TR Forums

Here are some things that PISS me off:

1) This is 2014 and we still have hard drives that might silently lose data AKA bitrot. We have ZFS to save us but it isn't a mainstream filesystem (Not on Windoze). ReFS: not ready for prime time. And it needs the slow abomination that is Storage Spaces to recover from bitrot errors. I ask you: how hard could it be to design a filesystem that maintains parity bits so that bitrot errors never occur? Why aren't the hard drive makers designing RELIABLE hard drives? They could design one and call it the Infallible Edition or something and put a premium on it. If it's better at retaining data than the current sucky drives, people WILL pay. EYE will PAY.

2) I still can't buy a Helium sealed 6TB hard drive for under $300! What's taking so long????????

3) Why aren't today's operating systems more like hypervisors???? I should be able to freeze any process, save its state to disk and resume it whenever I want.

4) Why aren't applications more intelligent? Instead of throwing up errors and shutting down when they encounter sudden failure of some sort, why don't they offer to work AROUND the failure condition? Give us options! Don't just bloody blow up in our faces, taking our data and sanity with you :S

5) What is up with hybrid drives? There is NO competition in this space. Just Seagate out there with their paltry 8GB of MLC in their hybrids. How long is that gonna last? There should be at least 32GB of high endurance MLC in a hybrid. Sell it at a premium dammit but SELL IT!

Igor_Kavinski wrote:

Here are some things that PISS me off:


2) I still can't buy a Helium sealed 6TB hard drive for under $300! What's taking so long????????

!

Because the gov't stupidly sold off our strategic supply of helium. There's less of it left, so it's not gonna be cheap.

And why helium, not argon, xenon, or nitrogen? Those are commonly used in zero-oxygen applications.

The future will come, just not as fast as you'd like.

1. "Reliable" hard drives don't really exist, and they can't. The actual area on the HDD disk that a bit occupies is so small that random space radiation will ruin the data. Unfortunately physics ruins it. Also, a file system that protects against such an occurrence? I thought RAID 1 fixed that.

2. Probably because of that flood that ruined HDD prices.

3. Had to look up hypervisor. It involves virtualization right? I don't think virtualization can have the same performance that regular stuff has. Also, I don't think the average user has a use for hibernating a program.

4. Because developers are humans usually being pushed by deadlines. It could also be a security thing. You really don't want a bug to start running code that isn't code and killing it right then and there may be the only way to guarantee that it doesn't.

5. Never understood hybrid drives. You want fast? SSD You want big? HDD Need both? Get both. SSDs have come down in price enough that a small SSD is still big enough for a lot of stuff.

Eh, I needed to type. I feel out of it for some reason.

@ Hz so good: Why use helium? Less drag. Why not just use nitrogen? It's basically just regular air. They use that lessened drag to improve power efficiency. Also a vacuum doesn't work because the external pressure would crush it.

Igor_Kavinski wrote:

Here are some things that PISS me off:

1) This is 2014 and we still have hard drives that might silently lose data AKA bitrot. We have ZFS to save us but it isn't a mainstream filesystem (Not on Windoze). ReFS: not ready for prime time. And it needs the slow abomination that is Storage Spaces to recover from bitrot errors. I ask you: how hard could it be to design a filesystem that maintains parity bits so that bitrot errors never occur? Why aren't the hard drive makers designing RELIABLE hard drives? They could design one and call it the Infallible Edition or something and put a premium on it. If it's better at retaining data than the current sucky drives, people WILL pay. EYE will PAY.

ZFS is available for Linux but its not mainstream there and its extremely unlikely it ever will be due to licensing issues. It can't be bundled with the kernel. Anyway, Btrfs exist and its really only a matter of tools maturity at this point. I haven't followed ReFS that closely (I'm a *nix admin, not a Windows one) but from my understanding it was available to use for non-OS partitions in the latest releases of Windows Server.

As far as "Infallible Edition" hard drives.... well that's simply impossible. As well, it's not as if the hard drive manufacturer is going to dictate which file system you choose.

2) I still can't buy a Helium sealed 6TB hard drive for under $300! What's taking so long????????

What are you expecting to happen? Hard drive manufacturers can charge a premium for these devices. They'll continue doing so until competition forces them otherwise and I don't see that happening soon.

3) Why aren't today's operating systems more like hypervisors???? I should be able to freeze any process, save its state to disk and resume it whenever I want.

Honestly I'm not sure what you're asking/wanting here. There's a number of ways to achieve your second sentence.

4) Why aren't applications more intelligent? Instead of throwing up errors and shutting down when they encounter sudden failure of some sort, why don't they offer to work AROUND the failure condition? Give us options! Don't just bloody blow up in our faces, taking our data and sanity with you :S

What? Depending on the type of work your application is performing it might simply be impossible to do what you're asking. Regardless, when it would be possible it's often an extremely challenging development task. Challenging development issues take time and therefore money. If such robustness is required/wanted then you're going to pay for it.

5) What is up with hybrid drives? There is NO competition in this space. Just Seagate out there with their paltry 8GB of MLC in their hybrids. How long is that gonna last? There should be at least 32GB of high endurance MLC in a hybrid. Sell it at a premium dammit but SELL IT!

Likely because the market demand for such a drive isn't honestly that large.

Ari Atari wrote:

Eh, I needed to type. I feel out of it for some reason.

@ Hz so good: Why use helium? Less drag. Why not just use nitrogen? It's basically just regular air. They use that lessened drag to improve power efficiency. Also a vacuum doesn't work because the external pressure would crush it.

Then what about argon and xenon? Those are used all the time in Glass glove boxes, when zero-oxygen environments are required for the experiments they're doing. And few of those actually use a vacuum.

What's the drag coefficient of other gases besides helium? Genuinely curious.

Ari Atari wrote:

The future will come, just not as fast as you'd like.

1. "Reliable" hard drives don't really exist, and they can't. The actual area on the HDD disk that a bit occupies is so small that random space radiation will ruin the data. Unfortunately physics ruins it. Also, a file system that protects against such an occurrence? I thought RAID 1 fixed that.

RAID 1 does not protect against Bitrot. There is a good, recent thread on this topic here: viewtopic.php?f=6&t=94296

I guess I'm more pissed because there just isn't enough innovation happening these days. We have a planet full of more than enough brilliant minds. They are just not putting their collective intelligence to good use and breaking new ground. I'm tired of seeing incremental improvements. I want mind blowing technology while I am still alive. Hate how we still have to program computers. Instead of wrestling with code, we should be able to design solutions intuitively in a visual manner. Like putting together Lego blocks. We are just so woefully behind what science fiction promised us in the 21st century. I hate how everyone uses the demand excuse to justify the lack of innovation. No one wanted an iPhone before it existed but that didn't stop Apple from making it. No one was crying for Android before they made it. Churning out more of the same with different flavors seems to be what everyone is doing these days. That sucks!

Hz so good wrote:

Ari Atari wrote:

Eh, I needed to type. I feel out of it for some reason.

@ Hz so good: Why use helium? Less drag. Why not just use nitrogen? It's basically just regular air. They use that lessened drag to improve power efficiency. Also a vacuum doesn't work because the external pressure would crush it.

Then what about argon and xenon? Those are used all the time in Glass glove boxes, when zero-oxygen environments are required for the experiments they're doing. And few of those actually use a vacuum.

What's the drag coefficient of other gases besides helium? Genuinely curious.

I think the issue is that you want a gas that is very light, and non-reactive. Helium is the second lightest gas. The lightest one (Hydrogen) is highly reactive and tends to go "boom" too easily. All of the other noble gases are heavier than Helium.

slowriot wrote:

Ari Atari wrote:

The future will come, just not as fast as you'd like.

1. "Reliable" hard drives don't really exist, and they can't. The actual area on the HDD disk that a bit occupies is so small that random space radiation will ruin the data. Unfortunately physics ruins it. Also, a file system that protects against such an occurrence? I thought RAID 1 fixed that.

RAID 1 does not protect against Bitrot. There is a good, recent thread on this topic here: viewtopic.php?f=6&t=94296

Yup. RAID-1 allows you to recover from a failed drive or unreadable sector. But if bits get flipped in a way that the drive's ECC doesn't detect (note: this is the definition of bitrot), you have no way of knowing which drive in the RAID-1 array has the *correct* copy. Heck, you probably won't even realize the error has occurred right away, since RAID-1 (typically) does not compare the copies on the two drives; it just reads the data from whichever drive's head is already positioned close enough to the desired track.

My turn:

It's 2014 and UPS doesn't have an app that allows you to GPS-locate the truck with your package on it. There's an app, but it doesn't track the trucks. If you have a tracking number you should be able to track the physical location of the package. You still have to wait for it to arrive at it's destination, of course, but at least you will know within ~1 to 2 hours time when it will show up.

And don't gimme that crap about security--BLAH BLAH, nobody's gonna "guess" my billion-digit-long tracking number at the precise right time. If that's still not enough security, then maybe have the app ask for the name of the intended recipient, or their birthday, whatever.

Igor_Kavinski wrote:

4) Why aren't applications more intelligent? Instead of throwing up errors and shutting down when they encounter sudden failure of some sort, why don't they offer to work AROUND the failure condition? Give us options! Don't just bloody blow up in our faces, taking our data and sanity with you :S

The most common type of application crash is caused by an invalid memory refefence, where the application has attempted to access a non-existent memory location or execute non-existent code. This is actually detected at the virtual memory manager level (down in the OS), not by the application. The OS has no idea what the application was trying to do; the *only* sensible thing for the OS to do in this situation is to terminate the application.

ssidbroadcast wrote:

My turn:

It's 2014 and UPS doesn't have an app that allows you to GPS-locate the truck with your package on it. There's an app, but it doesn't track the trucks. If you have a tracking number you should be able to track the physical location of the package. You still have to wait for it to arrive at it's destination, of course, but at least you will know within ~1 to 2 hours time when it will show up.

And don't gimme that crap about security--BLAH BLAH, nobody's gonna "guess" my billion-digit-long tracking number at the precise right time. If that's still not enough security, then maybe have the app ask for the name of the intended recipient, or their birthday, whatever.

Your impatience to receive that package doesn't negate the fact that there ARE security issues associated with shipping. Plus, if the truck was trackable by consumers, every person on earth who is itching to either get a package or turn in a package would try to track down the nearest driver en-route and completely screw with the logistics of the route mapping. Keep in mind these are companies whose logistics software does things like eliminate left turns from the route in order to reduce total road time and save fuel.

Igor_Kavinski wrote:

I guess I'm more pissed because there just isn't enough innovation happening these days. We have a planet full of more than enough brilliant minds. They are just not putting their collective intelligence to good use and breaking new ground. I'm tired of seeing incremental improvements. I want mind blowing technology while I am still alive. Hate how we still have to program computers. Instead of wrestling with code, we should be able to design solutions intuitively in a visual manner. Like putting together Lego blocks. We are just so woefully behind what science fiction promised us in the 21st century. I hate how everyone uses the demand excuse to justify the lack of innovation. No one wanted an iPhone before it existed but that didn't stop Apple from making it. No one was crying for Android before they made it. Churning out more of the same with different flavors seems to be what everyone is doing these days. That sucks!

There's plenty of innovation happening these days. It just isn't the type of "gee whiz" sci-fi leap in innovation you seem to be wanting. Almost all technological advances are evolutionary, not revolutionary (and have been, throughout history).

Igor_Kavinski wrote:

I guess I'm more pissed because there just isn't enough innovation happening these days. We have a planet full of more than enough brilliant minds. They are just not putting their collective intelligence to good use and breaking new ground. I'm tired of seeing incremental improvements. I want mind blowing technology while I am still alive. Hate how we still have to program computers. Instead of wrestling with code, we should be able to design solutions intuitively in a visual manner. Like putting together Lego blocks. We are just so woefully behind what science fiction promised us in the 21st century. I hate how everyone uses the demand excuse to justify the lack of innovation. No one wanted an iPhone before it existed but that didn't stop Apple from making it. No one was crying for Android before they made it. Churning out more of the same with different flavors seems to be what everyone is doing these days. That sucks!

This is entirely about perspective.

Just look at your comments on the iPhone and Android. The iPhone wasn't the first device to do anything. It just did it better than those before it. Likewise, Android was a direct response to the iPhone/iOS. Literally millions of people wanted devices like these, new they were possible and even thousands of them had designs for such devices before the iPhone came out.

Likewise, there are numerous efforts underway to make programming a more intuitive. There are numerous projects which do try to make it like fitting puzzles pieces of building blocks together.

Further, more than ever has collaboration become a key part in innovation. But here's the thing... when you have thousands of minds working towards a common goal you stop having these rare "big innovations" and instead it becomes a series of numerous small innovations and you stop losing that "wow" factor, which is really what you seem after.

Actually, there's probably quite a bit of crazy science fiction stuff being tested, it just can't be produced in a cost-efficient manner. They're not going to sell products unless they can turn a tidy profit.

slowriot wrote:

Ari Atari wrote:

The future will come, just not as fast as you'd like.

1. "Reliable" hard drives don't really exist, and they can't. The actual area on the HDD disk that a bit occupies is so small that random space radiation will ruin the data. Unfortunately physics ruins it. Also, a file system that protects against such an occurrence? I thought RAID 1 fixed that.

RAID 1 does not protect against Bitrot. There is a good, recent thread on this topic here: viewtopic.php?f=6&t=94296

Ah, my bad. I forgot that all the data looks the same. Who is right is just a matter of interpretation.

ludi wrote:

ssidbroadcast wrote:

My turn:

It's 2014 and UPS doesn't have an app that allows you to GPS-locate the truck with your package on it. There's an app, but it doesn't track the trucks. If you have a tracking number you should be able to track the physical location of the package. You still have to wait for it to arrive at it's destination, of course, but at least you will know within ~1 to 2 hours time when it will show up.

And don't gimme that crap about security--BLAH BLAH, nobody's gonna "guess" my billion-digit-long tracking number at the precise right time. If that's still not enough security, then maybe have the app ask for the name of the intended recipient, or their birthday, whatever.

Your impatience to receive that package doesn't negate the fact that there ARE security issues associated with shipping. Plus, if the truck was trackable by consumers, every person on earth who is itching to either get a package or turn in a package would try to track down the nearest driver en-route and completely screw with the logistics of the route mapping. Keep in mind these are companies whose logistics software does things like eliminate left turns from the route in order to reduce total road time and save fuel.

Then don't show me the exact position of the driver. Show me a blinking "district" of the driver's approximate location (ala Final Fight,) and an approximate delivery time within 60 minutes of accuracy. Not a big ask!

Can't you protect yourself from bitrot by making MD5 hashes of your important files and then checking them periodically? I suppose if you only keep one copy of the file that won't help but it's something...

MadManOriginal wrote:

Can't you protect yourself from bitrot by making MD5 hashes of your important files and then checking them periodically? I suppose if you only keep one copy of the file that won't help but it's something...

Yeah, that'll at least tell you if something is amiss.

just brew it! wrote:

Igor_Kavinski wrote:

I guess I'm more pissed because there just isn't enough innovation happening these days. We have a planet full of more than enough brilliant minds. They are just not putting their collective intelligence to good use and breaking new ground. I'm tired of seeing incremental improvements. I want mind blowing technology while I am still alive. Hate how we still have to program computers. Instead of wrestling with code, we should be able to design solutions intuitively in a visual manner. Like putting together Lego blocks. We are just so woefully behind what science fiction promised us in the 21st century. I hate how everyone uses the demand excuse to justify the lack of innovation. No one wanted an iPhone before it existed but that didn't stop Apple from making it. No one was crying for Android before they made it. Churning out more of the same with different flavors seems to be what everyone is doing these days. That sucks!

There's plenty of innovation happening these days. It just isn't the type of "gee whiz" sci-fi leap in innovation you seem to be wanting. Almost all technological advances are evolutionary, not revolutionary (and have been, throughout history).

Although there is certainly that. The US air force, in the past few years, have made a laser based weapon (!), the US Navy will be mounting rail guns to it's ships in the next few years, we're capable of growing non-complex body parts (ears, vocal cords), scientists have transported particles over a few feet (I believe. Too lazy too look up right now, but I believe it was a few feet), battery powered cars that don't totally suck (Tesla) and Hydrogen powered cars (!).

These things are absolutely "gee-whiz", but what a lot of people seem to be forgetting is the almighty $$. If money was no object, then sure, we'd have cold fusion by now, but many of the things above require billions of dollars. And there aren't that many people who have billions lying around...

ssidbroadcast wrote:

Then don't show me the exact position of the driver. Show me a blinking "district" of the driver's approximate location (ala Final Fight,) and an approximate delivery time within 60 minutes of accuracy. Not a big ask!

That's quite a bit to ask, honestly. I'd settle for a driver that actually knocks on the door instead of just leaving a tag and driving away.

It's not all $$$ that limits advancement (or widespread adoption). A great portion is actually figuring things out. Tesla has been more effective at getting EVs on the market because 1) that's all they do, and they have a low production capacity, and 2) they roll with the low volumes by making luxury products.

Ford and GM have taken a few stabs at the hybrid market, but notice how some of those are just hybrid versions of existing platforms? That's because (I assume) each company is producing hybrids on the same lines as "normal" vehicles, and they do that because of low volumes. But putting hybrids or EVs on the same line as "normal" vehicles means that you have to mimic the gas vehicle as much as possible instead of developing a process and architecture that makes sense for something with a hybrid powertrain. And even if Ford or GM came up with a good idea, if it means opening a new plant or refitting an old one, that idea is going to get shelved.

Space travel follows a similar trend. The US space program started with the simple goal of getting men into space (Mercury), followed that with a series of missions to figure out the techniques to get to the moon (Gemini), and finally ended with the moon missions (Apollo). SpaceX is doing something similar, too. Each of their launches is building up the technology and experience to eventually take astronauts to the ISS, and they're working on other technologies to limit cost while they're at it.

It's all very cool, but again, it isn't just the money. Every big advancement is going to bring a series of new problems to solve, and it takes a few iterations to move from "amazing prototype" to "real product."

derFunkenstein wrote:

ssidbroadcast wrote:

Then don't show me the exact position of the driver. Show me a blinking "district" of the driver's approximate location (ala Final Fight,) and an approximate delivery time within 60 minutes of accuracy. Not a big ask!

That's quite a bit to ask, honestly. I'd settle for a driver that actually knocks on the door instead of just leaving a tag and driving away.

Yeah, I agree -- doing a 60 minute window is probably a lot more difficult than it sounds. To make that work accurately you would need to know ahead of time the route the driver intends to take for the rest of the day, time and length of any breaks, traffic/weather conditions, and any other factors which could affect how long stops take (Do they need to use a hand truck to haul multiple loads of heavy packages into a business? Are they making any pickups?), etc.

Get it wrong and miss the window, and people will be more pissed off than they would've been if you hadn't given them a delivery window in the first place. Inaccurate information is worse than no information!

I have to agree with the file system issue. I'm genuinely confused about why Microsoft has not come up with a better file system, especially for server use. Doesn't "bitrot" become more porbable when you have more bits packed even closer together?

Not only that, F2FS is just getting off the ground as SSD's are becoming ever more popular. Wouldn't an SSD work better, faster if it did not have to compensate for the legacy file systems?

just brew it! wrote:

The most common type of application crash is caused by an invalid memory refefence, where the application has attempted to access a non-existent memory location or execute non-existent code. This is actually detected at the virtual memory manager level (down in the OS), not by the application. The OS has no idea what the application was trying to do; the *only* sensible thing for the OS to do in this situation is to terminate the application.

Actually in Unix-land the OS just delivers a segmentation violation signal (SIGSEGV). The OS default for this signal is for the application to exit, but you can install a signal handler for this signal and do other stuff.

In what I work on we have hardware doing DMA so need to shut off the hardware DMA before exiting, otherwise the memory would get freed back to the OS and then potentially handed out to something else that could suddenly find hardware scribbling over what the application was storing in the memory. After we shut off the DMA then we exit as per normal with the proper core dump so we have some hope of debugging and fixing what went wrong. Of course hilarity can ensue when your SIGSEGV handler also SIGSEGVs.

Losergamer04 wrote:

I have to agree with the file system issue. I'm genuinely confused about why Microsoft has not come up with a better file system, especially for server use. Doesn't "bitrot" become more porbable when you have more bits packed even closer together?

Not only that, F2FS is just getting off the ground as SSD's are becoming ever more popular. Wouldn't an SSD work better, faster if it did not have to compensate for the legacy file systems?

I think the error rate per bit stored has remained roughly constant; as the bits get packed closer together, drives use more sophisticated ECC codes to compensate. The problem is, we now store so much more data than we did just a few years ago that the probability of an undetected error occurring has gone way up in absolute terms.

I agree, we need better file systems. But there's a tradeoff too -- maintaining that redundancy reduces performance and takes up additional space on the storage device.

notfred wrote:

Actually in Unix-land the OS just delivers a segmentation violation signal (SIGSEGV). The OS default for this signal is for the application to exit, but you can install a signal handler for this signal and do other stuff.

In what I work on we have hardware doing DMA so need to shut off the hardware DMA before exiting, otherwise the memory would get freed back to the OS and then potentially handed out to something else that could suddenly find hardware scribbling over what the application was storing in the memory. After we shut off the DMA then we exit as per normal with the proper core dump so we have some hope of debugging and fixing what went wrong. Of course hilarity can ensue when your SIGSEGV handler also SIGSEGVs.

Ahh, my bad.

Yes, a SIGSEGV handler can be used to perform critical cleanup actions in the event of a fault.

However, it is still rather difficult (impossible?) to figure out what the application was doing to the extent necessary to recover gracefully. Internal data structures may be in an inconsistent state, the faulting thread may have been holding on to semaphores and/or large allocations of heap space, etc... and the potential set of pitfalls will be different for every application.

I suspect that in the vast majority of cases, the only sensible "recovery" action to a SIGSEGV is to perform any critical cleanup actions (like the one you described), display an error message, and restart the entire application. If the application also does checkpointing, this is actually not too bad.

Why does there have to be any gas in the big new drives at all....it is just freaking stupidity plain and simple. If you want to remove all air drag from any trace gas make sure the drive is air tight and build it in a vacuum. A vacuum is the most idea environment since they is zero air resistance.
Maybe I am wrong and they might need a little bit but it just does not seem logical. I believe a Vacuum is the answer...no wasted gasses etc. I think the benefits far out weigh the slight cost of manufacturing environment costs even though the vacuum could be added after the drive is built as a cheaper but possibly slower capacity...last I checked I do not own any factories.

vargis14 wrote:

Why does there have to be any gas in the big new drives at all....it is just freaking stupidity plain and simple. If you want to remove all air drag from any trace gas make sure the drive is air tight and build it in a vacuum. A vacuum is the most idea environment since they is zero air resistance.
Maybe I am wrong and they might need a little bit but it just does not seem logical. I believe a Vacuum is the answer...no wasted gasses etc. I think the benefits far out weigh the slight cost of manufacturing environment costs even though the vacuum could be added after the drive is built as a cheaper but possibly slower capacity...last I checked I do not own any factories.

HDD tech relies on an air (or helium, but could be any gas) bearing to keep the heads from directly contacting the platters. The rapidly spinning platters drag a thin layer of air around with them, and it is this air current that the heads ride on.

It is an elegant (and inexpensive) solution to precisely maintaining the microscopic (just a couple of nanometers) flying height of the head above the media.

just brew it! wrote:

...the microscopic (just a couple of nanometers)...

Wouldn't that make it nanoscopic?

Igor_Kavinski wrote:

Here are some things that PISS me off:

1) This is 2014 and we still have hard drives that might silently lose data AKA bitrot. We have ZFS to save us but it isn't a mainstream filesystem (Not on Windoze). ReFS: not ready for prime time. And it needs the slow abomination that is Storage Spaces to recover from bitrot errors. I ask you: how hard could it be to design a filesystem that maintains parity bits so that bitrot errors never occur? Why aren't the hard drive makers designing RELIABLE hard drives? They could design one and call it the Infallible Edition or something and put a premium on it. If it's better at retaining data than the current sucky drives, people WILL pay. EYE will PAY.

While this is a real issue in that it exists, it really isn't as bad as you're making it out to be. Hard drives are amazingly reliable for what they are doing. And frankly, the folks who actually need to worry about this issue just find it easier to include redundancy and a good backup plan, so there really isn't any demand for a "fix" to it.

2) I still can't buy a Helium sealed 6TB hard drive for under $300! What's taking so long????????

Physics. Right now 6TB is about the limit of a standard HDD because the data density even at that point is pretty incredible; at least for anything that's cost-effective to produce.

3) Why aren't today's operating systems more like hypervisors???? I should be able to freeze any process, save its state to disk and resume it whenever I want.

Because there isn't a demand for it. The ability to freeze a virtual machine in place is a real need; the ability to freeze a program in place just isn't a thing that the market is clamoring for yet.

4) Why aren't applications more intelligent? Instead of throwing up errors and shutting down when they encounter sudden failure of some sort, why don't they offer to work AROUND the failure condition? Give us options! Don't just bloody blow up in our faces, taking our data and sanity with you :S

Because in most cases when a program crashes it's because of an attempt to read or write at an invalid memory reference. At that point, the OS is actually what kills it; which is appropriate since the OS has no way of knowing what the app was trying to do. Anything else would be security risk.

5) What is up with hybrid drives? There is NO competition in this space. Just Seagate out there with their paltry 8GB of MLC in their hybrids. How long is that gonna last? There should be at least 32GB of high endurance MLC in a hybrid. Sell it at a premium dammit but SELL IT!

Frankly, I think the hybrid drive movement died the moment SSDs became cost effective. People who want speed will go with SSD; people who want space will go with HDDs (or multiple SSDs once the price gets even lower) or some variation of the two. The need for a single drive that combined the two is really predicated by the "speed" side being too expensive; and that isn't the case anymore.