Ray isn't. Both are proprietary but openly-licensable formats. In this regard, they're equivalent to each other, and to DVD and CD, not to mention SD, CompactFlash, MicroSD, etc. Incidentally, neither MemoryStick, SA-CD, MiniDisc, or UMD are Sony-proprietary formats. All of these are openly-licensable formats, just like DVD or CD. The only difference is in how many companies are backing the standard. Interestingly enough, more (and bigger) companies are backing Blu-Ray than HD-DVD. It is different from MemoryStick, MiniDisc, or UMD in this regard.
No sane Mac user buys a .0 release of an Apple OS. Most are wary of being Rev A of any new Apple hardware.
The reason Mac users stay with Apple isn't because Apple never makes mistakes. It's because they try to offer a good user experience, they fix things pretty quickly when something breaks, and they have a reasonable progress schedule, timely but not overly aggressive. Each release of OS X is an incremental step. They have enough new features to get excited about, but aren't jarring breaks from previous releases. And of course, each new release is faster, and after the first maintainence release more stable, than the previous one.
A lot of this has to do with Apple's development methodology. It's very evolutionary, and very incremental. When Leopard comes out, there will have been six major OS X releases in six years (10.0 in Spring '01 to 10.5 in Spring '07). In-between major releases, Apple does a minor release on average about every month or every other month. Each major release of OS X improves the key subsystems in substantial, but easy-to-digest ways. They introduce a few new APIs and deprecate some old ones, all while giving developers a chance to incrementally keep up with changes to the OS.
Let's use Carbon as an example. Carbon descends from the Quicktime for Windows codebase, and in 10.0 looked a lot like the OS 9 toolbox APIs. In 10.0, Apple introduced Quartz, but Carbon still used Quickdraw for most things. Apple also introduced Carbon Event Manager, replacing the old polling-based event API in Carbon. In 10.2, Apple added the HIView framework, based on composited Quartz drawing. In 10.3, Apple moved most of the standard widgets and standard menu handling code in Carbon over to HIView. In 10.4, Apple introduced the resolution independence APIs, and officially deprecated Quickdraw. Now, 10.5 will get rid of the remnants of Quickdraw in Carbon (eg: adding new methods for creating windows that take Quartz rects instead of Quickdraw rects), and fully support resolution independence. So in 10.5, crusty old Carbon is actually a throughly modern widget framework, supporting resolution independence, anti-aliased drawing, composited widgets, and an object-oriented substrate that serves as the basis for higher-level Cocoa APIs. The best part is that for developers who kept up with the little updates over the years, using the new Carbon features in 10.5 will be a piece of cake. By the end of 2007, the majority of actively-developed OS X applications will be utilizing these features. In contrast, Avalon is a complete break from Win32. Porting applications to Avalon is going to take a substantial amount of time, and it'll be years before the majority of common apps are based on the new APIs.
Both Dylan and Lisp are excellent procedural languages. Lisp makes a better imperative language than C. Yeha, the foldr and cdrs are there and whatnot, but you don't have to use them. Just start out using "dolist" (basically a foreach), and progress to "map" or "fold" where it makes sense.
Pre-PC, the market was for kids for the most part.
It was the original PlayStation that did that. Back in 1994, when it came out, the PC gaming market was not nearly big enough to have that kind of impact.
PC's brought 3D to the table.
It was arcades that brought 3D to the table, and even 3Dfx first became successful because of their involvement in making arcade hardware. The Voodoo 1 was the first popular 3D chipset for PCs, and it came out in October 1996, a full year and nine months after the PS1 brought arcade-style 3D gaming into the home. Even the much-delayed Nintendo 64, which was fully competitive with the Voodoo 1 courtesy of its SGI-designed graphics chip, came out several months earlier. And of course the N64 had Mario 64, which was the first game to fully utilize the 3D environment in ways most FPSs still don't allow you to do.
WoW is pretty big, but to put it into perspective, it wouldn't even break the top 10 of highest selling console games.
Okay, 3.1 GHz stock Vcore. What's the Core 2 shipping at? 2.93 GHz. Wow, seems right in line, doesn't it?
Being able to overclock lower-rated chips and being able to increase the clock on your top line are two completely different things. With conventional cooling, the Core2 isn't seeing much above 3.5 GHz, and its already shipping at 2.93 GHz. It'll go up another 10-20% as the process matures, but Intel isn't holding back in any sense.
This is actually the nature of Intel's P-M and Core designs. Unlike previous designs, they're aimed at a frequency target, then tuned for power consumption. Circuits that could run faster than the given frequency target are redesigned to save power, thus reducing the frequency headroom. So in general you're not going to see huge clock-speed increases on the same process. You saw the same thing with Dothan at 90nm and Yonah at 65nm. Neither gained much more than 130 MHz during their life times.
The TDP and voltage levels are part of the platform specification. Intel can't just up them without requiring motherboards, cooling units, etc, to be upgraded to handle the new spec. They might get away with it for some consumer level stuff, but not in the server market where Clovertown and Barcelona are competing. The server folks are going to want some substantial lead-time to rejigger everything to meet higher TDP and Vcore specs.
All Intel has to do is turn up the clock the day before Barcelona ships. We already know that the Core 2 Duo chips are very overclockable, and getting another 40% -- or even 50%+ out of them -- shouldn't be a problem.
The performance a chip can get with overclocking is way higher than what the manufacturer can deliver in final products. They have to be highly reliable at their specified clockspeed with (relatively) poor cooling, and while meeting the given voltage and thermal dissipation specifications. I've seen the Core 2 over-clock to 3.5 GHz (with conventional cooling) online, but how many of those are doing it at the stock Vcore while staying within the 65 watt TDP?
Linux is a preemptible kernel.
Yes, there were four cores. Two cores on one chip, two cores on the other. They key words from the titles you quoted are "platform" and "multi-socket", neither of which imply a single die with four cores.
Of course there were two sets of reasoning behind the expectation, and it is the difference between them that has led to this sorry affairs.
And by "we", I meant "them, not me". I could've told you this was coming back when the invasion was first announced. Why? I'm Bengali. I know first-hand that freedom has jack-shit to do with peacefulness, and that democracy is just a system of government, not a solution to all the political and social problems that are present in a country.
This is something that George Bush and many Americans fail to understand. There was a really choice quote in his State of the Union address this week:
"What every terrorist fears most is human freedom -- societies where men and women make their own choices, answer to their own conscience, and live by their hopes instead of their resentments. Free people are not drawn to violent and malignant ideologies -- and most will choose a better way when they are given a chance."
Bullshit. Studies have actually shown that democracies are no less warlike than non-democracies, they just don't fight each other. Iran is the most democratic government in the middle east, save for Israel. Its government was put into place by a popular revolution, and its infrastructure for elections is very established (read up on the political situation in Iran, and the backlash against Ahmadinejad's party in various elections). Of all the Islamic countries in that region, it's the one that looks most likely to look like a free democracy in the next few generations, without our intervention. However, that doesn't stop them from seeing the United States as a bitter enemy! And of course we can't forget Great Britain, the oldest existing free society, and their history of colonial rule and historical ideas of racial and cultural superiority.
Ultimately, it's not democracy that makes Westerners, well, Westernized. It's a couple of thousand years of shared culture, history, and civilization. Democracy is an expression of the underlying mindset of the West, not the underpinnings of that mindset. It took the West a long time to get its societies to the point where they could support democracy. It took France a 150 years to build a supportable Republic, and France had one of the longest traditions of liberal Enlightenment thinking in Europe! How could anybody be stupid enough to think that we could have pushed Iraq to do it in a few years, much less a few decades?
Seriously. How often do you encounter people in science who oppose the government being in science? It's exceedingly rare, for the simple reason that scientists realize that science is expensive and risky, and private industry often can't stomach it. There are even economic theories that show why government spending in research is necessary, based on the concept of "public goods", which benefit everyone but which are hard to get specific payment for.
But my question is, do manufacturers expect consumers to be able to understand all of this mess? What ever happened to plugging a game system into a TV?
When was this ever true? Even the original NES had people dealing with RF-output versus composite, and back then customers also had to deal with TVs that took mono-in, and figuring out that you could just connect the RCA jack for one stereo channel and it would work.
The irony here is that at least with Hussain there, we didn't have to worry about these things. The interests of the United States were better served with a low-level dictator in place than the current unpredictable and uncontrollable situation. Of course, we did not expect the dictator to be replaced by general chaos, but it seems that we did not realize that Hussain was the thing plugging up the dike.
You're missing the point. If the process technology had progressed as expected, the a fast P4 wouldn't have needed huge amounts of power. Look at Power6. It's about 130-watts at 5 GHz. Which is very good power dissipation considering that it has a ton of hardware (really wide busses, huge caches, massive SMP fabric) that Core 2 doesn't.
The point I'm trying to get across is that the P4's design isn't inherently bad, for a desktop/workstation chip. The problem was that it was designed for process technology that turned out to have very different power usage characteristics than were projected.
More than that. The fact that the Sputnik's were so huge relative to the American satellites revealed that their launch capability (read: warhead delivery technology) was far ahead of ours.
It's about performance, not MHz. Let's use SPECint as the metric. SPECint_rate scales almost perfectly with both clockspeed and core count. A P4 gets about 6.5 SPECint_rate/GHz/core, while a Core 2 gets about 11.5 SPECint_rate/GHz/core. So an 8 GHz P4 would get a score of 51.68, while a 3.4 GHz Core 2 would get 78.2.
The P4's single-core results would be substantially higher than the Core 2's single-core results, though. Interestingly, it points to what the P4 was originally designed to do: achieve high performance through high clockspeed. If process technology had met Intel's original projections, we'd have 6+ GHz P4s by now that would have been competitive with current Core 2 chips.
Of course the government printed those extra $3. There's a whole bunch of machinery in the Treasury and the Fed dedicated to keeping the money supply in line with increases in real GDP.
Actually, money is not zero sum at all over any finite time scale. If I invest $100 today, I can count on having about $103 a year from now (US GDP growth rate is 3.2%). Did those $3 comes from somebody else's pocket? No! The entire economy gained 3% worth of value over that time period.