Apple advertises the ability to view photos as one of the features of iPod nanos. Therefore, they are responsible for providing that feature. If the product, used as intended, is damaged in such a way that that feature becomes unusable, then Apple is responsible.
The basis of the lawsuit, which they will have to prove, is that benign usage causes damage. If Apple has made a product that is damaged such that a core function is impaired when used as designed, then the lawsuit is perfectly reasonable because doing that is illegal.
Also, I don't know if anyone else has noticed but there are two camps: those that claim their nanos don't have any problems, and those that say their screens scratch easily. Given that there are large numbers of both, I think the only reasonable conclusion that some percentage of nanos have a flaw that exposes them to damage. If the flaw is as bad as it seems, there would be virtually no unaffected users if it were uniform.
There are two issues. In the first, which Apple has already addressed, the LCD cracks very easily.
In the second, which is unrelated, the casing of the iPod including the screen scratch extremely easily. Apple denies this is a problem in the face of mounting evidence to the contrary.
"You're right. Intel, for example, would never hardwire the clock speed into a chip so you can't overclock it. HP would never add components to a printer to decrease its performance. You'll never find a motherboard with a NIC but no RJ45, Firewire support and no Firewire socket, but with traces for both leading to empty vias."
Hey, there's plenty of manufacturors for all those parts.
"Competition is good, I agree. It's a shame that the rules of the intellectual property game have been written so as to make an OS monopoly more or less inevitable, but at least you've got hardware choice if you choose the monopoly OS."
For the moment at least MacOS and Windows aren't the only choices, and even if software patents make the open source OSes unviable by themselves there's still the commercial distributions like Suse.
a) most OEMs won't go out of their way to reduce the capabilities of their computers b) even if they do there's lots of alternatives c) even if they all do I can build my own
My problem with Apple, over and above anything specific, is that machines are configured by policy, not by technological or economical constraints. They'll omit obvious features or even disable ones they get for free to implement this policy.
Even if certain classes of machine are attractive at some points in time, Apple policy prevents the platform from being a viable home.
"Second, the iPod still has Firewire support through the special cradle, IIRC."
Apple specs page - I don't see "firewire" or "1394" anywhere on that page. Arstechnica review - "Upon connecting the new iPod via Firewire, you'll see the same screen you get when you try to connect the nano via Firewire" said screen
It would appear you recall incorrectly.
"Also, most Macs since 1999 have shipped with USB, and I'm fairly certain that the two-year-old Macs have USB 2.0"
G4 iMacs and G3 iBooks did not. PowerMacs and PowerBooks have it... but artificially constraining the cheaper machines is one of my criticisms, and leaving USB 2.0 out of the cheaper machines is an example of that.
"If you want a tower, you're likely going after either a cheap box (covered by the Mini) or a powerhouse (covered by the G5 towers)."
My goal, along with the majority of tower owners, is a machine that isn't necessarily outrageously powerful but one that can be updated, and has better capabilities. Wanting dual displays or PCI slots does not imply wanting a second CPU or having a few thousand extra dollars lying around.
Also, an entry-level tower from Apple, if they decided to make one, would "just work" just fine. PowerMacs can do it, you only worry about the slots when you need to use them.
"Fourth, the CPU isn't that much slower than most modern chips"
Depends on how you define "slow". If you think 2-4 times slower (integer code) isn't that much slower... Well, then you have pretty low standards. I've used the faster G4s on integer-heavy stuff like web pages (particularly things like javascript) and it's pretty sad.
"I honestly love the display on my 12" PB (same as the one on the iB), having come from a world of much higher-res PC laptops"
If 1024x768 on 12" iBooks is okay, then 1280x1024 should be okay on a 14".
"It's not quite so simple as that. With balanced reads and writes (the common case), the 2.3 and the 2.5 can saturate the DDR2 bus. So for balanced code, you're looking at a non-trivial 30% bandwidth increase."
In my experience, code that does balanced reads and writes tends to be more latency sensitive than anything else.
"The previous G5 chipset had pretty bad latency compared to other chipsets, so improvements in the chipset likely offsets some of the latency increase."
Even if true, the bandwidth limitations prevent using DDR2 memory from being a good tradeoff. Similar optimizations with dual-DDR would have had lower latency.
"The cache is the biggie --- for workstation apps, its going to help quite a bit."
True... but if they can get a dual-core chip with 2 mb of cache, why couldn't they get a single-core chip with 1-2 mb of cache? Cache is forgiving of defects so yields and prices stay good. These aren't just up against the old competition, and if it turns out that they beat the old competition that's not that impressive. They're up against some very inexpensive PCs, and dual-core PCs that will leave these in the dust are available for a little more than half as much.
"Also, I wouldn't be surprised if the new machines don't cost Apple appreciably less than the old ones."
I would be shocked if that were the case, given the lower complexity of a system with only one CPU.
A G4 has a 167 mhz bus. DDR-333 memory was already too fast by a factor of 2. Add to that the fact that DDR2 has higher latency and you get a performance downgrade.
"Hello, the iBook, iMac, and Mac Mini easily fit within your price range, and all of them are very capable systems."
Depends on how you define "capable". The CPU in minis and iBooks is the slowest you can get in that class of machine from anyone, a 4200 rpm laptop drive is the slowest you can get, the resolution on iBooks sucks (particularly for the 14"). I think a VIA C3 processor might be slower in some areas, but the machines they come in are generally cheaper than even a mini.
As for the iMacs... they look pretty reasonable, but the lack of an entry-level tower is a problem. Because Apple disabled the dual-display feature of the video chipset and they lack PCI slots (or even laptop card slots) they are significantly less capable than entry-level towers. The lack of PCI is deffinitely an issue. Apple is willing to screw G3 iBook and G4 iMac users with machines less than 2 years old by dropping firewire support from iPods, and there's no reason to believe they won't be willing to do the same in 2 years when something else comes out.
An iMac might be okay for some users, but for enthusiasts or profesionals the lack of anything between an overpriced PowerMac and an less capable iMac is a problem.
"The new dual core will perform better, given the larger caches, on-chip cache-to-cache bus, and faster memory. If it was an Athlon 64, it could have performed worse, because you'd be going from dual independent memory controllers to a shared memory controller, but the G5's have a shared memory bus anyway, even when there are two seperate physical processors."
The G5's bus is actually 2x 32-bit unidirectional busses, so the single-dual-core chip can't saturate the DDR2 memory bus while a dual-single-core chips could have because they have independant busses to the chipset. Also, DDR2-533 memory is higher latency. You take a bandwidth hit, a latency hit, it costs Apple a lot less to make the machine and they're charging the same amount.
A G5's bus is 2x 32-bit unidirectional so they can't saturate the DDR2 bus, while two processors each with their own bus could have, so you take the latency hit without a bandwidth advantage.
Dual-core gives you most of the performance for a fraction of the price. Therefore, the performance hit relative to dual-single-core is usually worth it. However, Apple hasn't made the machines any cheaper. I'm sure their margins are much bigger, but it's not any better for us.
The only machine that's worth it is the dual-dual-core, but for anyone who doesn't need/doesn't want to pay for that, it's a really sad update.
"Read the comments in the code. It says if protothreads call other C routines and when in that context the called routine blocks then things don't work right."
Depends on the goal.
One of my favorite features of Python is generators. These can be used to implement generators in C. I don't care if it blocks because my goal is not concurrency.
"Sound travels approximately 30cm in one millisecond. Which means that even moving your head slightly will introduce significant delays. Unless you listen to your music sitting very rigidly with your head positioned _exactly_ between the speakers, I don't think you need to worry about millisecond delays."
I wouldn't worry about it, but I know people that do.
I think it has more to do with correctly positioning the nodes and anti-nodes of the interference pattern.
"Also, IP isn't my favored priority stream transport. I'd recommend a separate network for sound and I'd be weary of any delays incorporated in the IP transport. Think ping times! Also, encoding with the ADC does not include encapsulation into an IP packet, which can lead to worse lip-sync problems. Even 20ms delay makes me crazy (~1 frame). Of course, if its digital all-the-way, things can look brighter."
Worse than that, you're relying on the overall system to be synchronized. The enthusiasts that would care about quality enough to use this are the same ones that set up the positions of the speakers based on the wavelengths of the sound that they're listening to.
You'd have to have the speakers time-synchronized to substantially less than a millisecond and buffer the sound.
That's bee "10 years away" for as long as I can remember. It'll probably replace optical drives one of these days, but optical drives have certain disadvantages that you can't really do anything about. Primary storage will probably always be something else.
"It may be a quip, but at least in my experience, most people I know personally who use Linux use it out of hatred for Microsoft, and those that use BSD use it because they love Unix."
I don't care about Microsoft, and I use Linux in addition to BSD. What does that say about me?
"Just pointing out that unless a performance difference is glaring and sudden people tend not to mind or even notice. And I've seen crap Unix software and drivers too that users accepted as normal."
That depends on what you're doing. If it were an across-the-board 10% I probably wouldn't care, but it's not.
Linux 2.6 kernels are absoloutely outstanding at responsiveness, the only comparable experience I've had was with BeOS. It's pretty easy to tell whether or not (for example) your music skips under high load. Frankly, I've gotten used to being able to throw almost anything at my UP Linux machine without GUI or music being impacted perceptibly. Provided I'm not swapping, I haven't seen a slowdown under any load.
Nothing else out there at the moment seems to be able to do it. Windows can't do it, Mac can't do it, FreeBSD can't do it (even with ULE...), and OpenBSD certainly can't do it (my SSH sessions to my OpenBSD box get laggy when it's under high load).
Slashdot Mirror
Apple advertises the ability to view photos as one of the features of iPod nanos. Therefore, they are responsible for providing that feature. If the product, used as intended, is damaged in such a way that that feature becomes unusable, then Apple is responsible.
The basis of the lawsuit, which they will have to prove, is that benign usage causes damage. If Apple has made a product that is damaged such that a core function is impaired when used as designed, then the lawsuit is perfectly reasonable because doing that is illegal.
Also, I don't know if anyone else has noticed but there are two camps: those that claim their nanos don't have any problems, and those that say their screens scratch easily. Given that there are large numbers of both, I think the only reasonable conclusion that some percentage of nanos have a flaw that exposes them to damage. If the flaw is as bad as it seems, there would be virtually no unaffected users if it were uniform.
There are two issues. In the first, which Apple has already addressed, the LCD cracks very easily.
In the second, which is unrelated, the casing of the iPod including the screen scratch extremely easily. Apple denies this is a problem in the face of mounting evidence to the contrary.
Ability to see the screen is a core function (as Apple gave it photo functionality).
In the past, Apple has proven much more willing to take responsibility for deffective products when doing so will make a lawsuit go away.
"You're right. Intel, for example, would never hardwire the clock speed into a chip so you can't overclock it. HP would never add components to a printer to decrease its performance. You'll never find a motherboard with a NIC but no RJ45, Firewire support and no Firewire socket, but with traces for both leading to empty vias."
Hey, there's plenty of manufacturors for all those parts.
"Competition is good, I agree. It's a shame that the rules of the intellectual property game have been written so as to make an OS monopoly more or less inevitable, but at least you've got hardware choice if you choose the monopoly OS."
For the moment at least MacOS and Windows aren't the only choices, and even if software patents make the open source OSes unviable by themselves there's still the commercial distributions like Suse.
a) most OEMs won't go out of their way to reduce the capabilities of their computers
b) even if they do there's lots of alternatives
c) even if they all do I can build my own
Fair enough, they're now slightly more than 2 years old.
My problem with Apple, over and above anything specific, is that machines are configured by policy, not by technological or economical constraints. They'll omit obvious features or even disable ones they get for free to implement this policy.
Even if certain classes of machine are attractive at some points in time, Apple policy prevents the platform from being a viable home.
"Second, the iPod still has Firewire support through the special cradle, IIRC."
Apple specs page - I don't see "firewire" or "1394" anywhere on that page.
Arstechnica review - "Upon connecting the new iPod via Firewire, you'll see the same screen you get when you try to connect the nano via Firewire"
said screen
It would appear you recall incorrectly.
"Also, most Macs since 1999 have shipped with USB, and I'm fairly certain that the two-year-old Macs have USB 2.0"
G4 iMacs and G3 iBooks did not. PowerMacs and PowerBooks have it... but artificially constraining the cheaper machines is one of my criticisms, and leaving USB 2.0 out of the cheaper machines is an example of that.
"If you want a tower, you're likely going after either a cheap box (covered by the Mini) or a powerhouse (covered by the G5 towers)."
My goal, along with the majority of tower owners, is a machine that isn't necessarily outrageously powerful but one that can be updated, and has better capabilities. Wanting dual displays or PCI slots does not imply wanting a second CPU or having a few thousand extra dollars lying around.
Also, an entry-level tower from Apple, if they decided to make one, would "just work" just fine. PowerMacs can do it, you only worry about the slots when you need to use them.
"Fourth, the CPU isn't that much slower than most modern chips"
Depends on how you define "slow". If you think 2-4 times slower (integer code) isn't that much slower... Well, then you have pretty low standards. I've used the faster G4s on integer-heavy stuff like web pages (particularly things like javascript) and it's pretty sad.
"I honestly love the display on my 12" PB (same as the one on the iB), having come from a world of much higher-res PC laptops"
If 1024x768 on 12" iBooks is okay, then 1280x1024 should be okay on a 14".
"It's not quite so simple as that. With balanced reads and writes (the common case), the 2.3 and the 2.5 can saturate the DDR2 bus. So for balanced code, you're looking at a non-trivial 30% bandwidth increase."
In my experience, code that does balanced reads and writes tends to be more latency sensitive than anything else.
"The previous G5 chipset had pretty bad latency compared to other chipsets, so improvements in the chipset likely offsets some of the latency increase."
Even if true, the bandwidth limitations prevent using DDR2 memory from being a good tradeoff. Similar optimizations with dual-DDR would have had lower latency.
"The cache is the biggie --- for workstation apps, its going to help quite a bit."
True... but if they can get a dual-core chip with 2 mb of cache, why couldn't they get a single-core chip with 1-2 mb of cache? Cache is forgiving of defects so yields and prices stay good. These aren't just up against the old competition, and if it turns out that they beat the old competition that's not that impressive. They're up against some very inexpensive PCs, and dual-core PCs that will leave these in the dust are available for a little more than half as much.
"Also, I wouldn't be surprised if the new machines don't cost Apple appreciably less than the old ones."
I would be shocked if that were the case, given the lower complexity of a system with only one CPU.
"Doesn't Freescale have a G4 with a 200 MHz bus?"
They do, but Apple didn't use it.
A G4 has a 167 mhz bus. DDR-333 memory was already too fast by a factor of 2. Add to that the fact that DDR2 has higher latency and you get a performance downgrade.
"Hello, the iBook, iMac, and Mac Mini easily fit within your price range, and all of them are very capable systems."
Depends on how you define "capable". The CPU in minis and iBooks is the slowest you can get in that class of machine from anyone, a 4200 rpm laptop drive is the slowest you can get, the resolution on iBooks sucks (particularly for the 14"). I think a VIA C3 processor might be slower in some areas, but the machines they come in are generally cheaper than even a mini.
As for the iMacs... they look pretty reasonable, but the lack of an entry-level tower is a problem. Because Apple disabled the dual-display feature of the video chipset and they lack PCI slots (or even laptop card slots) they are significantly less capable than entry-level towers. The lack of PCI is deffinitely an issue. Apple is willing to screw G3 iBook and G4 iMac users with machines less than 2 years old by dropping firewire support from iPods, and there's no reason to believe they won't be willing to do the same in 2 years when something else comes out.
An iMac might be okay for some users, but for enthusiasts or profesionals the lack of anything between an overpriced PowerMac and an less capable iMac is a problem.
"The new dual core will perform better, given the larger caches, on-chip cache-to-cache bus, and faster memory. If it was an Athlon 64, it could have performed worse, because you'd be going from dual independent memory controllers to a shared memory controller, but the G5's have a shared memory bus anyway, even when there are two seperate physical processors."
The G5's bus is actually 2x 32-bit unidirectional busses, so the single-dual-core chip can't saturate the DDR2 memory bus while a dual-single-core chips could have because they have independant busses to the chipset. Also, DDR2-533 memory is higher latency. You take a bandwidth hit, a latency hit, it costs Apple a lot less to make the machine and they're charging the same amount.
A G5's bus is 2x 32-bit unidirectional so they can't saturate the DDR2 bus, while two processors each with their own bus could have, so you take the latency hit without a bandwidth advantage.
Dual-core gives you most of the performance for a fraction of the price. Therefore, the performance hit relative to dual-single-core is usually worth it. However, Apple hasn't made the machines any cheaper. I'm sure their margins are much bigger, but it's not any better for us.
The only machine that's worth it is the dual-dual-core, but for anyone who doesn't need/doesn't want to pay for that, it's a really sad update.
"Read the comments in the code. It says if protothreads call other C routines and when in that context the called routine blocks then things don't work right."
Depends on the goal.
One of my favorite features of Python is generators. These can be used to implement generators in C. I don't care if it blocks because my goal is not concurrency.
"Sound travels approximately 30cm in one millisecond. Which means that even moving your head slightly will introduce significant delays. Unless you listen to your music sitting very rigidly with your head positioned _exactly_ between the speakers, I don't think you need to worry about millisecond delays."
I wouldn't worry about it, but I know people that do.
I think it has more to do with correctly positioning the nodes and anti-nodes of the interference pattern.
"Also, IP isn't my favored priority stream transport. I'd recommend a separate network for sound and I'd be weary of any delays incorporated in the IP transport. Think ping times! Also, encoding with the ADC does not include encapsulation into an IP packet, which can lead to worse lip-sync problems. Even 20ms delay makes me crazy (~1 frame). Of course, if its digital all-the-way, things can look brighter."
Worse than that, you're relying on the overall system to be synchronized. The enthusiasts that would care about quality enough to use this are the same ones that set up the positions of the speakers based on the wavelengths of the sound that they're listening to.
You'd have to have the speakers time-synchronized to substantially less than a millisecond and buffer the sound.
That's bee "10 years away" for as long as I can remember. It'll probably replace optical drives one of these days, but optical drives have certain disadvantages that you can't really do anything about. Primary storage will probably always be something else.
Never attribute to malice that which can be adequately explained by stupidity.
"It may be a quip, but at least in my experience, most people I know personally who use Linux use it out of hatred for Microsoft, and those that use BSD use it because they love Unix."
I don't care about Microsoft, and I use Linux in addition to BSD. What does that say about me?
1.4 was unsupported when I gave up on OpenBSD as my dev workstation. That's spiffy, it'll help to have that available on the OpenBSD machine. Thanks.
"Just pointing out that unless a performance difference is glaring and sudden people tend not to mind or even notice. And I've seen crap Unix software and drivers too that users accepted as normal."
That depends on what you're doing. If it were an across-the-board 10% I probably wouldn't care, but it's not.
Linux 2.6 kernels are absoloutely outstanding at responsiveness, the only comparable experience I've had was with BeOS. It's pretty easy to tell whether or not (for example) your music skips under high load. Frankly, I've gotten used to being able to throw almost anything at my UP Linux machine without GUI or music being impacted perceptibly. Provided I'm not swapping, I haven't seen a slowdown under any load.
Nothing else out there at the moment seems to be able to do it. Windows can't do it, Mac can't do it, FreeBSD can't do it (even with ULE...), and OpenBSD certainly can't do it (my SSH sessions to my OpenBSD box get laggy when it's under high load).
"Yet users will take a hit by running a virus scanner without complaint because they're used to the slowdown."
Sure. Windows users.