In fact, it probably ranks as top 25% believable quotes on Slashdot. His writing style could use a little work (who am I to judge, I write like crap too), but his technical eval seems legit. It sounds pretty logical to me:
You got a fast device (SSD) on a fast bus (SATA3) connected to a slower bus (PCIe 2.0 1x) connected to the Northbridge. What's the result? Bottleneck at the slowest part (PCIe). Toss in an overhead estimate of 20%, and you get a practical top speed of 400MB/sec.
When switching down to smaller packet sizes like 4K read/writes, overhead increases since you'll need to send more bytes down the bus in order to mark the smaller packets. (big packets == less overhead in a well designed system, see GigE Jumbo Frames)
A 20% estimate isn't unrealistic either. USB2.0 might be 480 mb/sec (60MB/sec) but it's got a 30% typical overhead penalty giving you a top speed around 40MB/sec.
AMD and Intel can make the SATA controller closer to the CPU by placing it in the Northbridge. This removes both the PCIe 1X bandwidth problem as well as gives them the opportunity to lower all overhead at the same time by abstracting out the PCIe bus if they choose to. Obvious advantage.
I'm kinda amazed there are slashdotters who don't recognize basic computer architecture problems like this. Like, for anybody wondering why there's no USB3 controllers on Apple's Macs, it isn't because they're trying to rip you off and sell it to you later. It's because, in this case, their engineers know better. Wait for Northbridge integration, anything less is subpar USB3.
Apple dongle doesn't support writing from an app as far as I know. On the other hand, you could toss videos onto it and import them into the camera roll to play if I wanted, I guess.
Pretty much everybody I know of who bought an Android phone did it because they didn't want to switch to AT&T at the time. (back then, the verizon iphone was just a legend)
Heck, I was given a free original iPhone, tried AT&T for 20 days, said "fuck this dropped call shit", and got myself a WinMo 6 phone on Verizon.
If carriers were that equal, yeah, then switching would be fine. But if you get more dropped calls in 1 week on AT&T than you did in 2 years of Verizon, switching is a big deal.
Well written and much more diplomatic than what I would have said:
"Haha, suck it, Mr. We-don't-need-no-stable-kernel-ABI! Not so much fun being on the other end of the 'can't we program to a uniform standard' problem, is it, Linus?"
Google is providing incentives to handset manufacturers to get clearance with Google with regards to what they're allowed to do in order to install some Google apps like the Market and Navigation. So it's not entirely a "do whatever with it for free" situation.
Let's get some quick info about a WD Velociraptor. (http://www.anandtech.com/show/2508) 98.4MB/sec avg sequential. 4.6ms seek time.
During 4.6ms, you can do 463.5KB of transfer. To do a 4K transfer, it's 0.04ms.
Now consider some sort of file lookup. What do you need to do? You'll need to seek to the part of the disk which has your file system's tables, which probably points to a few tree nodes. This is equivalent to walking a 4 node linked list, I guess. Each of these would be a seek, and a 4KB read. At this rate, you're spending 18.4ms moving the heads, and 0.16ms actually reading. Afterwards, you can then chug along at your blazing 98.4MB/sec reading the file.... which might be just 128KB. (1.27ms)
Out of a total of 19.83ms, 1.27ms is actually spent doing your file transfer.
Going with a crappy SSD. Say 0.03ms seek time, and 98.4MS/sec sequential (so I don't have to keep typing in numbers into a calculator). This means we're doing a 0.12ms for a seek, 0.16ms traversing the filesystem, and then 1.27ms reading the file.
Out of a total of 1.55ms... well, we're already 12.7 times faster than the hard drive, so uh, yeah.
So you're probably thinking, okay, but hey, this is for a random 128KB read. And we're trying to copy sequentially to another image, right?
Well, if you're doing multiple reads of the same template to multiple drives, and you just so happen to be doing these out of sync, then copying 128KB to each target drive will require a read from different locations, yeah? Which means every chunk (depending on how you schedule them) will incur this 12x penalty.
You could take advantage of caching. But caching won't work if your cache overflows before the next drive can use the already read data. So you'll either see caching as a win, or no help.
Of course, you can make bigger copy block sizes. This cuts the seek penalty down quite a bit. But then again, this 12x improvement in the hypothetical example is a hypothetically truly crappy SSD. Getting a Sandforce or Marvell controller gets you read speeds up 4x better than what I used to calculate this example.
A 9x improvement doesn't sound so unbelievable any more.
No doubt Amazon is very mainstream... but I highly doubt that a significant number of people buy their cell phones from Amazon.
To put it bluntly, I might not be the most popular guy in the world, but in the last 10 years, I don't think I've ever heard of even an acquaintance who bought their phone off Amazon.com. (And I've talked with enough people about phones to even know a person who bought an android phone off some used phone swapping site that seemed kinda shady but still worked out.)
I dunno.... it's starting to look pretty much the same to me too. They're both the process of getting root on a device that wasn't allowing you root access on.
The way I see it, it's not a false dichotomy. Why? Pot, kettle, black.
For anybody who has a reasonably open mind, they'd see this: BOTH the Linux and Apple communities have a both a technological aspect and a religious aspect. To find shame that one of these communities has that aspect, should be to find shame in both communities.
The driving force behind adoption of both camps is pretty much opposite to each other.
Linux favors GPL code which is about as anti-large-company as you can get, which pairs well with the linux camp's religious aspect, mostly driven by Stallman.
Apple favors closed systems and tight control, all in the name of a coherent user experience and the "one true way" sort of philosophy, mostly driven by Jobs.
I find it funny that you try to argue false dichotomy on the "tech as religion" front while simultaneously implying that Apple directly implies "unethical" especially when it's only from a GNU/RMS supporter's viewpoint. Your implied falsehood that Apple raided Jason's house is just as abusive.
Simply put, use of technology from each camp has not only the disadvantages that people dredge up, but significant advantages as well. It's just a shame that you don't see it that way.
For those who've become "true believers" of their respective communities... um, good for you; glad you found your calling. But always remember, not everybody cares about their tech the same way you do. For some people, the freedom of code matters. For some people, it doesn't. For some people, the walled garden is a blessing. For some people, it's bondage.
In case you're wondering: I use to be a Gentoo and Debian fan. I switched all my servers to OpenSolaris. And while I'm typing this on a Macbook Pro; I still cringe when I see the "true believer mac fanboys." If it isn't clear by now, I agree with jimicus when he says that his computer should "get the hell out of the way so I can do something useful."
If Linux were GPLv3, you'd see a mass exodus for any reasonably powerful embedded system. And Android would probably get ported to a BSD by Google. So yeah, no patent suits would result, but mostly because nobody'd use it anymore.
True. In fact, every non-low end phone sold in the last years has a triplet of sensors for each one of accelerometer, magnetometer and gyroscope, I think.
Name some from 2006 and 2007 and before which include all 3. Even the original iPhone had only accelerometers. The compass and gyro didn't come until later.
Which is bad for everybody, even Apple users: the iPhone already costs $650 - the next iPhone might cost even more...
What's to say patents like these can't be used to counter other patents as leverage against those who sue Apple, and therefore decrease future costs?
I'm not saying patents are good or bad. I'm saying that patents can be used in various ways. Just because a few of the ways are unfriendly doesn't mean they'll automatically be used in that manner.
Besides, the $650 price tag for a smartphone isn't unusual. Pretty much every decent-quality smartphone costs that much regardless of the OS. Selling it at component cost would be cheap, but then you can't pay for R&D.
This "good always wins in the end" thing certainly isn't catering to Chinese. It's an American thing.
Afterall you can watch "The Departed" and then the original "Infernal Affairs." I'd have to say the Chinese original is better because there's no stupid "good guy wins" ending.
And while we're at it, half the Korean dramas have another formulaic ending: the "everybody but one major character dies in the end, and he/she is left standing to either mope alone or wonder wtf happened." Personally, I think it's fun to watch the storyline tie up like that.
Sure, it's nice that KHTML code contributed to text/image layout in WebKit. But KHTML itself isn't all that cool, nor pushing the envelope. Apple chose it because it was simply the least crufty, most easily understood, and not GPL. As much as we should be thankful that KHTML was written, it was not groundbreaking.
Between roughly Netscape 2 and Google Chrome, I'd have to argue that the entire ecosystem of web browsers had practically nothing groundbreaking. To me, the history of the web browser goes something like this: 1991: WorldWideWeb was released. First web browser. Utilizes tagged text for automatic layout and linking to other pages. 1993: NCSA Mosaic. Brings inlined images and established the traditional UI we expect today. (back, forward, bookmarks, address bar, stop, refresh) 1996: Netscape 2. Added scripting via Javascript. 1996-2008 - "Warring States Period": Tons of people trying to establish proprietary tags, incompatible formatting (XHTML for the lose), and binary plugin technologies. Popularized IE6. Made Flash relevant. Mozilla became bloated. IE6 became the easiest way to transmit viruses. 2008: Google Chrome. Added sandboxing.
I guess cars should come with the hood locked, and only the dealer has the keys?
Funny you should say that, I got a story. A friend of mine left his car lights on and the battery ran out. Needed a jumpstart. He couldn't figure out how to open the hood. Several friends tried to help him too and it took them somewhere around 15 minutes to open the hood.
Basically, cars don't need to come with the hood locked by the dealer. Whether or not you're able to open the hood yourself sets the benchmark. (and if you think these guys are complete idiots: The car owner's a physicist. First friend to try to help is employed by Google; specialized in NetSec.)
And... iPhones don't need to come with a easy way to use another store, cuz figuring out how to jailbreak it (as opposed to running somebody else's tool) means you probably understand enough to be safe. Running somebody else's jailbreaking tool at least means you understand that stuff that goes wrong is probably caused by something you did.
And I found a fascinating discussion of patching the Linux memory management system to save power by shutting down parts of the RAM that aren't needed! Which version of UNIX did they copy that from? Or will you concede that this is new?
Compaq/DEC Tru64. While doing it to save power is a interesting idea, all the groundwork necessary to do so existed well before Linux became popular. So while it'll take some work to implement in Linux, it ought to be easy in many old UNIXs if it wern't for the fact that power usage didn't really matter to those users.
In Tru64's case, to ability to shut down chunks of RAM is necessary when your hardware supports hotswapping RAM modules.
Solaris/SPARC does this too.
I'm admitting the idea is novel and new, at least to me. But also noting that all the hard work necessary to make it happen on Linux compared to old-school UNIX is what makes Linux "behind the curve."
Except do you count Apple as part of the FOSS community? It's easy to see that a lot of people don't. As far as I recall, practically nobody used KHTML until Apple forked it.
The FOSS community has definitely produced a lot of code, no doubt. But CharlyFoxtrot wasn't asking about quantity of code, but rather creativity/innovation. I expect a FOSS alternative to Dropbox will be little more than cloning Dropbox, and a less usable user interface.
Pushing the envelope seems to come more from the Open Source community (non-GNU), like ZFS, Apache, and PostgreSQL. From the FOSS side, we have GCC, which is definitely something to be thankful for. But looks to be passed up by LLVM soon.
The GPL bothers me too. If I had the choice to utilize an open source project that was GPL licensed and BSD/Apache/MIT licensed, I'd pick the less restrictive license and still release my changes. The GPL is simply something I'd rather not deal with because it imposes restrictions on code that the project does not own.
LGPL is fine, although I do find it to be a problem when in a proprietary project and on platforms where dynamic linking isn't possible. In that specific case, LGPL essentially becomes the GPL again.
Simply put, I feel the choice to add functionality, minor or major from another project, should not impose restrictions outside of that project. I'm perfectly okay with making changes in a project and contributing them back, even handing over control of copyright to the project owner. But having that escape the project boundaries and impose restrictions on code not sourced from the project (and usually not something you'd integrate back into the project) is simply something I'm not cool with.
From what I heard, when iPhones started needing versions of iTunes that don't run on 10.4 while 10.5 was already not available, it was possible to call in to Apple's tech support, tell them you can't find 10.5, and they'd send you 10.5 for free.
I run ZFS on my personal file server with questionably old 160GB drives.
There isn't a "fsck" per say, but you should do "zpool scrub" periodically, which is kinda similar.
What it actually does it read all your data and checks the checksum. If the checksum fails, it'll know to try whatever it can within your settings to make redundant copies of it (if you allowed it to).
What this is useful for is more to see if your drives are about to kick the bucket out from under you.:P
Slashdot Mirror
In fact, it probably ranks as top 25% believable quotes on Slashdot. His writing style could use a little work (who am I to judge, I write like crap too), but his technical eval seems legit. It sounds pretty logical to me:
You got a fast device (SSD) on a fast bus (SATA3) connected to a slower bus (PCIe 2.0 1x) connected to the Northbridge. What's the result? Bottleneck at the slowest part (PCIe). Toss in an overhead estimate of 20%, and you get a practical top speed of 400MB/sec.
When switching down to smaller packet sizes like 4K read/writes, overhead increases since you'll need to send more bytes down the bus in order to mark the smaller packets. (big packets == less overhead in a well designed system, see GigE Jumbo Frames)
A 20% estimate isn't unrealistic either. USB2.0 might be 480 mb/sec (60MB/sec) but it's got a 30% typical overhead penalty giving you a top speed around 40MB/sec.
AMD and Intel can make the SATA controller closer to the CPU by placing it in the Northbridge. This removes both the PCIe 1X bandwidth problem as well as gives them the opportunity to lower all overhead at the same time by abstracting out the PCIe bus if they choose to. Obvious advantage.
I'm kinda amazed there are slashdotters who don't recognize basic computer architecture problems like this. Like, for anybody wondering why there's no USB3 controllers on Apple's Macs, it isn't because they're trying to rip you off and sell it to you later. It's because, in this case, their engineers know better. Wait for Northbridge integration, anything less is subpar USB3.
Apple dongle doesn't support writing from an app as far as I know.
On the other hand, you could toss videos onto it and import them into the camera roll to play if I wanted, I guess.
There some other manufacturer that had a dongle that does more:
http://www.gizmag.com/zoomit-sd-card-reader/14115/
Pretty much everybody I know of who bought an Android phone did it because they didn't want to switch to AT&T at the time. (back then, the verizon iphone was just a legend)
Heck, I was given a free original iPhone, tried AT&T for 20 days, said "fuck this dropped call shit", and got myself a WinMo 6 phone on Verizon.
If carriers were that equal, yeah, then switching would be fine. But if you get more dropped calls in 1 week on AT&T than you did in 2 years of Verizon, switching is a big deal.
Well written and much more diplomatic than what I would have said:
"Haha, suck it, Mr. We-don't-need-no-stable-kernel-ABI! Not so much fun being on the other end of the 'can't we program to a uniform standard' problem, is it, Linus?"
*sigh* I would so mod this up if I had points.
Google is providing incentives to handset manufacturers to get clearance with Google with regards to what they're allowed to do in order to install some Google apps like the Market and Navigation. So it's not entirely a "do whatever with it for free" situation.
If you want to see Google going shady, how about the Motorola/Skyhook deal?
http://www.engadget.com/2011/05/10/internal-emails-reveal-googles-desperation-over-skyhooks-andro/
Let's get some quick info about a WD Velociraptor. (http://www.anandtech.com/show/2508)
98.4MB/sec avg sequential.
4.6ms seek time.
During 4.6ms, you can do 463.5KB of transfer.
To do a 4K transfer, it's 0.04ms.
Now consider some sort of file lookup. What do you need to do?
You'll need to seek to the part of the disk which has your file system's tables, which probably points to a few tree nodes. This is equivalent to walking a 4 node linked list, I guess. Each of these would be a seek, and a 4KB read.
At this rate, you're spending 18.4ms moving the heads, and 0.16ms actually reading.
Afterwards, you can then chug along at your blazing 98.4MB/sec reading the file.... which might be just 128KB. (1.27ms)
Out of a total of 19.83ms, 1.27ms is actually spent doing your file transfer.
Going with a crappy SSD. Say 0.03ms seek time, and 98.4MS/sec sequential (so I don't have to keep typing in numbers into a calculator).
This means we're doing a 0.12ms for a seek, 0.16ms traversing the filesystem, and then 1.27ms reading the file.
Out of a total of 1.55ms... well, we're already 12.7 times faster than the hard drive, so uh, yeah.
So you're probably thinking, okay, but hey, this is for a random 128KB read. And we're trying to copy sequentially to another image, right?
Well, if you're doing multiple reads of the same template to multiple drives, and you just so happen to be doing these out of sync, then copying 128KB to each target drive will require a read from different locations, yeah? Which means every chunk (depending on how you schedule them) will incur this 12x penalty.
You could take advantage of caching. But caching won't work if your cache overflows before the next drive can use the already read data. So you'll either see caching as a win, or no help.
Of course, you can make bigger copy block sizes. This cuts the seek penalty down quite a bit. But then again, this 12x improvement in the hypothetical example is a hypothetically truly crappy SSD. Getting a Sandforce or Marvell controller gets you read speeds up 4x better than what I used to calculate this example.
A 9x improvement doesn't sound so unbelievable any more.
Why does Bluetooth HID/SPP need a custom kernel?
Isn't it possible to write a userspace daemon to handle serial communications to a Bluetooth radio?
No doubt Amazon is very mainstream... but I highly doubt that a significant number of people buy their cell phones from Amazon.
To put it bluntly, I might not be the most popular guy in the world, but in the last 10 years, I don't think I've ever heard of even an acquaintance who bought their phone off Amazon.com. (And I've talked with enough people about phones to even know a person who bought an android phone off some used phone swapping site that seemed kinda shady but still worked out.)
I dunno.... it's starting to look pretty much the same to me too. They're both the process of getting root on a device that wasn't allowing you root access on.
If you disagree, then elaborate, plz?
The way I see it, it's not a false dichotomy. Why? Pot, kettle, black.
For anybody who has a reasonably open mind, they'd see this:
BOTH the Linux and Apple communities have a both a technological aspect and a religious aspect. To find shame that one of these communities has that aspect, should be to find shame in both communities.
The driving force behind adoption of both camps is pretty much opposite to each other.
Linux favors GPL code which is about as anti-large-company as you can get, which pairs well with the linux camp's religious aspect, mostly driven by Stallman.
Apple favors closed systems and tight control, all in the name of a coherent user experience and the "one true way" sort of philosophy, mostly driven by Jobs.
I find it funny that you try to argue false dichotomy on the "tech as religion" front while simultaneously implying that Apple directly implies "unethical" especially when it's only from a GNU/RMS supporter's viewpoint. Your implied falsehood that Apple raided Jason's house is just as abusive.
Simply put, use of technology from each camp has not only the disadvantages that people dredge up, but significant advantages as well. It's just a shame that you don't see it that way.
For those who've become "true believers" of their respective communities... um, good for you; glad you found your calling. But always remember, not everybody cares about their tech the same way you do. For some people, the freedom of code matters. For some people, it doesn't. For some people, the walled garden is a blessing. For some people, it's bondage.
In case you're wondering: I use to be a Gentoo and Debian fan. I switched all my servers to OpenSolaris. And while I'm typing this on a Macbook Pro; I still cringe when I see the "true believer mac fanboys."
If it isn't clear by now, I agree with jimicus when he says that his computer should "get the hell out of the way so I can do something useful."
If Linux were GPLv3, you'd see a mass exodus for any reasonably powerful embedded system. And Android would probably get ported to a BSD by Google.
So yeah, no patent suits would result, but mostly because nobody'd use it anymore.
True. In fact, every non-low end phone sold in the last years has a triplet of sensors for each one of accelerometer, magnetometer and gyroscope, I think.
Name some from 2006 and 2007 and before which include all 3. Even the original iPhone had only accelerometers. The compass and gyro didn't come until later.
Which is bad for everybody, even Apple users: the iPhone already costs $650 - the next iPhone might cost even more...
What's to say patents like these can't be used to counter other patents as leverage against those who sue Apple, and therefore decrease future costs?
I'm not saying patents are good or bad. I'm saying that patents can be used in various ways. Just because a few of the ways are unfriendly doesn't mean they'll automatically be used in that manner.
Besides, the $650 price tag for a smartphone isn't unusual. Pretty much every decent-quality smartphone costs that much regardless of the OS. Selling it at component cost would be cheap, but then you can't pay for R&D.
Mainstream microkernels? You mean QNX?
What's this HURD thing? ;)
i dunno man. you could also blame apple for not using usb. in fact you could also blame yourself for buying a shitty mp3 player.
USB2 wasn't available back then.
Isn't this something like how the iPhone 4's wavefront coding camera works?
This "good always wins in the end" thing certainly isn't catering to Chinese. It's an American thing.
Afterall you can watch "The Departed" and then the original "Infernal Affairs." I'd have to say the Chinese original is better because there's no stupid "good guy wins" ending.
And while we're at it, half the Korean dramas have another formulaic ending: the "everybody but one major character dies in the end, and he/she is left standing to either mope alone or wonder wtf happened." Personally, I think it's fun to watch the storyline tie up like that.
Um, no. My Nexus S came with it out of the box December last year. Source code was available then too.
Sure, it's nice that KHTML code contributed to text/image layout in WebKit.
But KHTML itself isn't all that cool, nor pushing the envelope. Apple chose it because it was simply the least crufty, most easily understood, and not GPL. As much as we should be thankful that KHTML was written, it was not groundbreaking.
Between roughly Netscape 2 and Google Chrome, I'd have to argue that the entire ecosystem of web browsers had practically nothing groundbreaking. To me, the history of the web browser goes something like this:
1991: WorldWideWeb was released. First web browser. Utilizes tagged text for automatic layout and linking to other pages.
1993: NCSA Mosaic. Brings inlined images and established the traditional UI we expect today. (back, forward, bookmarks, address bar, stop, refresh)
1996: Netscape 2. Added scripting via Javascript.
1996-2008 - "Warring States Period": Tons of people trying to establish proprietary tags, incompatible formatting (XHTML for the lose), and binary plugin technologies. Popularized IE6. Made Flash relevant. Mozilla became bloated. IE6 became the easiest way to transmit viruses.
2008: Google Chrome. Added sandboxing.
I guess cars should come with the hood locked, and only the dealer has the keys?
Funny you should say that, I got a story. A friend of mine left his car lights on and the battery ran out. Needed a jumpstart.
He couldn't figure out how to open the hood. Several friends tried to help him too and it took them somewhere around 15 minutes to open the hood.
Basically, cars don't need to come with the hood locked by the dealer. Whether or not you're able to open the hood yourself sets the benchmark.
(and if you think these guys are complete idiots: The car owner's a physicist. First friend to try to help is employed by Google; specialized in NetSec.)
And... iPhones don't need to come with a easy way to use another store, cuz figuring out how to jailbreak it (as opposed to running somebody else's tool) means you probably understand enough to be safe. Running somebody else's jailbreaking tool at least means you understand that stuff that goes wrong is probably caused by something you did.
And I found a fascinating discussion of patching the Linux memory management system to save power by shutting down parts of the RAM that aren't needed! Which version of UNIX did they copy that from? Or will you concede that this is new?
Compaq/DEC Tru64.
While doing it to save power is a interesting idea, all the groundwork necessary to do so existed well before Linux became popular. So while it'll take some work to implement in Linux, it ought to be easy in many old UNIXs if it wern't for the fact that power usage didn't really matter to those users.
In Tru64's case, to ability to shut down chunks of RAM is necessary when your hardware supports hotswapping RAM modules.
Solaris/SPARC does this too.
I'm admitting the idea is novel and new, at least to me. But also noting that all the hard work necessary to make it happen on Linux compared to old-school UNIX is what makes Linux "behind the curve."
Except do you count Apple as part of the FOSS community? It's easy to see that a lot of people don't.
As far as I recall, practically nobody used KHTML until Apple forked it.
The FOSS community has definitely produced a lot of code, no doubt. But CharlyFoxtrot wasn't asking about quantity of code, but rather creativity/innovation.
I expect a FOSS alternative to Dropbox will be little more than cloning Dropbox, and a less usable user interface.
Pushing the envelope seems to come more from the Open Source community (non-GNU), like ZFS, Apache, and PostgreSQL.
From the FOSS side, we have GCC, which is definitely something to be thankful for. But looks to be passed up by LLVM soon.
The GPL bothers me too. If I had the choice to utilize an open source project that was GPL licensed and BSD/Apache/MIT licensed, I'd pick the less restrictive license and still release my changes. The GPL is simply something I'd rather not deal with because it imposes restrictions on code that the project does not own.
LGPL is fine, although I do find it to be a problem when in a proprietary project and on platforms where dynamic linking isn't possible. In that specific case, LGPL essentially becomes the GPL again.
Simply put, I feel the choice to add functionality, minor or major from another project, should not impose restrictions outside of that project.
I'm perfectly okay with making changes in a project and contributing them back, even handing over control of copyright to the project owner. But having that escape the project boundaries and impose restrictions on code not sourced from the project (and usually not something you'd integrate back into the project) is simply something I'm not cool with.
From what I heard, when iPhones started needing versions of iTunes that don't run on 10.4 while 10.5 was already not available, it was possible to call in to Apple's tech support, tell them you can't find 10.5, and they'd send you 10.5 for free.
I run ZFS on my personal file server with questionably old 160GB drives.
There isn't a "fsck" per say, but you should do "zpool scrub" periodically, which is kinda similar.
What it actually does it read all your data and checks the checksum. If the checksum fails, it'll know to try whatever it can within your settings to make redundant copies of it (if you allowed it to).
What this is useful for is more to see if your drives are about to kick the bucket out from under you. :P
ZFS is awesome.