Just because a vehicle doesn't have people doesn't really mean that you can ignore risks. One of the recent "failed" launch vehicles (I don't recall if it was a Delta or a Titan) blew up. Ignoring the cost of the vehicle and launch itself, the *payload* that it had been commissioned to launch was worth several *billion* dollars.
Now, if your unmanned vehicles are in any way unreliable, it doesn't take too many of those before customers stop using your services (private industry), or government inquiries shut you down (government service).
Why do I say that? Because it sounds like you're not a licensed electrician. If you were, you probably wouldn't be asking these questions.
Now, what does that matter? While many/most places will let you work on your *own home* without being an electrician, I doubt there are any in the US (and probably most of the industrial world) that will let you work on your employer's electrical system.
So, here's what can happen: 5 years down the road, a fire starts. The fire investigators trace it back to faulty wiring. Woops! The work wasn't done by an electrician, and you didn't have it inspected. Guess who's liable.
Now, in the off-chance that your municipality WILL let you do the work and pass it off, the answer is easy: You bite the bullet and rewire everything. A pain in the butt, but not as large as having a fire - or even as much as tripping circuits to important systems at inopportune moments.
What's up with their CDRW drives? I've been buying Sony CDRW drives for my machines for years, and I've never had a problem with them under any OS, and never had one fail.
You really can't paint Sony with a broad brush. Some of their products are fantastic, while others are just rubbish. You just have to know which are which!
Is 1.26 million of them really that hard? Contrary to the belief of people who have never been west of the Mississipi river, there is a LOT of land in the country.
Shoot, look at it as a boon to the farmers. Have them put them throughout their crop fields (because they're tall, the footprint in the field is miniscule). Not only would it provide clean power, the farmers could make a few bucks off of it, and the way you hear farmers complaining, you'd think that a few extra bucks would be a good thing for them.
Of course, then maybe you'd hear from the out-of-work coal miners. : )
1) Specialized hardware will run at least an order of magnitude faster than generic hardware. What a 300 MHz GeForce FX will render at 30-60 FPS, a 3 GHz CPU couldn't render at 3 FPS, no matter how much you tweak it. Really.
2) GPU rendering isn't software at all. The great, overwhelming bulk of the work is all done in hardware on the GPU. Perhaps you should look into GPU design a little bit.
As for tweaking for an individual platform, id already does that on their games - in the parts of code where performance is critical and takes non-trivial time, they often use assembler code tailored for the type of CPU that the game has detected. And guess what! D3 also detects the video card used, and uses different rendering paths based on that.
It's really not like you can take a magic wand and "tweak" your program and get tons of performance out of it. The real reason why consoles perform so well even with anemic processers is because they mate those anemic processers with very powerful video processers, which means hardware rendering. They're using their transistor count where it really matters, for the video.
If you really think that you can *practically* do these things in software, I'll tell you what: Do it. When you magically find some way to take these games that need high-end video cards and make them render in software, you'll make a killing.
It's not a matter of tweaking, it's simply that (a) the Xbox doesn't have the hardware capabilities in the graphics chip that are *required* for full Doom3 effects, and (b) software rendering, no matter how tweaked, just won't do it.
... but without programmable vertex shaders and the like, you're still not going to get all of the graphics, especially the shadows - which have always been touted as one of the largest parts of the game's experience. Even dropping to 640x480 won't help there. That right there would seem to contradict their statement.
The Xbox version of Doom III will be basically the complete experience that PC gamers will get, according to Hollenshead Aside from the limited amount of RAM in the console there are no technical bottlenecks in converting the PC Doom III to work on the Xbox. As reported earlier, Hollenshead said that developer Vicarious Visions is the primary developer behind the Xbox port.
So... it's going to be the complete experience on the Xbox? You'll get the "complete" experience on something with an outdated CPU, tremendously slow FSB, very limitted memory, and an outdated video card? Sheesh, and all of this time I had heard that you needed fancy new video cards to even enable some of the video features. I guess I wuz lied to! (tongue in cheek)
Know the whole world knows! All of us IT types are really all sitting around in a room filled with exotic dancers, pool tables, video games, and food. We don't really do anything. And it all gets so boring that we go LOOKING for systems to give us more work to do!
In reality, 99% of the IT people that I know would practically sell their left arm to have systems which required less of their time.
The guy is talking about bursts of 42,000 hits per day, and talking about it "bringing their system down". Now I could see that on Windows, but not on Linux.
Now, before you think I'm talking out of my posterior orifice, when my company was young and bright, we had a server built on a single 450 MHz Pentium 2, and 256 megs of RAM. It ran both Apache and PostgreSQL. Many of our pages were database-driven, which of course is a much larger load on the server than simple static pages.
That little machine would peak out at around 60,000 hits per day. At that point, it was slow enough to be self-limitting, but there was never any fear (or realization) of having the machine "brought down".
So, still "back in the day", I replaced it with a dual P3/650. That machine would peak out at around 100,000 hits/day (database driven!), without much problem at all. Also, as time goes on, and we develop new apps that make further use of our data, we tend to need more power to generate every page. Even still, we could crank out 40,000 hits per day on what would now be a relatively anemic server.
Now, with 7 front-end web servers and a dedicated DB machine, we crank out 5 million hits/day without problem. And even when our systems have been IMMENSELY overloaded from both legitimate and illegitimate traffic, the systems have still responded, and never once have I ever worried about a machine "going down" from the load. Failed hardware, perhaps, but not the load.
but honestly who besides me cares about(let alone has heard of):
"Back in the day", BeachHead and BeachHead 2 were the games for me. I went in the store, and saw a new game called BeachHead. I almost flipped my wig, I was so excited. I grabbed the box, turned it over to look at the screenshots...... it had absolutely nothing to do with the original. I never thought that a mere memory of a C64 video game from 20 years ago could leave me with such an empty feeling.
Q: How can you tell an extroverted computer geek from an introverted computer geek?
A: The introverted computer geek will look at his shoes while he talks to you. The extroverted computer geek will look at your shoes while he talks to you.
There are lots of stories about how company XYZ is using Linux. However, this one has potential for a *real* benefit to Linux. Why?
Well, when Oracle, with cash flowing out of it's orifices, finds something in Linux that they'd really like to have improved, they have plenty of resources to improve it, which benefits Linux.
If some small, third-world government adopts Linux, that's great. But they're still not going to give anywhere near as much back to Linux as companies like IBM have been able to. Oracle stands a pretty good chance of giving quite a bit back as well, and I think Linux will be much better off for it.
Actually, I don't have a link handy, I saw it on a documentary show. While they wouldn't show the actual computing center, they showed the coolant pipes going to the radiators. Normally, hearing "they showed us the coolant pipes!" would be a pretty big let-down, but these pipes were pretty impressive in and of itself.
Also, the NSA doesn't just walk down to Fry's and buy a couple of P4's when they need more computing power, they have their own foundry facilities in order to produce specialized chips, and when you're producing customized hardware to attack a problem, you get a LOT more done than trying to do it on a general-purpose CPU.
Actually, with good algorithms, knowing the plaintext doesn't really help you. Witness the RC5-64 contest, where the plaintext was known, and with over 300,000 people working in parallel, it still took nearly 5 years!
However, the idea of keeping government supercomputers busy for weeks is a bit far-fetched. From what I understand, the NSA's "deep thought" supercomputer can crack an RC5 key in a few minutes. There's something to be said about a computer that needs 20,000 tons of coolant!
Whether it's a public-key algorithm or a symmetric cypher, either way, there is a key involved, and protecting that key is, as you seem to realize, paramount.
My argument is that finding a way to retrieve the key in question will prove to be rather trivial, certainly more trivial than trying to brute-force the encryption.
Encryption algorythms are sufficiently advanced that key management is the real issue: Trying to brute-force it can be very difficult, but finding out the private key (which makes decryption trivial) can often be relatively easy. So, even if they used reasonably strong encryption, chances are that they won't succeed at protecting the private keys.
However, I suspect that their encryption isn't really all that strong. Doing strong encryption at speeds necessary to sustain IDE transfers (up to 50 megabytes/second *per drive*) is fairly serious stuff, especially if you want to be able to do it at sufficiently low latencies. Hardware-encryption boards that truly do strong encryption at much slower speeds than that are pretty pricey, usually at least four figures.
You're *exactly* the type of person who will make the best tester, precisely because it *will* crash for you. It's the tiny bugs that only show up under bizarre/rare combinations of features and usage that can be the most pesky.
Now, I'm not saying you should run a non-stable version on your server, but what about setting up a spare machine simply to replicate what's being done on your server?
Not only will it help out kernel development, it will also mean that you will get a stable kernel for your server sooner.
I would have thought that the more people that go through the process that I have over the last few weeks the better.
I'm not trying to be mean. Really. But that's not better. What's better is when more people search out and read the docs before they whip out the compiler and become part of the problem. It's really a great thing that you understand more about SMTP, but you could have gone about it in a much better way.
I don't think that the "people like you" was unfair. Harsh, but not unfair. I am glad for you, however, for being humble enough to admit and learn from your mistakes. It sounds like you'll do a much better job of things from now on.
I've turned down some attractive deals with large, key providers for *years* precisely because they blatantly tolerated spammers.
As far as I'm concerned, if the netblock in question was blacklisted with the RBLS that are taken more seriously, it was precisely because the provider didn't take any sort of action to contain or prevent spamming. And if you sign up with a provider like that, well, don't come crying to me when you're affected, too.
All the time that went by before they knew they were on the blackhole list nearly led them to bankrupty.
It takes less than five minutes to see if you're on the major blacklists, and any administrator who doesn't do it on a regular basis simply isn't worth his pay. I certainly don't have any sympathy for them.
I also have a hard time believing that they simply went about their business for that long without realizing what was going on. How brain-dead do you have to be to realize that a particular person never responds to your email? How long does it take you to realize that SEVERAL people never respond to your email? And for the email problem to truly cause bankruptcy, you're talking about some very important email: The kind that you don't just send and forget. If my users think that someone isn't getting their email, you bet I'll hear about it. And you bet I'll track down the reason.
Really, your description of them makes them sound completely incompetent. For the sake of those involved (and the rest of the world), I hope that's not so.
I'm the first in line to point out that DJB has an ego the size of a larger continent, and has one of the most severe lack of personal skills I've ever had the displeasure to witness.
I've also dropped enough of my pride to look past that, and see that he has written some astoundingly fine code.
You can argue about his personal skills, egomania, and other deficiencies all you like but you just can't argue with the fact that qmail works, and works awfully well. I can't honestly say that I've never had the thought to use a different MTA just because of my disagreement with so much about Dan. But every time, when looking at the technical aspects, I've realized that I'd just be cutting off my nose to spite my face.
AOL annoys me the most, they block ranges of addresses that are dynamically allocated by ISPs and as a result I can't mail any AOL users. That's probably no big deal, I just feel descriminated against.
AOL certainly aren't the only ones. I do it. And I'll tell you why: Exactly because of people like you.
For ever *legitimate* email message that comes from a dial-up IP address, I (honestly) get 10,000 + that are spam.
As long as there are (a) Distribution-installed MTA's that are open relays by default, and (b) people that won't read the documentation, nothing's likely to change.
It's nothing personal, really. It's just that open relays on cable modems make up such an incredibly significant portion of the spam that I get, it's an obvious decision.
There must be scope for a simple "Setting up your own mail server" FAQ.
That's odd, I thought there were plenty of docs on how to set up your own mail server. I know that I've never had any trouble when I looked for them.
In Switzerland, pretty much *everyone* has a firearm. Now, think to yourself: Are you going to cause trouble in a society like that? You certainly wouldn't think of sticking up that cafe, now would you?
Now, move your thoughts to America. The gun laws serve only to take guns away from honest citizens, while doing very little to keep them away from criminals. Think about it: You're a criminal, you've got your gun. You know that the honest folks don't. Now how scared are you of sticking up a cafe?
There are still places where guns are common-place. Guess what! Things go along without problems.
I've also lived in countries where guns were extremely difficult to come by - if the laws didn't stop you, the economics of the situation probably would. Guess what! The murder rate was astronomical compared to the United States. Serial killings, mass killings, murder/suicides, family dispute killings, they all happened - and they happened a lot. Just because a gun isn't available doesn't change a person's predisposition to violence.
Slashdot Mirror
Just because a vehicle doesn't have people doesn't really mean that you can ignore risks. One of the recent "failed" launch vehicles (I don't recall if it was a Delta or a Titan) blew up. Ignoring the cost of the vehicle and launch itself, the *payload* that it had been commissioned to launch was worth several *billion* dollars.
Now, if your unmanned vehicles are in any way unreliable, it doesn't take too many of those before customers stop using your services (private industry), or government inquiries shut you down (government service).
steve
600 volts is fine, but give a 9+ KV source a try. You'd be amazed at the things arcs will go through, and (usually) light on fire. : )
steve
Why do I say that? Because it sounds like you're not a licensed electrician. If you were, you probably wouldn't be asking these questions.
Now, what does that matter? While many/most places will let you work on your *own home* without being an electrician, I doubt there are any in the US (and probably most of the industrial world) that will let you work on your employer's electrical system.
So, here's what can happen: 5 years down the road, a fire starts. The fire investigators trace it back to faulty wiring. Woops! The work wasn't done by an electrician, and you didn't have it inspected. Guess who's liable.
Now, in the off-chance that your municipality WILL let you do the work and pass it off, the answer is easy: You bite the bullet and rewire everything. A pain in the butt, but not as large as having a fire - or even as much as tripping circuits to important systems at inopportune moments.
steve
What's up with their CDRW drives? I've been buying Sony CDRW drives for my machines for years, and I've never had a problem with them under any OS, and never had one fail.
You really can't paint Sony with a broad brush. Some of their products are fantastic, while others are just rubbish. You just have to know which are which!
steve
Is 1.26 million of them really that hard? Contrary to the belief of people who have never been west of the Mississipi river, there is a LOT of land in the country.
Shoot, look at it as a boon to the farmers. Have them put them throughout their crop fields (because they're tall, the footprint in the field is miniscule). Not only would it provide clean power, the farmers could make a few bucks off of it, and the way you hear farmers complaining, you'd think that a few extra bucks would be a good thing for them.
Of course, then maybe you'd hear from the out-of-work coal miners. : )
steve
1) Specialized hardware will run at least an order of magnitude faster than generic hardware. What a 300 MHz GeForce FX will render at 30-60 FPS, a 3 GHz CPU couldn't render at 3 FPS, no matter how much you tweak it. Really.
2) GPU rendering isn't software at all. The great, overwhelming bulk of the work is all done in hardware on the GPU. Perhaps you should look into GPU design a little bit.
As for tweaking for an individual platform, id already does that on their games - in the parts of code where performance is critical and takes non-trivial time, they often use assembler code tailored for the type of CPU that the game has detected. And guess what! D3 also detects the video card used, and uses different rendering paths based on that.
It's really not like you can take a magic wand and "tweak" your program and get tons of performance out of it. The real reason why consoles perform so well even with anemic processers is because they mate those anemic processers with very powerful video processers, which means hardware rendering. They're using their transistor count where it really matters, for the video.
If you really think that you can *practically* do these things in software, I'll tell you what: Do it. When you magically find some way to take these games that need high-end video cards and make them render in software, you'll make a killing.
steve
It's not a matter of tweaking, it's simply that (a) the Xbox doesn't have the hardware capabilities in the graphics chip that are *required* for full Doom3 effects, and (b) software rendering, no matter how tweaked, just won't do it.
steve
... but without programmable vertex shaders and the like, you're still not going to get all of the graphics, especially the shadows - which have always been touted as one of the largest parts of the game's experience. Even dropping to 640x480 won't help there. That right there would seem to contradict their statement.
steve
Here's a quote:
The Xbox version of Doom III will be basically the complete experience that PC gamers will get, according to Hollenshead Aside from the limited amount of RAM in the console there are no technical bottlenecks in converting the PC Doom III to work on the Xbox. As reported earlier, Hollenshead said that developer Vicarious Visions is the primary developer behind the Xbox port.
So... it's going to be the complete experience on the Xbox? You'll get the "complete" experience on something with an outdated CPU, tremendously slow FSB, very limitted memory, and an outdated video card? Sheesh, and all of this time I had heard that you needed fancy new video cards to even enable some of the video features. I guess I wuz lied to! (tongue in cheek)
steve
Know the whole world knows! All of us IT types are really all sitting around in a room filled with exotic dancers, pool tables, video games, and food. We don't really do anything. And it all gets so boring that we go LOOKING for systems to give us more work to do!
In reality, 99% of the IT people that I know would practically sell their left arm to have systems which required less of their time.
steve
The guy is talking about bursts of 42,000 hits per day, and talking about it "bringing their system down". Now I could see that on Windows, but not on Linux.
Now, before you think I'm talking out of my posterior orifice, when my company was young and bright, we had a server built on a single 450 MHz Pentium 2, and 256 megs of RAM. It ran both Apache and PostgreSQL. Many of our pages were database-driven, which of course is a much larger load on the server than simple static pages.
That little machine would peak out at around 60,000 hits per day. At that point, it was slow enough to be self-limitting, but there was never any fear (or realization) of having the machine "brought down".
So, still "back in the day", I replaced it with a dual P3/650. That machine would peak out at around 100,000 hits/day (database driven!), without much problem at all. Also, as time goes on, and we develop new apps that make further use of our data, we tend to need more power to generate every page. Even still, we could crank out 40,000 hits per day on what would now be a relatively anemic server.
Now, with 7 front-end web servers and a dedicated DB machine, we crank out 5 million hits/day without problem. And even when our systems have been IMMENSELY overloaded from both legitimate and illegitimate traffic, the systems have still responded, and never once have I ever worried about a machine "going down" from the load. Failed hardware, perhaps, but not the load.
steve
but honestly who besides me cares about(let alone has heard of):
... it had absolutely nothing to do with the original. I never thought that a mere memory of a C64 video game from 20 years ago could leave me with such an empty feeling.
"Back in the day", BeachHead and BeachHead 2 were the games for me. I went in the store, and saw a new game called BeachHead. I almost flipped my wig, I was so excited. I grabbed the box, turned it over to look at the screenshots...
steve
Q: How can you tell an extroverted computer geek from an introverted computer geek?
A: The introverted computer geek will look at his shoes while he talks to you. The extroverted computer geek will look at your shoes while he talks to you.
There are lots of stories about how company XYZ is using Linux. However, this one has potential for a *real* benefit to Linux. Why?
Well, when Oracle, with cash flowing out of it's orifices, finds something in Linux that they'd really like to have improved, they have plenty of resources to improve it, which benefits Linux.
If some small, third-world government adopts Linux, that's great. But they're still not going to give anywhere near as much back to Linux as companies like IBM have been able to. Oracle stands a pretty good chance of giving quite a bit back as well, and I think Linux will be much better off for it.
steve
Actually, I don't have a link handy, I saw it on a documentary show. While they wouldn't show the actual computing center, they showed the coolant pipes going to the radiators. Normally, hearing "they showed us the coolant pipes!" would be a pretty big let-down, but these pipes were pretty impressive in and of itself.
Also, the NSA doesn't just walk down to Fry's and buy a couple of P4's when they need more computing power, they have their own foundry facilities in order to produce specialized chips, and when you're producing customized hardware to attack a problem, you get a LOT more done than trying to do it on a general-purpose CPU.
steve
Actually, with good algorithms, knowing the plaintext doesn't really help you. Witness the RC5-64 contest, where the plaintext was known, and with over 300,000 people working in parallel, it still took nearly 5 years!
However, the idea of keeping government supercomputers busy for weeks is a bit far-fetched. From what I understand, the NSA's "deep thought" supercomputer can crack an RC5 key in a few minutes. There's something to be said about a computer that needs 20,000 tons of coolant!
steve
Whether it's a public-key algorithm or a symmetric cypher, either way, there is a key involved, and protecting that key is, as you seem to realize, paramount.
My argument is that finding a way to retrieve the key in question will prove to be rather trivial, certainly more trivial than trying to brute-force the encryption.
steve
Encryption algorythms are sufficiently advanced that key management is the real issue: Trying to brute-force it can be very difficult, but finding out the private key (which makes decryption trivial) can often be relatively easy. So, even if they used reasonably strong encryption, chances are that they won't succeed at protecting the private keys.
However, I suspect that their encryption isn't really all that strong. Doing strong encryption at speeds necessary to sustain IDE transfers (up to 50 megabytes/second *per drive*) is fairly serious stuff, especially if you want to be able to do it at sufficiently low latencies. Hardware-encryption boards that truly do strong encryption at much slower speeds than that are pretty pricey, usually at least four figures.
steve
You're *exactly* the type of person who will make the best tester, precisely because it *will* crash for you. It's the tiny bugs that only show up under bizarre/rare combinations of features and usage that can be the most pesky.
Now, I'm not saying you should run a non-stable version on your server, but what about setting up a spare machine simply to replicate what's being done on your server?
Not only will it help out kernel development, it will also mean that you will get a stable kernel for your server sooner.
steve
I would have thought that the more people that go through the process that I have over the last few weeks the better.
I'm not trying to be mean. Really. But that's not better. What's better is when more people search out and read the docs before they whip out the compiler and become part of the problem. It's really a great thing that you understand more about SMTP, but you could have gone about it in a much better way.
I don't think that the "people like you" was unfair. Harsh, but not unfair. I am glad for you, however, for being humble enough to admit and learn from your mistakes. It sounds like you'll do a much better job of things from now on.
steve
I've turned down some attractive deals with large, key providers for *years* precisely because they blatantly tolerated spammers.
As far as I'm concerned, if the netblock in question was blacklisted with the RBLS that are taken more seriously, it was precisely because the provider didn't take any sort of action to contain or prevent spamming. And if you sign up with a provider like that, well, don't come crying to me when you're affected, too.
All the time that went by before they knew they were on the blackhole list nearly led them to bankrupty.
It takes less than five minutes to see if you're on the major blacklists, and any administrator who doesn't do it on a regular basis simply isn't worth his pay. I certainly don't have any sympathy for them.
I also have a hard time believing that they simply went about their business for that long without realizing what was going on. How brain-dead do you have to be to realize that a particular person never responds to your email? How long does it take you to realize that SEVERAL people never respond to your email? And for the email problem to truly cause bankruptcy, you're talking about some very important email: The kind that you don't just send and forget. If my users think that someone isn't getting their email, you bet I'll hear about it. And you bet I'll track down the reason.
Really, your description of them makes them sound completely incompetent. For the sake of those involved (and the rest of the world), I hope that's not so.
who isn't widely regarded as an obnoxious twit
I'm the first in line to point out that DJB has an ego the size of a larger continent, and has one of the most severe lack of personal skills I've ever had the displeasure to witness.
I've also dropped enough of my pride to look past that, and see that he has written some astoundingly fine code.
You can argue about his personal skills, egomania, and other deficiencies all you like but you just can't argue with the fact that qmail works, and works awfully well. I can't honestly say that I've never had the thought to use a different MTA just because of my disagreement with so much about Dan. But every time, when looking at the technical aspects, I've realized that I'd just be cutting off my nose to spite my face.
steve
AOL annoys me the most, they block ranges of addresses that are dynamically allocated by ISPs and as a result I can't mail any AOL users. That's probably no big deal, I just feel descriminated against.
AOL certainly aren't the only ones. I do it. And I'll tell you why: Exactly because of people like you.
For ever *legitimate* email message that comes from a dial-up IP address, I (honestly) get 10,000 + that are spam.
As long as there are (a) Distribution-installed MTA's that are open relays by default, and (b) people that won't read the documentation, nothing's likely to change.
It's nothing personal, really. It's just that open relays on cable modems make up such an incredibly significant portion of the spam that I get, it's an obvious decision.
There must be scope for a simple "Setting up your own mail server" FAQ.
That's odd, I thought there were plenty of docs on how to set up your own mail server. I know that I've never had any trouble when I looked for them.
steve
The answer is extremely simple, really.
In Switzerland, pretty much *everyone* has a firearm. Now, think to yourself: Are you going to cause trouble in a society like that? You certainly wouldn't think of sticking up that cafe, now would you?
Now, move your thoughts to America. The gun laws serve only to take guns away from honest citizens, while doing very little to keep them away from criminals. Think about it: You're a criminal, you've got your gun. You know that the honest folks don't. Now how scared are you of sticking up a cafe?
There are still places where guns are common-place. Guess what! Things go along without problems.
I've also lived in countries where guns were extremely difficult to come by - if the laws didn't stop you, the economics of the situation probably would. Guess what! The murder rate was astronomical compared to the United States. Serial killings, mass killings, murder/suicides, family dispute killings, they all happened - and they happened a lot. Just because a gun isn't available doesn't change a person's predisposition to violence.
Second, if you paid Linus as much as you paid Cisco, I'm sure he'd be delighted.
I haven't paid Cisco a dime. I bought the routers second-hand, and don't have a service contract.
steve