1. I have linux set to autoreboot on panic - and this usually works. This is a mythtv box and I don't want it to sit around not recording stuff for three days as a result of a panic. 2. The console may not display the error since I'm running nvidia X11 drivers. 3. Even if it did display the error, it isn't like I'm going to copy the dump to paper or take a photo of the screen...
I can't find a way to log panics in any way - short of putting the console on some device that would independently support logging. Granted, I can see why the kernel devs wouldn't want to try to write to disk after a panic - who knows what would happen. Then again, maybe there should be some facility to allocate a tiny partition or something to hold kernel core dumps that would be safer to write to if something went wrong...
That's ok - when you have a transmission line carrying 4000 amps of power and the resistance ramps up from 0 ohms/cm to 10000 ohms/cm due to the pressure drop the chemical reactivity of the conductor will probably seem like small potatoes...
Actually - is there really any difference between this and allowing a filesystem to have any amount of delay before syncing the write cache?
Right now linux will hold buffers in memory to try to minimize disk IO. If you could tell it to try to cache as much of a filesystem as possible and take as long as it wants to flush buffers then you'd essentially have this.
They talk about power failures as a potential problem. That doesn't worry me so much - just buy a UPS (a trivial expense if you're going to waste this much RAM). I'd be more concerned about kernel panics. I've had about one per week for the last few week ever since upgrading to avoid the vmsplice bug. I'm sure my array of somewhat-atypical hardware doesn't help, but neither does the fact that linux no longer has a real development branch...
I think it depends on your needs. If the access is read-only and the data isn't sensitive then the embedded string isn't a problem.
I'd say that in my experience users actually having accounts on database servers is pretty uncommon. Most applications just connect to the database using an obfuscated password, or they have a business-logic tier that does the data manipulation.
I agree completely that single-user database accounts are far more secure, but they can be a lot more difficult to maintain and as a result they don't get used much.
I get a chuckle at work. For similar reasons there is a lot of encouragement to avoid conducting business over email in favor of meetings which offers the personal touch (and which aren't discoverable).
On the other hand, you also get all kinds of communications at work pointing out the benefits of working in a virtual environment where tools such as email and document managements systems can help reduce the need for in-person meetings (often with the cost of travel).
And of course it seems like everybody is already in meetings all day every day, so the only way to actually get anything done is via email unless you're an executive that can just tell everybody to just skip all their other meetings.
"If it can't be sourced, it shouldn't be on Wikipedia." - that's where "deletionism" comes from. If you can't prove the accuracy of a page, should it exist?
Except that they've very picky regarding sources. Give me a source for the fact that Slashdot posted a discussion on the "Battle for Wikipedia's Soul". Nope - Slashdot is not an acceptable source - and yet it is obviously factually true that this happened, and any idiot could look it up.
Uh, an article that gets deleted and resurrected a dozen times? Shouldn't that indicate that there is something fundamentally wrong with the process? Maybe we need a three-delete rule - or better still a one-delete rule? If an article gets resurrected it never gets deleted again for x years?
Suppose I wanted to write up an article on some chemical intermediate and the various methods by which it is produced and used industrially, and other useful information. The article might not be of interest outside the chemical industry, but it would be informative to anybody who happens to encounter the topic and wants to quickly learn about it.
Would that be considered non-notable and fodder for deletion?
I can certainly see the reason for debate when regarding detail on works of fiction (although even then I don't see the harm in keeping that stuff - the fanatics who maintain it will likely police it more thoroughly than most editors review articles on even important topics). I'm concerned though that Wikipedia will turn into a repository of common knowledge and not expand beyond this...
Actually, it will go over with quite a few scientists. The ones arrogant enough to think that they're far more likely to get the prize than any of the 10,000 other scientists graduating in the same year as them...:)
Don't get me wrong - I'm not saying all scientists are like this (I have a science degree myself), but it is a VERY common attitude among the academic elite. When I went from a small undergrade college into a large university for grad school I noticed a big difference in the teaching. The latter emphasized names quite a bit - any time you talked about anything you talked about who did the work. The former really only mentioned names if they were of long-term historical significance. It wasn't just one professor in either institution either. My conclusion - the large university had a more elite atmosphere, and the who's were important since their goal was to be one of the who's that the next generation of scientists would need to know about. The small college really just taught the material, and did research on small projects that didn't have a lot of funding and which weren't really in competition with the academic elite. They knew they weren't the who's, and it didn't really bother them.
Then the academic elite wonders why nobody in the general public cares to listen to them. Sure, part of it is that they're less in tune with some concepts that really do matter, but part of it is that they really don't care for the attitude. In fact, many in the public would rather see the cure for cancer left undiscovered than to have to listen to somebody lord it over them that they were the one to discover it...:)
I don't think that would happen on a grand scale. Set-top boxes won't have that feature, and any that do have it will be sued into oblivion and excluded from distribution channels like the local walmart.
You could do something like this today with DVRs, and yet it doesn't happen - for precisely these reasons.
The TV execs don't need to keep EVERYBODY from skipping ads. If they get 95% of the public to watch the ads (or even 50%) they can make a fortune. Most people won't bother setting up all kinds of 3rd party software to hack their media player.
And I'm not proposing that the polished interface prevent fast-forwarding through ads. If you make the ads too obnoxious then people will find the adblock equivalent and your ads will go poof entirely.
Just look at the lessons of google - they use simple text-based ads and make a fortune. Nobody bothers to block them because #1 - it isn't easy, and #2 - it isn't worth it since you aren't chasing flash widgets all over your webpage while trying to read an article.
The same way they do it on TV. Survey a sample of households and generate ratings and charge accordingly.
How do you decide what your house is worth when you sell it? Simple - you offer a price and haggle until it is worked out. Happens every day in ad agencies worldwide...
The "back door"is being paid for by ads. Record all you want. The question is, can content producers survive in a world hostile to any means of them recouping their costs?
Yes, and the online content would be as well. They're already surviving in the world you describe - you can get most shows today ad-free, and yet almost nobody does. Oh sure, the average slashdotter might, but I'm talking about the other 99.999% of folks who have money to spend on advertised products.
Right. Much like the NYT distributing their content for the price of signing up, and see how they're taking over the market.
Uh, the online news market is dominated by probably 3-4 companies (I'm talking about the content and the ads - not the portal people visit through). To the extent that they're losing out it is to companies like google who are doing exactly what I'm suggesting the TV networks should do. All of them were traditional news networks before the internet came along. I don't see your point. No one network would beat out all its peers by doing online - but they could make a lot more money this way.
Apple TV.
Uh, what will Apple TV do? Make it easy for people to download TV shows with random filenames posted to random distribution networks by random people? Easier than obtaining the TV from a couple of TV networks distributing shows via standardized protocols over big pipes with lots of infrastructure behind them? I'm sure the networks would give Apple a cut for every referral - the button to watch Battlestar Galactica from the official sources will be bold and on page 1, and the option to configure browsing through random files on TPB will be buried on configuration page 12...
Yeah right! (linked to TPB)
Ok, go ahead and schedule 10 TV shows to auto-download all episodes from TPB so that your 80-year-old grandmother can just click on the show they want and watch it on their TV using a remote control (not a keyboard). Oh wait - the 10 shows don't have any metadata, and the filenames aren't consistent, and a few are posts by guys who didn't bother to seed.
Sure, TPB works, but not well. And it won't have the Gardening special that aired last night or anything not of interest to geeks (who make up all of 1% of the population).
And TPB exists now, and for whatever reason 99% of everybody doesn't use it. Maybe everybody you know does, but most people don't. So this isn't a new threat. And going online will probably actually help to combat it, as opposed to networks sticking their heads in the sand.
They have to deal with that reality anyway. DVRs are owned by something like 20% of the viewing public. And they're probably owned by 90% of the kinds of folks who spend lots of money on the kinds of stuff advertised on TV.
I just don't see the downside to online distribution. All the "negatives" associated with it are already here today. Networks have to deal with that stuff already. So why not at least capture some of the upside of the online world?
Uh, I stated that the online content would contain ads.
Sure, people could strip them out and make a competing version since there is no DRM. But:
1. How many people would bother to even look for the ad-free version? Sure, most people here would, but that just means TV execs will have to live with losing 0.001% of their profits... 2. Who is going to bother to strip all those ads and redistribute? Sure, maybe for Battlestar Galactica, but most geeks with time to kill don't do this stuff for the other 95% of TV programming. 3. Sure, TPB and all will still be around, but who would use it if they could just click and get the polished version? The network version would have hundreds of seeds most likely on high-bandwidth pipes. And nice fat clients and TV grids and side-content as well.
I don't see why this couldn't work. After all - in theory all those pirate channels exist today, and yet 99% of the viewing public watches it TV by turning on a TV. The online versions would only have more and better content, so why would piracy suddenly be such a threat?
I'm surprised this hasn't already taken off for TV. Here's why:
1. Right now networks can only own one station per market. With HD they can in theory broadcast multiple streams on it, but only a few. With online distribution they could put out as much content as they would like.
2. Right now anybody can record and redistribute the off-the-air content. So, DRM is trying to lock up the front door when the back door is already wide open.
3. Right now due to inefficient distribution schemes shows only run in a local market, creating a huge demand for online content. Typically this content lacks commercials, and is ignored when calculating ratings even if it did.
4. If a TV station made it EASY to download their shows with full commercials they'd take over the market overnight. The big networks could collaborate to make it easy to watch their shows just like watching TV. Who would mess around with nzb files and all that when you could just fire up your online "Tivo" and it has already downloaded everything you're interested in. The polished experience would give them 99% of the market all the time.
5. Sure, in theory somebody could find some way to redistribute their content and strip out all the commercials, but the scale of this task except for a few shows would be hard to match with the level of polish that the networks could deliver. They would still own copyright so they would only need to deal with distributed bands of unpaid volunteers redistributing their work - if anybody tried to organize they could be dealt with in court. The court cases would be stronger since the networks could convine local governments that they are actually genuinely trying to get their content to everyone (right now some countries turn a blind eye to copyright violation since it enables their consumers to get access to TV they wouldn't ever see otherwise).
It seems like the TV execs are missing a huge opportunity that they could just own without issue if they just stepped out and took advantage of it.
The US tends to use elegant designs that just barely work, and then engineer 40 layers of redundancy to keep anything from going wrong.
The USSR tended to use simple designs that are inherently more stable, so that when things do go wrong they're less likely to cause a critical failure.
In the shuttle they have 5 computers so that the chances of the computer going out are minimal. However, if the computers do go out they're probably toast - the shuttle can't just land anywhere and it is pretty touchy with regard to re-entry profile. On the other hand, the Russians could land their capsules with 1950s technology in the driver's seat if they needed to - it doesn't need to land on a runway - so as long as it doesn't come down too steep it will land somewhere. In fact, a capsule in LEO will essentially "land" on its own even if it had no crew or computer just due to air friction eventually de-orbiting it (granted, the chutes might not deploy, but until it hits ground it would otherwise be a normal re-entry).
Both approaches have their pros/cons. The US approach can generally support more bells/whistles, the Russian approach is a lot cheaper and "just works" with the caveat that even a stable design will fail if you are too cheap. Unless you need the bells/whistles the simpler design is the better one...
I'm not quite sure why there is such a drive to put CPU fabs in 3rd world countries. Sure, the labor is cheaper, but my understanding is that the capital costs are measured in hundreds of millions of dollars, or even billions. I'm guessing the manpower to operate such a fab is fairly low. So, where are the savings?
Maybe if you want to dump solvents in the local creek it might make sense, or if there were some crazy tax law you were trying to dodge. However, in the latter case at least I'm sure you can convince some Congressmen to fix the tax code if you put the fab in their state. No getting around dumping the solvents in a civilized nation, but really - do companies like Intel/etc really need to resort to this?
Good point - you could probably land a Soyuz without a working computer if you had to in a pinch (assuming you could manually trigger the engines and have some idea what your speed was). Deorbital burns from LEO aren't nearly as touchy as the Apollo re-entries. And not having to hit a particular point on the earth is a big plus.
The shuttle would be a death trap if you did the re-entry just fine but ended up 100 miles away from the nearest airport with gargantuan runways. I'm not sure how well it would do ditching in an ocean or field. Parachutes are a lot nicer that way - in fact light aircraft are sometimes equipped with huge parachutes for exactly this reason - just pull a lever and you glide down to Earth.
A magic user would have on average 2.5 hps at level 1. A magic user could cast exactly one spell all day. A magic user couldn't wear armor. A goblin with a sword would do an average of 4.5hp damage per hit. Do the math... Even a fighter only had 4.5hp on average - good for one hit (granted, a harder hit with decent armor). You couldn't really play the first few levels because one bad die roll would be the end of you, and the magic users basically just had to hide the whole time and maybe light off a light spell to pass the time.
Cricket is just downright dangerous. I know a guy from the UK who plays often and he has had some really nasty hand injuries from catching the ball incorrectly.
However, it is also entertaining. Particularly when said fellow took time to explain the whole game to an American audience. We found all kinds of new uses for common vocabulary terms like silly, square leg, off, etc...
I agree that access to memory does open avenues of attack. However, the program code would likely also be encrypted.
An electron microscope can reveal the design of a CPU, but not the contents of its cache/registers/etc. And I doubt a running CPU could be put into an EM even if it could.
Any keys that are discovered would likely be unique to that PC - it is unlikely that with TCPM that you could open an entire platform just by getting a single system's private keys. However, if you were quiet about it you could download lots of protected media and de-protect it and distribute it. Once you were exposed, however, any keys you did have would be revoked and you'd have only what you've gotten so far...
The whole system is fundamentally flawed, but they have gone pretty far to make it awfully hard to circumvent. The big gap is that once a given piece of data is cracked, it can be freely copied by all.
Yup - at work we have lots of systems that are subject to various government regulations, and which contain data that could become evidence in a lawsuit.
We take all kinds of care to document everything about these systems and their reliability, and we have retention schedules for everything and we follow them. While in a court case we might attempt to limit the scope of discovery we ultimately would be prepared to defend our data. Otherwise a computer log isn't evidence any more than a piece of paper typed up on a typewriter 10 minutes before the trial.
Looking at the laundry list, I saw one or two items that might have been a little broad, but most of this stuff is directly limited to the scope of the issues at hand and the reliability of the evidence. If there were 25 precedents that this particular software was bulletproof the plaintiffs might get the scope of discovery narrowed down a little further (maybe just to demonstrate that the software is the software that is considered reliable), but as things currently stand I'd be surprised if the judge didn't order the plaintiff to produce the supporting evidence or have their documentation ruled inadmissible (which would pretty-much gut their case).
If you're in the left lane and overtaking traffic in the right lane, you're fine in my book.
If you're in the left lane, and just keeping pace with traffic in the right lane, and there is somebody behind you, you should try to move over for them when safe to do so.
It is just courtesy. I have all the patience in the world for somebody who passes somebody else slower than I might. I get really annoyed when two cars drive in side-by-side formation for 5 miles. As long as you're passing I don't care what your absolute speed is.
It isn't like every byte in memory needs an elaborate set of security permissions, and that we have 500 devices on the bus whose memory access needs change every 15 seconds.
Your network card needs access to the network buffer, your sound card needs access to the sound buffer, your hard drive needs access to the drive buffer, etc. The device driver would just grant access to a few MB of memory to each device and keep its buffers in that space. You might only have a dozen ranges defined in a typical system. It isn't like you'd need to adjust the permissions every time the buffer grows or shrinks a kilobyte - just keep unaddressed space around it and give it plenty of virtual address space to grow/shrink.
Right now we effectively have a table with one entry - allow all. Sure, controlling access down to the byte might be nice to have, but even chopping RAM into a couple of segments would be an improvement over what we have now.
And this wouldn't be a general-purpose CPU or anything like that. It wouldn't have any processing power - it would just get requests in on one side and pass them out the other after a quick logic operation on the address and source.
I'm sure it isn't completely trivial, but I imagine that components in the PC are already doing far more work than this at full bus speeds.
Certainly a compromise would be some kind of memory controller chip that could keep a lookup table of memory access permissions. Devices would authenticate requests on some way and the controller would allow DMA access or not. This wouldn't need to impact the CPU (other than maintaining the table - which probably would be fairly short as most peripherals don't need arbitrary memory access), and a simple controller could be pretty fast.
Slashdot Mirror
Alas - when it happens:
1. I have linux set to autoreboot on panic - and this usually works. This is a mythtv box and I don't want it to sit around not recording stuff for three days as a result of a panic.
2. The console may not display the error since I'm running nvidia X11 drivers.
3. Even if it did display the error, it isn't like I'm going to copy the dump to paper or take a photo of the screen...
I can't find a way to log panics in any way - short of putting the console on some device that would independently support logging. Granted, I can see why the kernel devs wouldn't want to try to write to disk after a panic - who knows what would happen. Then again, maybe there should be some facility to allocate a tiny partition or something to hold kernel core dumps that would be safer to write to if something went wrong...
That's ok - when you have a transmission line carrying 4000 amps of power and the resistance ramps up from 0 ohms/cm to 10000 ohms/cm due to the pressure drop the chemical reactivity of the conductor will probably seem like small potatoes...
Actually - is there really any difference between this and allowing a filesystem to have any amount of delay before syncing the write cache?
Right now linux will hold buffers in memory to try to minimize disk IO. If you could tell it to try to cache as much of a filesystem as possible and take as long as it wants to flush buffers then you'd essentially have this.
They talk about power failures as a potential problem. That doesn't worry me so much - just buy a UPS (a trivial expense if you're going to waste this much RAM). I'd be more concerned about kernel panics. I've had about one per week for the last few week ever since upgrading to avoid the vmsplice bug. I'm sure my array of somewhat-atypical hardware doesn't help, but neither does the fact that linux no longer has a real development branch...
I think it depends on your needs. If the access is read-only and the data isn't sensitive then the embedded string isn't a problem.
I'd say that in my experience users actually having accounts on database servers is pretty uncommon. Most applications just connect to the database using an obfuscated password, or they have a business-logic tier that does the data manipulation.
I agree completely that single-user database accounts are far more secure, but they can be a lot more difficult to maintain and as a result they don't get used much.
I get a chuckle at work. For similar reasons there is a lot of encouragement to avoid conducting business over email in favor of meetings which offers the personal touch (and which aren't discoverable).
:)
On the other hand, you also get all kinds of communications at work pointing out the benefits of working in a virtual environment where tools such as email and document managements systems can help reduce the need for in-person meetings (often with the cost of travel).
And of course it seems like everybody is already in meetings all day every day, so the only way to actually get anything done is via email unless you're an executive that can just tell everybody to just skip all their other meetings.
Gotta love progress!
"If it can't be sourced, it shouldn't be on Wikipedia." - that's where "deletionism" comes from. If you can't prove the accuracy of a page, should it exist?
Except that they've very picky regarding sources. Give me a source for the fact that Slashdot posted a discussion on the "Battle for Wikipedia's Soul". Nope - Slashdot is not an acceptable source - and yet it is obviously factually true that this happened, and any idiot could look it up.
Uh, an article that gets deleted and resurrected a dozen times? Shouldn't that indicate that there is something fundamentally wrong with the process? Maybe we need a three-delete rule - or better still a one-delete rule? If an article gets resurrected it never gets deleted again for x years?
What about trivial non-fictional information?
Suppose I wanted to write up an article on some chemical intermediate and the various methods by which it is produced and used industrially, and other useful information. The article might not be of interest outside the chemical industry, but it would be informative to anybody who happens to encounter the topic and wants to quickly learn about it.
Would that be considered non-notable and fodder for deletion?
I can certainly see the reason for debate when regarding detail on works of fiction (although even then I don't see the harm in keeping that stuff - the fanatics who maintain it will likely police it more thoroughly than most editors review articles on even important topics). I'm concerned though that Wikipedia will turn into a repository of common knowledge and not expand beyond this...
Actually, it will go over with quite a few scientists. The ones arrogant enough to think that they're far more likely to get the prize than any of the 10,000 other scientists graduating in the same year as them... :)
:)
Don't get me wrong - I'm not saying all scientists are like this (I have a science degree myself), but it is a VERY common attitude among the academic elite. When I went from a small undergrade college into a large university for grad school I noticed a big difference in the teaching. The latter emphasized names quite a bit - any time you talked about anything you talked about who did the work. The former really only mentioned names if they were of long-term historical significance. It wasn't just one professor in either institution either. My conclusion - the large university had a more elite atmosphere, and the who's were important since their goal was to be one of the who's that the next generation of scientists would need to know about. The small college really just taught the material, and did research on small projects that didn't have a lot of funding and which weren't really in competition with the academic elite. They knew they weren't the who's, and it didn't really bother them.
Then the academic elite wonders why nobody in the general public cares to listen to them. Sure, part of it is that they're less in tune with some concepts that really do matter, but part of it is that they really don't care for the attitude. In fact, many in the public would rather see the cure for cancer left undiscovered than to have to listen to somebody lord it over them that they were the one to discover it...
I don't think that would happen on a grand scale. Set-top boxes won't have that feature, and any that do have it will be sued into oblivion and excluded from distribution channels like the local walmart.
You could do something like this today with DVRs, and yet it doesn't happen - for precisely these reasons.
The TV execs don't need to keep EVERYBODY from skipping ads. If they get 95% of the public to watch the ads (or even 50%) they can make a fortune. Most people won't bother setting up all kinds of 3rd party software to hack their media player.
And I'm not proposing that the polished interface prevent fast-forwarding through ads. If you make the ads too obnoxious then people will find the adblock equivalent and your ads will go poof entirely.
Just look at the lessons of google - they use simple text-based ads and make a fortune. Nobody bothers to block them because #1 - it isn't easy, and #2 - it isn't worth it since you aren't chasing flash widgets all over your webpage while trying to read an article.
The lesson: don't be too greedy!
The same way they do it on TV. Survey a sample of households and generate ratings and charge accordingly.
How do you decide what your house is worth when you sell it? Simple - you offer a price and haggle until it is worked out. Happens every day in ad agencies worldwide...
The "back door"is being paid for by ads. Record all you want. The question is, can content producers survive in a world hostile to any means of them recouping their costs?
Yes, and the online content would be as well. They're already surviving in the world you describe - you can get most shows today ad-free, and yet almost nobody does. Oh sure, the average slashdotter might, but I'm talking about the other 99.999% of folks who have money to spend on advertised products.
Right. Much like the NYT distributing their content for the price of signing up, and see how they're taking over the market.
Uh, the online news market is dominated by probably 3-4 companies (I'm talking about the content and the ads - not the portal people visit through). To the extent that they're losing out it is to companies like google who are doing exactly what I'm suggesting the TV networks should do. All of them were traditional news networks before the internet came along. I don't see your point. No one network would beat out all its peers by doing online - but they could make a lot more money this way.
Apple TV.
Uh, what will Apple TV do? Make it easy for people to download TV shows with random filenames posted to random distribution networks by random people? Easier than obtaining the TV from a couple of TV networks distributing shows via standardized protocols over big pipes with lots of infrastructure behind them? I'm sure the networks would give Apple a cut for every referral - the button to watch Battlestar Galactica from the official sources will be bold and on page 1, and the option to configure browsing through random files on TPB will be buried on configuration page 12...
Yeah right! (linked to TPB)
Ok, go ahead and schedule 10 TV shows to auto-download all episodes from TPB so that your 80-year-old grandmother can just click on the show they want and watch it on their TV using a remote control (not a keyboard). Oh wait - the 10 shows don't have any metadata, and the filenames aren't consistent, and a few are posts by guys who didn't bother to seed.
Sure, TPB works, but not well. And it won't have the Gardening special that aired last night or anything not of interest to geeks (who make up all of 1% of the population).
And TPB exists now, and for whatever reason 99% of everybody doesn't use it. Maybe everybody you know does, but most people don't. So this isn't a new threat. And going online will probably actually help to combat it, as opposed to networks sticking their heads in the sand.
They have to deal with that reality anyway. DVRs are owned by something like 20% of the viewing public. And they're probably owned by 90% of the kinds of folks who spend lots of money on the kinds of stuff advertised on TV.
I just don't see the downside to online distribution. All the "negatives" associated with it are already here today. Networks have to deal with that stuff already. So why not at least capture some of the upside of the online world?
Uh, I stated that the online content would contain ads.
Sure, people could strip them out and make a competing version since there is no DRM. But:
1. How many people would bother to even look for the ad-free version? Sure, most people here would, but that just means TV execs will have to live with losing 0.001% of their profits...
2. Who is going to bother to strip all those ads and redistribute? Sure, maybe for Battlestar Galactica, but most geeks with time to kill don't do this stuff for the other 95% of TV programming.
3. Sure, TPB and all will still be around, but who would use it if they could just click and get the polished version? The network version would have hundreds of seeds most likely on high-bandwidth pipes. And nice fat clients and TV grids and side-content as well.
I don't see why this couldn't work. After all - in theory all those pirate channels exist today, and yet 99% of the viewing public watches it TV by turning on a TV. The online versions would only have more and better content, so why would piracy suddenly be such a threat?
I'm surprised this hasn't already taken off for TV. Here's why:
1. Right now networks can only own one station per market. With HD they can in theory broadcast multiple streams on it, but only a few. With online distribution they could put out as much content as they would like.
2. Right now anybody can record and redistribute the off-the-air content. So, DRM is trying to lock up the front door when the back door is already wide open.
3. Right now due to inefficient distribution schemes shows only run in a local market, creating a huge demand for online content. Typically this content lacks commercials, and is ignored when calculating ratings even if it did.
4. If a TV station made it EASY to download their shows with full commercials they'd take over the market overnight. The big networks could collaborate to make it easy to watch their shows just like watching TV. Who would mess around with nzb files and all that when you could just fire up your online "Tivo" and it has already downloaded everything you're interested in. The polished experience would give them 99% of the market all the time.
5. Sure, in theory somebody could find some way to redistribute their content and strip out all the commercials, but the scale of this task except for a few shows would be hard to match with the level of polish that the networks could deliver. They would still own copyright so they would only need to deal with distributed bands of unpaid volunteers redistributing their work - if anybody tried to organize they could be dealt with in court. The court cases would be stronger since the networks could convine local governments that they are actually genuinely trying to get their content to everyone (right now some countries turn a blind eye to copyright violation since it enables their consumers to get access to TV they wouldn't ever see otherwise).
It seems like the TV execs are missing a huge opportunity that they could just own without issue if they just stepped out and took advantage of it.
I guess the difference is this:
The US tends to use elegant designs that just barely work, and then engineer 40 layers of redundancy to keep anything from going wrong.
The USSR tended to use simple designs that are inherently more stable, so that when things do go wrong they're less likely to cause a critical failure.
In the shuttle they have 5 computers so that the chances of the computer going out are minimal. However, if the computers do go out they're probably toast - the shuttle can't just land anywhere and it is pretty touchy with regard to re-entry profile. On the other hand, the Russians could land their capsules with 1950s technology in the driver's seat if they needed to - it doesn't need to land on a runway - so as long as it doesn't come down too steep it will land somewhere. In fact, a capsule in LEO will essentially "land" on its own even if it had no crew or computer just due to air friction eventually de-orbiting it (granted, the chutes might not deploy, but until it hits ground it would otherwise be a normal re-entry).
Both approaches have their pros/cons. The US approach can generally support more bells/whistles, the Russian approach is a lot cheaper and "just works" with the caveat that even a stable design will fail if you are too cheap. Unless you need the bells/whistles the simpler design is the better one...
I'm not quite sure why there is such a drive to put CPU fabs in 3rd world countries. Sure, the labor is cheaper, but my understanding is that the capital costs are measured in hundreds of millions of dollars, or even billions. I'm guessing the manpower to operate such a fab is fairly low. So, where are the savings?
Maybe if you want to dump solvents in the local creek it might make sense, or if there were some crazy tax law you were trying to dodge. However, in the latter case at least I'm sure you can convince some Congressmen to fix the tax code if you put the fab in their state. No getting around dumping the solvents in a civilized nation, but really - do companies like Intel/etc really need to resort to this?
Good point - you could probably land a Soyuz without a working computer if you had to in a pinch (assuming you could manually trigger the engines and have some idea what your speed was). Deorbital burns from LEO aren't nearly as touchy as the Apollo re-entries. And not having to hit a particular point on the earth is a big plus.
The shuttle would be a death trap if you did the re-entry just fine but ended up 100 miles away from the nearest airport with gargantuan runways. I'm not sure how well it would do ditching in an ocean or field. Parachutes are a lot nicer that way - in fact light aircraft are sometimes equipped with huge parachutes for exactly this reason - just pull a lever and you glide down to Earth.
I have to agree with you there.
A magic user would have on average 2.5 hps at level 1. A magic user could cast exactly one spell all day. A magic user couldn't wear armor. A goblin with a sword would do an average of 4.5hp damage per hit. Do the math... Even a fighter only had 4.5hp on average - good for one hit (granted, a harder hit with decent armor). You couldn't really play the first few levels because one bad die roll would be the end of you, and the magic users basically just had to hide the whole time and maybe light off a light spell to pass the time.
Cricket is just downright dangerous. I know a guy from the UK who plays often and he has had some really nasty hand injuries from catching the ball incorrectly.
However, it is also entertaining. Particularly when said fellow took time to explain the whole game to an American audience. We found all kinds of new uses for common vocabulary terms like silly, square leg, off, etc...
I agree that access to memory does open avenues of attack. However, the program code would likely also be encrypted.
An electron microscope can reveal the design of a CPU, but not the contents of its cache/registers/etc. And I doubt a running CPU could be put into an EM even if it could.
Any keys that are discovered would likely be unique to that PC - it is unlikely that with TCPM that you could open an entire platform just by getting a single system's private keys. However, if you were quiet about it you could download lots of protected media and de-protect it and distribute it. Once you were exposed, however, any keys you did have would be revoked and you'd have only what you've gotten so far...
The whole system is fundamentally flawed, but they have gone pretty far to make it awfully hard to circumvent. The big gap is that once a given piece of data is cracked, it can be freely copied by all.
Yup - at work we have lots of systems that are subject to various government regulations, and which contain data that could become evidence in a lawsuit.
We take all kinds of care to document everything about these systems and their reliability, and we have retention schedules for everything and we follow them. While in a court case we might attempt to limit the scope of discovery we ultimately would be prepared to defend our data. Otherwise a computer log isn't evidence any more than a piece of paper typed up on a typewriter 10 minutes before the trial.
Looking at the laundry list, I saw one or two items that might have been a little broad, but most of this stuff is directly limited to the scope of the issues at hand and the reliability of the evidence. If there were 25 precedents that this particular software was bulletproof the plaintiffs might get the scope of discovery narrowed down a little further (maybe just to demonstrate that the software is the software that is considered reliable), but as things currently stand I'd be surprised if the judge didn't order the plaintiff to produce the supporting evidence or have their documentation ruled inadmissible (which would pretty-much gut their case).
IANAL though...
If you're in the left lane and overtaking traffic in the right lane, you're fine in my book.
If you're in the left lane, and just keeping pace with traffic in the right lane, and there is somebody behind you, you should try to move over for them when safe to do so.
It is just courtesy. I have all the patience in the world for somebody who passes somebody else slower than I might. I get really annoyed when two cars drive in side-by-side formation for 5 miles. As long as you're passing I don't care what your absolute speed is.
It isn't like every byte in memory needs an elaborate set of security permissions, and that we have 500 devices on the bus whose memory access needs change every 15 seconds.
Your network card needs access to the network buffer, your sound card needs access to the sound buffer, your hard drive needs access to the drive buffer, etc. The device driver would just grant access to a few MB of memory to each device and keep its buffers in that space. You might only have a dozen ranges defined in a typical system. It isn't like you'd need to adjust the permissions every time the buffer grows or shrinks a kilobyte - just keep unaddressed space around it and give it plenty of virtual address space to grow/shrink.
Right now we effectively have a table with one entry - allow all. Sure, controlling access down to the byte might be nice to have, but even chopping RAM into a couple of segments would be an improvement over what we have now.
And this wouldn't be a general-purpose CPU or anything like that. It wouldn't have any processing power - it would just get requests in on one side and pass them out the other after a quick logic operation on the address and source.
I'm sure it isn't completely trivial, but I imagine that components in the PC are already doing far more work than this at full bus speeds.
Certainly a compromise would be some kind of memory controller chip that could keep a lookup table of memory access permissions. Devices would authenticate requests on some way and the controller would allow DMA access or not. This wouldn't need to impact the CPU (other than maintaining the table - which probably would be fairly short as most peripherals don't need arbitrary memory access), and a simple controller could be pretty fast.