You've got me there. Well, I've hacked on the kernel, more like hacked it up a bit, but nothing that matters to the rest of the world. I haven't used the inline assmebly features of GCC at all.
That still doesn't stop Intel from implementing those features, so my second statement may yet happen. That would, of course, require it to have an assembler that understands the as(1) syntax.
Doesn't matter who regulates the filling of them, the problem is WHO defines what a wetland is. There is another poster who seems to think that with the EPA, 14 days of continued "wet" area... e.g. ponds, puddles, snowmelt, whatever, makes it a wetland.
While, I don't think that is exactly correct, the current definitions are still absurd, and still used as a way for a small number of people to stop a large number of people from using their own land.
Furthermore, the Corps does NOT, as far as I can tell, regulate wetlands within incorporated areas, and NOTHING prevents other agencies from stopping you, even if the Corps would give you a permit.
I know this because my family has 7 acres of land in Minnesota that we bought 40 years ago, with the idea of subdividing it and selling it in the future.
Now that future is here, and when we went to do that, suddenly it is a wetlands (It is lakeshore property, of course it's fscking wet!), and nobody bothered to inform us over the 40 years that it had become one, nobody gave us an option to appeal its status, and the land is worthless to anyone. You can't even camp there.
The land would be worth hundreds of thousands of dollars we could use it the way we bought it. Right now, it is worth... Nothing.
Of all the ridiculous, stupid things... a wetlands that is artificial, when removed, is an EPA violation? JFC, things have gotten out of hand.
You'd think that the artificial wetlands WOULD be an EPA violation, since it is not supposed to be there and has obviously changed the area dramatically.
Welcome to era where environmentalist whackos can take your property rights if your water pipe has a leak and makes a pond of standing water in your back yard. Suddenly, your property is a wetlands, and we all know you can't damage a wetlands. The poor tadpoles might need to go find a new home.
And the damage to your land equity isn't even reimbursed most places.
This Intel document, also available in PDF indeed states that the Intel compiler will not compile the Linux kernel, but I think the kernel could be modified to not use the GNU extensions that the Intel compiler does not support. I wouldn't be surprised it someone isn't already working on that.
Intel also claims to be interested in improving compatibility with GCC, so I think it's just a matter of time before it will work.
Frankly, Red Hat pays a lot of people to work on open source software, including Alan Cox. Red Hat doesn't necessarily make that much money on other people's work, compared to how much work they do.
Don't deny that they give a lot back to the community, as well.
And, what I meant to say along with that was, he was in the German military long before Hitler was in power. They may have sworn loyalty to Hitler, but it was only because they were sworn already to the physical country and whatever government happens to be in power.
People who joined after Hitler came to power did have a choice, true, but many of them were looking for a paying job when they became soldiers. This was in the Great Depression, you know.
Why did so many German soldiers go to war for Hitler?
Umm, it might have had something to do with their oath to their country.
Almost all soldiers anywhere in the world take an oath stating they will obey orders and protect and serve their country. It's not like an oath means much these days. ("I didn't have sex with that woman Monica Lewinsky") However, sixty years ago, people actually kept their word and their oaths.
In the U.S., the oath is to "support and defend the Constitution" and "to obey the orders of the President." Thus, if the government is overthrown, the military is under no obligation to obey orders from the new government, consequently, if congress threw out the Constitutuon and gave absolute power to anyone, they wouldn't have an automatic loyal, sworn in military. Unlike when Hitler got absolute power in Germany, because their oath was to their physical country, not the specific government.
The video card is probably not putting out RF emissions
Yes it is. All signals of any kind that are not D.C. and have sufficiently fast frequency emit RF, and any kind of switch to on or off (digital) WILL emit RF.
That includes flicking the power switch on ANY device, and the digital signals going across your cable.
At any rate, this isn't the problem with the method described, the problem is the LIGHT from the CRT, since it can be sampled and dupicated.
LCDs do not do this, because they don't scan, so your LCD is safe from this kind of eavesdropping.
it doesn't have to do any "guessing", all it has to do is look for the horizontal vertical blanking periods, which are pretty easy. You can have the horizontal timing after two scan lines, and the vertical timings after two refreshes.
As far as dynamically changing scan rates in software... that won't work, and would probably damage the monitor, if someone managed to do it.
This isn't meant to capture one "screen" of information, it's meant to give you a duplicate, real-time image of what's on the target monitor.
It was poor rendering of dimensional analysis, which really requires mathML or some other way to "pretty print."
That formula is: y Dollars = x MHz * 2 Dollars/MHz... the MHz units cancel, leaving Dollars.
I would have written it like this: used laptops seem to cost $2 per MHz, e.g. 100 MHz * 2 dollars/MHz = 200 dollars
Of course, everyone could have figured it out without the e.g., so it's overcomplicated and verbose anyway. It could be simply written as "$2 per MHz."
Re:And how are they supposed to measure this?
on
More on MPEG4
·
· Score: 2
It's still $7.50 per movie, not $6.00 for a 100-minute one, and $7.50 for a 120-minute one.
Re:And how are they supposed to measure this?
on
More on MPEG4
·
· Score: 1
Ahh, but thet is in the U.K., and that just gives another reason why it was the Right Idea to stage that little rebellion we call the "Revolitionary War."
[taking tongue out of cheek now]
Re:And how are they supposed to measure this?
on
More on MPEG4
·
· Score: 3, Interesting
Ahh, BUT the car's gas mileage is BETTER on freeways than it is on city streets, and in fact, when idling in a traffic jam because the state did not keep their end of the bargain uses a LOT less gas.
Gas tax is more associated with "mileage" as opposed to "time." It even works out with your math, "you use a certain amount of gas per hour, and the size of the vehicle (roughly) determines how much gas you burn", making it fair for larger vehicles which use more gas putting their extra weight per mile.
If they charged for "time", then it would cost loads more. I am okay with paying for actual use, which is mileage driven on the surface of the road, which is covered by the gas tax. If my car is stopped, either by me, or by the fault of the state not adequately designing the road, there is no way I should have to pay it.
I can also support normal mileage-based toll roads for the same way... they are charging for miles driven on the road, not time spent.
Just imagine if they checked the time your car entered the freeway, and then when you got off, two hours later, they charged you for two hours of road use. Sounds okay? Okay, what if that trip, at legal speeds, only should have taken 30 minutes, but the freeway was backed up due to an accident or something?
I bet you'd be so friggin pissed off.
Now, how that relates to the article is, if they charged you for "viewing time", and you say... took a piss during the MPEG-4 movie you were watching, do they refund you? Or do they charge? What if you fall asleep? How do they measure whether you were actually *viewing*?
And how are they supposed to measure this?
on
More on MPEG4
·
· Score: 4, Insightful
How the hell are they going to measure "viewing time"? What if I sneeze and briefly aren't looking at the video, do they charge me for that?
Buncha bullcrap. I'm tired of this crap that tries to wring money out of you for time spent doing something. Subscription software, pay-per-minute viewing/listening, and the like.
What next, the state is going to charge me for every minute I'm on the freeway?
I've got an idea. Let's make a computer that charges me $.02/minute for as long as I'm sitting in front of it.
I just wish more people would get sick of this crap, and write their congressman as I have done. There are too many idiots out there who just miss everything as it goes on by.
I bet they care when they get thier first bill for per-minute charges of movie viewing. By then, it'll be too late.
The graphed output of the solar cell would resemble a rectified sine wave with a D.C. offset though, right?
I'm not an EE, but I've done enough basic electronics to where I think I've got this one.
A few posts up in this thread I was arguing that light output from the incandescent is never zero when it has AC voltage applied, and someone thought I might have been implying "not modulated," and that's what started this.:)
It's times like this I wish I owned an oscilloscope.
I was arguing that the bulb is never in an "off" state, even at the zero points, as even though it is cooling, it never goes to zero output, which is in fact, "off".
I wasn't trying to say they don't modulate... I was going for don't have zero output at any point on the cycle. I figure the light output looks like a rectified sine wave with a D.C. offset, when put on a scope.
This is opposed to LEDs and flourescents, which have light output resembling a square wave, and definite are emitting zero light for part of the cycle.
Wish I had an oscilliscope right now. I've suspected the telltale LEDs on my new external modem... but haven't had a way to check since I currently don't own or have access to a scope.
Because the technology isn't around, does not mean it isn't right around the corner, and does not mean that it doesn't exist elsewhere.
I spent enough time working for the government to know that there is technology out there that YOU don't know exists, and probably never will until someone who isn't a government contractor re-invents it for some civilian use.
I am absolutely positive that now that someone has said this is possible, someone else is working on a way to do it even as we talk about it.
Your class III equipment might be safe for now, but do you think they are safe for long?
I didn't read his paper myself, but I certainly agree that it's possible. One would only need to read the intensity of the light, convert that to a video signal, synced with the rate of the display you are spying on, which is easy because of the verticle blanking period.
Wish I'd thought of it first. I'm looking for a thesis idea....
And today, we are going to learn about math and electricity:
AC current flows in a sine wave. Now, I will assume you know what a sine curve looks like.
At a sine curve's peaks, at pi/2 radians from zero in either direction on the unit circle, the absolute unit is 1. Its zero is at zero.
Now, it is only zero at zero degrees. At all other times it is NOT zero, and thus, current is flowing. On a cycle of pi radians, there are an infinite number of points where current is flowing, and only THREE where it is zero, and "stopped" as you say. Since an incandescent bulb is resistant no matter the volage, and has a slow cooling time, the bulb is infinitely "on" for the complete cycle, because it does not turn "off" during the infinitely small zero points of the curve.
Now, the reason LEDs pulse is because their switching speed is near-instantaneous, and they only flow current in one direction.
Flourescents are similar, but generally more apparent in their flickering because of "threshold voltage", which basically, increases the size of the zero points on the curve, because light output is effectively zero for input voltages less than a certain amount. LEDs have a threshold voltage too, but it's a lot smaller percentage generally, for zero light output.
-- Someone decently technically savvy builds his/her own PC with an AMD chip; You don't explain what exactly the is that you think AMD chips have. Furthermore, your list of steps taken toward system incompatibility ends with: -- Said person goes and manually messes with IRQ settings, thus wreaking havoc on the poor commputer that functioned perfectly before.
Which is the real cause of the problem you are describing.
I just like to take the time to point out that I use an AMD processor, and that the last Intel-based system I ever owned was a P200MMX. My machines (self-built) are ALWAYS reliable, and do not have problems with ANY version of Windows, including ACPI support.
In fact, I usually find that many stability problems are directly related to the quality and condition of the hardware, such as:
Accidentally put a small scratch on the motherboard? Looks okay, probably didn't cut a trace, right? No, but if it got down to the trace, you just created a point of extra impedance in the trace... future stability problem. In fact, even if there is no obvious damage, if you dropped the end of the screwdriver onto the motherboard, it may have caused subtle damage to the circuitry that can show up as stability problems.
I once saw a system come in that had problems, only to find a loose screw under the motherboard.
Simple way to improve a system's stability (physically, and in software): Put in ALL the screws that belong in the case. ALL of the drive mounting screws. ALL hardware mounting screws. Do NOT put in one here and there just to keep things tied down... put them ALL in. Not only do they help anchor hardware and dissipate vibrations from moving parts, they provide a ground path for shielding, and shielding from electrical noise is important. Thumbscrews are fine, and I recommend them.
Another one: Don't "loop" cables that are too long. Always use cables that are the correct length for the application. "Looped" cables create larger magnetic fields than ones that are not. Magnetic fields can induce spurious voltage potential in nearby circuits.
On that subject, keep the cables as far from the surface of the motherboard as possible, for the same reason. Use good quality cables. Also, I've heard of more problems from rounded IDE cables over flat ones.
Tip on RAM: Always use high-quality, name-brand memory, not no-name junk from god-knows-what-fourth-world-country. Memory that is even the slightest out of spec can cause intermittent problems.
Fans and cooling: Where possible, lways install dust filters where air intake occurs. For intake fans on the back of a computer, there are "snap on" filters that can be mounted exterior to the fan. Clean filters regularly, and blow any dust out of the computer periodically.
When installing fans, and multiple placement options are possible, think of a place that gets greatest airflow. Every other fan should be an exhaust fan, not counting the power supply fan. Try to think about air current in a system.
Make sure there is enough cooling for the hard drives, as they can get very hot. My policy is one additional fan for every two drives installed.
I didn't mean to turn this into a class on system design, but that is how you build a rock-stable system. I've built computer systems for myself and for others since my first '286 way back. I DO have experience here.
Slashdot Mirror
You're not a kernel hacker, are you?
You've got me there. Well, I've hacked on the kernel, more like hacked it up a bit, but nothing that matters to the rest of the world. I haven't used the inline assmebly features of GCC at all.
That still doesn't stop Intel from implementing those features, so my second statement may yet happen. That would, of course, require it to have an assembler that understands the as(1) syntax.
Doesn't matter who regulates the filling of them, the problem is WHO defines what a wetland is. There is another poster who seems to think that with the EPA, 14 days of continued "wet" area... e.g. ponds, puddles, snowmelt, whatever, makes it a wetland.
While, I don't think that is exactly correct, the current definitions are still absurd, and still used as a way for a small number of people to stop a large number of people from using their own land.
Furthermore, the Corps does NOT, as far as I can tell, regulate wetlands within incorporated areas, and NOTHING prevents other agencies from stopping you, even if the Corps would give you a permit.
I know this because my family has 7 acres of land in Minnesota that we bought 40 years ago, with the idea of subdividing it and selling it in the future.
Now that future is here, and when we went to do that, suddenly it is a wetlands (It is lakeshore property, of course it's fscking wet!), and nobody bothered to inform us over the 40 years that it had become one, nobody gave us an option to appeal its status, and the land is worthless to anyone. You can't even camp there.
The land would be worth hundreds of thousands of dollars we could use it the way we bought it. Right now, it is worth... Nothing.
My point is, that this is bullshit.
Use hexadecimal for numbered parts. Two decimal digits is enough for enumerating 100 things, two hex digits is enough for 256.
Two, how about subdividing the domain name further: rather than servername.domainname.tld, try servername.subdomainname.domainname.tld.
e.g. web-0ff2.group-02.mycompany.com
Of all the ridiculous, stupid things... a wetlands that is artificial, when removed, is an EPA violation? JFC, things have gotten out of hand.
You'd think that the artificial wetlands WOULD be an EPA violation, since it is not supposed to be there and has obviously changed the area dramatically.
Welcome to era where environmentalist whackos can take your property rights if your water pipe has a leak and makes a pond of standing water in your back yard. Suddenly, your property is a wetlands, and we all know you can't damage a wetlands. The poor tadpoles might need to go find a new home.
And the damage to your land equity isn't even reimbursed most places.
"Turd"
Talk about wrong choice in potential buzzwords. If the first few "production" releases suck, everyone's going to be calling it GNU/Turd.
If only M$ had named Windows something like "Bliss", then we could call it "M$ Piss."
This Intel document, also available in PDF indeed states that the Intel compiler will not compile the Linux kernel, but I think the kernel could be modified to not use the GNU extensions that the Intel compiler does not support. I wouldn't be surprised it someone isn't already working on that.
Intel also claims to be interested in improving compatibility with GCC, so I think it's just a matter of time before it will work.
Frankly, Red Hat pays a lot of people to work on open source software, including Alan Cox. Red Hat doesn't necessarily make that much money on other people's work, compared to how much work they do.
Don't deny that they give a lot back to the community, as well.
And, what I meant to say along with that was, he was in the German military long before Hitler was in power. They may have sworn loyalty to Hitler, but it was only because they were sworn already to the physical country and whatever government happens to be in power.
People who joined after Hitler came to power did have a choice, true, but many of them were looking for a paying job when they became soldiers. This was in the Great Depression, you know.
You know, my great granduncle was General Heinz Guderian.
Why did so many German soldiers go to war for Hitler?
Umm, it might have had something to do with their oath to their country.
Almost all soldiers anywhere in the world take an oath stating they will obey orders and protect and serve their country. It's not like an oath means much these days. ("I didn't have sex with that woman Monica Lewinsky") However, sixty years ago, people actually kept their word and their oaths.
In the U.S., the oath is to "support and defend the Constitution" and "to obey the orders of the President." Thus, if the government is overthrown, the military is under no obligation to obey orders from the new government, consequently, if congress threw out the Constitutuon and gave absolute power to anyone, they wouldn't have an automatic loyal, sworn in military. Unlike when Hitler got absolute power in Germany, because their oath was to their physical country, not the specific government.
pet peeve:
The video card is probably not putting out RF emissions
Yes it is. All signals of any kind that are not D.C. and have sufficiently fast frequency emit RF, and any kind of switch to on or off (digital) WILL emit RF.
That includes flicking the power switch on ANY device, and the digital signals going across your cable.
At any rate, this isn't the problem with the method described, the problem is the LIGHT from the CRT, since it can be sampled and dupicated.
LCDs do not do this, because they don't scan, so your LCD is safe from this kind of eavesdropping.
it doesn't have to do any "guessing", all it has to do is look for the horizontal vertical blanking periods, which are pretty easy. You can have the horizontal timing after two scan lines, and the vertical timings after two refreshes.
As far as dynamically changing scan rates in software... that won't work, and would probably damage the monitor, if someone managed to do it.
This isn't meant to capture one "screen" of information, it's meant to give you a duplicate, real-time image of what's on the target monitor.
CRTs will (probably), eventually be a thing of the past and replaced with somthing that doesn't have a scan timing to be deciphered.
Uh, you mean like, LCDs?
It was poor rendering of dimensional analysis, which really requires mathML or some other way to "pretty print."
That formula is: y Dollars = x MHz * 2 Dollars/MHz... the MHz units cancel, leaving Dollars.
I would have written it like this: used laptops seem to cost $2 per MHz, e.g. 100 MHz * 2 dollars/MHz = 200 dollars
Of course, everyone could have figured it out without the e.g., so it's overcomplicated and verbose anyway. It could be simply written as "$2 per MHz."
It's still $7.50 per movie, not $6.00 for a 100-minute one, and $7.50 for a 120-minute one.
Ahh, but thet is in the U.K., and that just gives another reason why it was the Right Idea to stage that little rebellion we call the "Revolitionary War."
[taking tongue out of cheek now]
Ahh, BUT the car's gas mileage is BETTER on freeways than it is on city streets, and in fact, when idling in a traffic jam because the state did not keep their end of the bargain uses a LOT less gas.
Gas tax is more associated with "mileage" as opposed to "time." It even works out with your math, "you use a certain amount of gas per hour, and the size of the vehicle (roughly) determines how much gas you burn", making it fair for larger vehicles which use more gas putting their extra weight per mile.
If they charged for "time", then it would cost loads more. I am okay with paying for actual use, which is mileage driven on the surface of the road, which is covered by the gas tax. If my car is stopped, either by me, or by the fault of the state not adequately designing the road, there is no way I should have to pay it.
I can also support normal mileage-based toll roads for the same way... they are charging for miles driven on the road, not time spent.
Just imagine if they checked the time your car entered the freeway, and then when you got off, two hours later, they charged you for two hours of road use. Sounds okay? Okay, what if that trip, at legal speeds, only should have taken 30 minutes, but the freeway was backed up due to an accident or something?
I bet you'd be so friggin pissed off.
Now, how that relates to the article is, if they charged you for "viewing time", and you say... took a piss during the MPEG-4 movie you were watching, do they refund you? Or do they charge? What if you fall asleep? How do they measure whether you were actually *viewing*?
How the hell are they going to measure "viewing time"? What if I sneeze and briefly aren't looking at the video, do they charge me for that?
Buncha bullcrap. I'm tired of this crap that tries to wring money out of you for time spent doing something. Subscription software, pay-per-minute viewing/listening, and the like.
What next, the state is going to charge me for every minute I'm on the freeway?
I've got an idea. Let's make a computer that charges me $.02/minute for as long as I'm sitting in front of it.
I just wish more people would get sick of this crap, and write their congressman as I have done. There are too many idiots out there who just miss everything as it goes on by.
I bet they care when they get thier first bill for per-minute charges of movie viewing. By then, it'll be too late.
The graphed output of the solar cell would resemble a rectified sine wave with a D.C. offset though, right?
:)
I'm not an EE, but I've done enough basic electronics to where I think I've got this one.
A few posts up in this thread I was arguing that light output from the incandescent is never zero when it has AC voltage applied, and someone thought I might have been implying "not modulated," and that's what started this.
It's times like this I wish I owned an oscilloscope.
I was arguing that the bulb is never in an "off" state, even at the zero points, as even though it is cooling, it never goes to zero output, which is in fact, "off".
I wasn't trying to say they don't modulate... I was going for don't have zero output at any point on the cycle. I figure the light output looks like a rectified sine wave with a D.C. offset, when put on a scope.
This is opposed to LEDs and flourescents, which have light output resembling a square wave, and definite are emitting zero light for part of the cycle.
Wish I had an oscilliscope right now. I've suspected the telltale LEDs on my new external modem... but haven't had a way to check since I currently don't own or have access to a scope.
Because the technology isn't around, does not mean it isn't right around the corner, and does not mean that it doesn't exist elsewhere.
I spent enough time working for the government to know that there is technology out there that YOU don't know exists, and probably never will until someone who isn't a government contractor re-invents it for some civilian use.
I am absolutely positive that now that someone has said this is possible, someone else is working on a way to do it even as we talk about it.
Your class III equipment might be safe for now, but do you think they are safe for long?
I didn't read his paper myself, but I certainly agree that it's possible. One would only need to read the intensity of the light, convert that to a video signal, synced with the rate of the display you are spying on, which is easy because of the verticle blanking period.
Wish I'd thought of it first. I'm looking for a thesis idea....
And today, we are going to learn about math and electricity:
AC current flows in a sine wave. Now, I will assume you know what a sine curve looks like.
At a sine curve's peaks, at pi/2 radians from zero in either direction on the unit circle, the absolute unit is 1. Its zero is at zero.
Now, it is only zero at zero degrees. At all other times it is NOT zero, and thus, current is flowing. On a cycle of pi radians, there are an infinite number of points where current is flowing, and only THREE where it is zero, and "stopped" as you say. Since an incandescent bulb is resistant no matter the volage, and has a slow cooling time, the bulb is infinitely "on" for the complete cycle, because it does not turn "off" during the infinitely small zero points of the curve.
Now, the reason LEDs pulse is because their switching speed is near-instantaneous, and they only flow current in one direction.
Flourescents are similar, but generally more apparent in their flickering because of "threshold voltage", which basically, increases the size of the zero points on the curve, because light output is effectively zero for input voltages less than a certain amount. LEDs have a threshold voltage too, but it's a lot smaller percentage generally, for zero light output.
-- Someone decently technically savvy builds his/her own PC with an AMD chip;
You don't explain what exactly the is that you think AMD chips have. Furthermore, your list of steps taken toward system incompatibility ends with:
-- Said person goes and manually messes with IRQ settings, thus wreaking havoc on the poor commputer that functioned perfectly before.
Which is the real cause of the problem you are describing.
I just like to take the time to point out that I use an AMD processor, and that the last Intel-based system I ever owned was a P200MMX. My machines (self-built) are ALWAYS reliable, and do not have problems with ANY version of Windows, including ACPI support.
In fact, I usually find that many stability problems are directly related to the quality and condition of the hardware, such as:
Accidentally put a small scratch on the motherboard? Looks okay, probably didn't cut a trace, right? No, but if it got down to the trace, you just created a point of extra impedance in the trace... future stability problem. In fact, even if there is no obvious damage, if you dropped the end of the screwdriver onto the motherboard, it may have caused subtle damage to the circuitry that can show up as stability problems.
I once saw a system come in that had problems, only to find a loose screw under the motherboard.
Simple way to improve a system's stability (physically, and in software): Put in ALL the screws that belong in the case. ALL of the drive mounting screws. ALL hardware mounting screws. Do NOT put in one here and there just to keep things tied down... put them ALL in. Not only do they help anchor hardware and dissipate vibrations from moving parts, they provide a ground path for shielding, and shielding from electrical noise is important. Thumbscrews are fine, and I recommend them.
Another one: Don't "loop" cables that are too long. Always use cables that are the correct length for the application. "Looped" cables create larger magnetic fields than ones that are not. Magnetic fields can induce spurious voltage potential in nearby circuits.
On that subject, keep the cables as far from the surface of the motherboard as possible, for the same reason. Use good quality cables. Also, I've heard of more problems from rounded IDE cables over flat ones.
Tip on RAM: Always use high-quality, name-brand memory, not no-name junk from god-knows-what-fourth-world-country. Memory that is even the slightest out of spec can cause intermittent problems.
Fans and cooling: Where possible, lways install dust filters where air intake occurs. For intake fans on the back of a computer, there are "snap on" filters that can be mounted exterior to the fan. Clean filters regularly, and blow any dust out of the computer periodically.
When installing fans, and multiple placement options are possible, think of a place that gets greatest airflow. Every other fan should be an exhaust fan, not counting the power supply fan. Try to think about air current in a system.
Make sure there is enough cooling for the hard drives, as they can get very hot. My policy is one additional fan for every two drives installed.
I didn't mean to turn this into a class on system design, but that is how you build a rock-stable system. I've built computer systems for myself and for others since my first '286 way back. I DO have experience here.