Before the instruction you posted, I found this explanation in TFA:
An enormous amount of the world's commerce operates on this assumption, including your daily credit card transactions. However, if this makes you uneasy, that's OK: ZFS provies a 'verify' option that performs a full comparison of every incoming block with any alleged duplicate to ensure that they really are the same, and ZFS resolves the conflict if not. To enable this variant of dedup, just specify 'verify' instead of 'on':
I fail to see how someone can sit down and rationally decide whether their data will be more susceptible to hash collisions or not. While I would be very surprised if any two blocks on my computer hash to the same value in spite of being different, it seems to me that someone's going to get hit by this sooner rather than later. And what a nasty way to find hash collisions! Who would have thought my Aunt's chocolate cake recipe had the same SHA1 as hello.jpg from goatse.cx!.
On one hand, 2^256 is a damn big keyspace. I've heard people say a collision is about as likely as winning every lottery in the world simultaneously, and then doing it again next week. Bug give enough computers with enough blocks enough time, and find a SHA1 collision you will. Depending on what kind of data it happens to, you might not even notice it.
Maybe he isn't. Perhaps he's drawing from his experience over the last decade of vendor after vendor treating FOSS like a red-headed freckled stepchild.
You sir, just won that pissing contest. The only person who could take it away from you now would need to post the portion of Windows source code that emits the error.
Hey, give them credit - I mean, you also get a tray app that tells you when to drop another hundred on ink cartridges, and slows down your login giving you time to pop down the shop and get some. That alone makes it worth the DVD-sized download!
By your (un)logic, MS released the 2000th version of their operating system about a decade ago. Porsche scrapped 910 different designs before setting on the immortal 911. Sun Microsystems gave up development of 5.11 in 1995 and started selling old copies from their archives, beginning with 2.6, then 2.7, before shocking the world and jumping all the way up to 8!
I'm not so much worried about the bugs that I am "unlikely to find", so much as the bugs others are likely to find.
Of course, I'm probably being paranoid, after all MS have told me time and time again that Win7 is going to be more secure, and I can be pig-headed at times.
Do you often bump into the slashdot filter "That exact same comment has already been posted" much?
It's just that I could swear I've seen you trot out that same story dozens of times. Let's say you did buy Leopard when it came out, and upgraded your machine. You'd still be supported now, and you'd still have spent well less than a copy of XP Pro. You'd get install media that would just work for you and not require you to prove you have the previous OS, and would work on other friend's macs should they unfortunately lose their copy.
Users got a disproportionately good run out of XP. Back when people were talking about "Longhorn" I was almost confident XP was only going to be in the game for a couple more years, I'm sure that was Microsoft's hope at the time too. I don't think the support lifecycle of Win7 will be anywhere near as long.
Again, consider if you have have bought Leopard (10.5) when it came out. Now imagine if, at the same time, you instead decided to switch to PC, so you get a new one. At the time, you would have got Vista. Which boat would you rather be in now? And how handy is that Vista-OEM install DVD going to be? Better not lose it though, chance are you won't be able to use anything else to recover your install with the only key you've got. Likewise, you won't be able to use your DVD to reinstall someone else's machine because the probability of finding another person with the same model is practically zero.
It's hard to compare the deals you get between OSX and Windows completely on price, but if you have have upgraded just once over that time, you'd be getting a much better deal in terms of support, and still come in cheaper than the purchase of a Windows OS.
Even some logged in users have to fill out captchas. It's quite a bit of work convincing slashcode that you're not a bot and have a ready supply of crack for the moderation.
Okay, while I haven't found one specifically about the Lunar Landings (in the NZ/Australia app store) there is one (cr)app called "Conspiracy Video!" for NZ$1.29 covering many uhm, 'debated' events in history, the Lunar Landings, JFK, 9/11 of course, and even up to the minute coverage with Michael Jackson.
Like I said, I fully expected by now someone would have made a Moon Landing Hoax app by now - I was close - perhaps nobody has, perhaps Apple won't let it fly.
No you do not want your vehicle helpfully straightening you out of a skid if you were in that skid in the first place to avoid something like a pedestrian or a B-train coming out of nowhere.
Agree - there's so many reasons to RAID and different levels of availability. Some people would shudder at the absence of a hot-spare drive, never mind not immediately hot-swapping in a replacement. Here's a few (in my opinion) simplified "reasons" and ways to do RAID.
A software mirror on cold-plugging drives (what I do on my personal workstation)
Essentially saves the day when the inevitable happens, and your client is with you. Your hourly rate keeps getting charged out, you finish the job, send your client home, and get your RAID straightened in first order. Phew! Cheap insurance, that's all that can be said.
Hardware/Software RAID 1 or 5 or similar, hot swappable, hopefully something that passes as a DR plan
We're rich, we can afford to keep cold spares or are guaranteed their fast delivery. Even scheduled downtime is expensive for us, so we'd rather not ever shut down if we can avoid it.
All out hardware RAID or SAN with hot-spares at the ready, DR plan that's been tested.
As above, but to the next level. If we lose that RAID and don't recover within 24 hours, we might as well pack up. History is riddled with example of companies that have proven this.
This new Mac Mini server suits me, in terms of the trade-off between cost and [availa|relia]bility. Admittedly prying open a Mini to replace a teensy weensy drive at 5PM Friday isn't my idea of fun, but I wouldn't be in the trade if I didn't thoroughly hate myself already.
Actually, BFS performance is shitty. No, really shitty.
Damn. And here was me thinking that a name like "Brain Fuck Scheduler" would get some attention and give Linux just the push it needs to succeed on the desktop.
Out of context I'd agree with you, but the question posed can't really be answered without identifying the underlying codec, so it's as good a time as any to point this out. GP probably knew he'd be called up for being a dick, but posted anyway.
For goodness' sake the term is downsampling - we speak of it distinctly from compression for reasons I'd hoped were obvious.
There's no algorithm involved - you simply discard the information regardless of its contents, compression algo's don't do that. The resulting data is downsampled, not compressed, because the loss of quality reflects the reduction in data size directly. It is a lower-resolution uncompressed copy.
Compressors use computational power. Downsampling can be as simple as electrically 'ignoring' certain bits or samples.
Gee, if anyone could work a car analogy in here that'd be great, thanks.
Slashdot Mirror
I fail to see how someone can sit down and rationally decide whether their data will be more susceptible to hash collisions or not. While I would be very surprised if any two blocks on my computer hash to the same value in spite of being different, it seems to me that someone's going to get hit by this sooner rather than later. And what a nasty way to find hash collisions! Who would have thought my Aunt's chocolate cake recipe had the same SHA1 as hello.jpg from goatse.cx!.
On one hand, 2^256 is a damn big keyspace. I've heard people say a collision is about as likely as winning every lottery in the world simultaneously, and then doing it again next week. Bug give enough computers with enough blocks enough time, and find a SHA1 collision you will. Depending on what kind of data it happens to, you might not even notice it.
Maybe he isn't. Perhaps he's drawing from his experience over the last decade of vendor after vendor treating FOSS like a red-headed freckled stepchild.
You sir, just won that pissing contest. The only person who could take it away from you now would need to post the portion of Windows source code that emits the error.
Hey, give them credit - I mean, you also get a tray app that tells you when to drop another hundred on ink cartridges, and slows down your login giving you time to pop down the shop and get some. That alone makes it worth the DVD-sized download!
Or has has used his guest account, and was one of the 99% of users not affected.
It shouldn't. Windows 7 has barely been out a week.
From memory that was the only Win95 with USB "support".
By your (un)logic, MS released the 2000th version of their operating system about a decade ago. Porsche scrapped 910 different designs before setting on the immortal 911. Sun Microsystems gave up development of 5.11 in 1995 and started selling old copies from their archives, beginning with 2.6, then 2.7, before shocking the world and jumping all the way up to 8!
You always deliberately misinterpret people when idiosyncrasies of the language show up? Yikes!
Of course, I'm probably being paranoid, after all MS have told me time and time again that Win7 is going to be more secure, and I can be pig-headed at times.
It's just that I could swear I've seen you trot out that same story dozens of times. Let's say you did buy Leopard when it came out, and upgraded your machine. You'd still be supported now, and you'd still have spent well less than a copy of XP Pro. You'd get install media that would just work for you and not require you to prove you have the previous OS, and would work on other friend's macs should they unfortunately lose their copy.
Users got a disproportionately good run out of XP. Back when people were talking about "Longhorn" I was almost confident XP was only going to be in the game for a couple more years, I'm sure that was Microsoft's hope at the time too. I don't think the support lifecycle of Win7 will be anywhere near as long.
Again, consider if you have have bought Leopard (10.5) when it came out. Now imagine if, at the same time, you instead decided to switch to PC, so you get a new one. At the time, you would have got Vista. Which boat would you rather be in now? And how handy is that Vista-OEM install DVD going to be? Better not lose it though, chance are you won't be able to use anything else to recover your install with the only key you've got. Likewise, you won't be able to use your DVD to reinstall someone else's machine because the probability of finding another person with the same model is practically zero.
It's hard to compare the deals you get between OSX and Windows completely on price, but if you have have upgraded just once over that time, you'd be getting a much better deal in terms of support, and still come in cheaper than the purchase of a Windows OS.
Even some logged in users have to fill out captchas. It's quite a bit of work convincing slashcode that you're not a bot and have a ready supply of crack for the moderation.
I hate to break this to you, but the only people taking /. moderation seriously lately are you and the GP. It's time to stop, and read at -1.
Oh really?
Like I said, I fully expected by now someone would have made a Moon Landing Hoax app by now - I was close - perhaps nobody has, perhaps Apple won't let it fly.
I'm going to check now because I'd be quite surprised if there wasn't already one.
No you do not want your vehicle helpfully straightening you out of a skid if you were in that skid in the first place to avoid something like a pedestrian or a B-train coming out of nowhere.
Up up down down left right left right B A start to start the engine
All that software? Sorry, pal, it's some of it, at best.
A software mirror on cold-plugging drives (what I do on my personal workstation)
Essentially saves the day when the inevitable happens, and your client is with you. Your hourly rate keeps getting charged out, you finish the job, send your client home, and get your RAID straightened in first order. Phew! Cheap insurance, that's all that can be said.
Hardware/Software RAID 1 or 5 or similar, hot swappable, hopefully something that passes as a DR plan
We're rich, we can afford to keep cold spares or are guaranteed their fast delivery. Even scheduled downtime is expensive for us, so we'd rather not ever shut down if we can avoid it.
All out hardware RAID or SAN with hot-spares at the ready, DR plan that's been tested.
As above, but to the next level. If we lose that RAID and don't recover within 24 hours, we might as well pack up. History is riddled with example of companies that have proven this.
This new Mac Mini server suits me, in terms of the trade-off between cost and [availa|relia]bility. Admittedly prying open a Mini to replace a teensy weensy drive at 5PM Friday isn't my idea of fun, but I wouldn't be in the trade if I didn't thoroughly hate myself already.
Damn. And here was me thinking that a name like "Brain Fuck Scheduler" would get some attention and give Linux just the push it needs to succeed on the desktop.
Out of context I'd agree with you, but the question posed can't really be answered without identifying the underlying codec, so it's as good a time as any to point this out. GP probably knew he'd be called up for being a dick, but posted anyway.
No, they said it was a "double-blind" test.
There's no algorithm involved - you simply discard the information regardless of its contents, compression algo's don't do that. The resulting data is downsampled, not compressed, because the loss of quality reflects the reduction in data size directly. It is a lower-resolution uncompressed copy.
Compressors use computational power. Downsampling can be as simple as electrically 'ignoring' certain bits or samples.
Gee, if anyone could work a car analogy in here that'd be great, thanks.