I use TPLink's TL-WR1043ND, however with openwrt, not dd-wrt (although it is supported by ddwrt since feb. 2011 build). My main reason was gigabit ethernet followed by flash (8MB) and price (I think it's one of the cheapest gigabit routers) I'm using it as a main router for over 2-3 months, with wifi for android phone and 2 laptops (only 1 laptop with 802.11n).
This is all about W8 logo. W8 will run on BIOS/unsecure UEFI (just like W7), but MS will not allow the OEM to use the W8 logo. If the OEM will have enough complaints, they will abandon MS requirements or offer firmware with the option. Maybe unsupported/beta firmware. I'm sure the OEMs will "create" a W8 "ready" logo.
I second the TP-WR1043ND. It was the cheapest GB LAN router I could find in my country from the OpenWRT list. I'm using it on a 20/2.5Mbits/s VDSL line and it works great (although I do limit my torrent speeds to 800/150KB/s down/up, served from another PC).
I read just a few line in the article, but apart from history, it seems the article is about navigation SW (with GPS HW), not pure GPS. I remember I just used 2-3 times the "navigation" part of the SW, the rest was just birds-eye view of my current location.
My ubuntu server did this by default (I don't remember which 8.x was used for installing). I remember when I replaced my mini-itx with a laptop (HDD went inside laptop) to continue as my router and had to shuffle MACs (2nd ethernet was a USB bridge which was unchanged since it was present on both HWs).
You said it yourself..."without warning them". That is not what racing drivers do.
But what you are talking about is driver preference, not manufacturer designs. Usually the driver prefers a quieter engine, and that means low rpm. And he usually drives it at low rpm too. For gasoline engines, that means 80% of times below peak torque rpm. My engine has peak torque at 4K rpm, and that is usually the latest I shift. So to answer your question, I don't prefer high-revving engines for every-day driving. But the topic is about car manufacturers and their reliability.
And about your horsepower and torque, remember that what the manufacturers specify are maximums. You will always have some horsepower and torque, which are affected by pedal position and rpm. And if you calculate the HP at peak torque and rpm, you will see that you have somewhere between 50-70% of your peak HP, depending on engine. And for best acceleration, forget about engine torque, it's torque at wheels you want. And you will always maximize that when you are around peak engine power.
And a high-rpm engine will always keep you on edge, as you need to downshift prior to making a high-acceleration manoeuvre.
While I don't disagree completely, there is one more detail: engine maintenance. A 10-year old engine NEEDS some parts replaced. In 10 years of running (let's say 7000 miles/year) you will have to change the seals, otherwise you will have lots of exhaust gases in the oil and lots of burnt oil. Keeping on running that is not healthy even on the short term. And considering todays 15-20 year old cars, I would say you argument is not valid.
But 6000rpm is not uncommon. My car has a 1.6L engine with a 6400rpm cut-off. That means they need to guarantee the engine works flawlessly up to 6400rpm, with any fuel from the specified range (95 or 98 in Europe). It does not matter how I drive it. The manufacturer specifies peak HP at 6000rpm, so it has to deliver it. Failiures are allowed only from wear, not electronic. And for motorcycles.....my city/touring Yamaha has a 12000rpm cut-off. But the 1L Fazer has peak power at 12500rpm.
Actually there is...at least I know about the Infineon Tricore 1782...it's rated for automotive use (-40C to 150C), is build in 90nm technology and works at 133MHz or 180MHz. I don't really know why they have 2 speed ratings...possibly the reccomended values for power envelope, so the project decides what to use. But you probably won't see anything based on it until mid-2011, as it's a pretty new processor.
Please read the article again.....It talks about the microcontroller of the engine, not the suspension. The suspension ECU does not need the 150C grade. And 1ms control for engine is way too slow. At 5000rpm you have 12ms/rotation, that means the engine turns 30 degrees/ms. When you control just spark advance, 30 degrees is difference between knocking (too high advance, very dangerous to the engine) and burning the fuel in the catalyzer (too low advance, dangerous to the catalyzer). And now also add the fuel injection for the cylinder that is in intake phase. Even more, you can have an airflow prediction algorithm so you don't keep the same fuel flow for the intake phase, so you get as close to the perfect mix in the cylinder. In other words, the suspension is something that you notice immediately, but the engine output is very hard to notice. You cannot notice a 10nm difference in engine output, unless you are a race driver and comparing lap times. But even the suspension can affect the lap times.
Engine Control Units ARE close to the engine. If you've seen some in the passenger compartment, it was either an old car, or it was not and engine ECU (ECU is used as "electronical control unit" in the automotive industry). This automotive grade comes for those units that are directly clamped to the engine. While many are just in the engine compartment (shorter cables for high-current control), some are actually clamped to the engine block. That means it should work normally at 90C, even at 110-120C. This temperature is linked to the maximum temperature a coolant can achieve (which is water-based) before turning to vapor. But the water is constantly running, so heat buildup cannot be achieved for the coolant, but that does not exclude the engine block reaching 100+C.
It may be, but that is mechanical and hidraulic only. Do remember that around 90s and 2K they started targeting "consumerism" (or what you call that). About the cars, I can tell you that all SW needs to be available 15 years from the launch date. So any bugs found after 5 years, you can still repair them. But after 15 years, you do what you want with the archives. But that means that the 1st cars with ECUs (pre 1995) just started to "expire". I doesn't mean they won't work. The car manufacturer is responsible of having enough ECUs to cover those 15 years (or more) for replacements, but the idea is to make the people drop the old cars and drive the new, less polutant cars. The pollution only came around 2000 (at elast in Europe).
Excuse me, but x86 is used only on desktop. Only where compatibility is a gain (like running 20-year-old SW without changing...don't laugh....it's still done). Car algorithms are evolving each year, so the SW is constantly recompiled. So they are not burdened by x86. Even your GPS device is NOT x86. You cell phone is NOT x86. It seems you are one of the people who buy the "big number" instead of the performance. And by performance, I don't refer solely to number-crunching. I mean what you need, like lowest power for the "enough" speed. Compatibility is being able to take a 2008 Camaro engine ECU and putting it in a 1998 Camaro. You just cannot do that. So when talking about car computers, forget everything you know about Windows....it's useless.
I think I just bit a flamebait. But the only part you are right is with the placement under the seat. While I don't know for older models (pre-90s ???), that was never true for engine ECUs. I know a specific model of gearbox that has the ECU in passenger compartment, but all post-2000 engine ECUs are clamped (or close) to the engine block. That is because what is controlled by the ECU has a high current, and the wires can be cut. So the shorter the wires, the less chance of cutting them. Also, I really don't see how you can say then engine compartment stays cooler...... This is only true for a car in the sun which was NOT started that day. The engine water will get to 80-90C within minutes of driving. So the engine compartment can have 50C very easily even in winter.
Since he is quoting 16-year-olds who start playing games like GTA, I think they should consider the teenager psichology: they always want to show-off. It's a constant competition, and that is the #1 reason of teenage car accidents. I know, I tipped-over a car after 6 month of driving license (18 years). Locky for me my passengers were only bruised.
PS: I play simulators now. At that time I was a big NFS fan.
Do you mean the source code for the Apache web server itself? Hasn't that always been available? Since when has it been a closed source product like IIS?
The problem is they could do changes and reviews posing as other long-time trusted developers. This is a vulnerability in the chain of trust. If such an attack would be discovered too late, all the latest updates could be left vulnerable. For example, if the attack is discovered 1 month later, some back-door patches could get in the stable (production) release.
So this is like cancer: discovered in time, the consequences can be controlled, but discovered too late, all hell breaks loose.
I'm personally using: iptables with "-A INPUT -p tcp -m tcp --dport 22 -m hashlimit --hashlimit-upto 5/min --hashlimit-mode srcip --hashlimit-name sshd -j ACCEPT" and a default DROP rule. The only problem with this is that it does not make permanent blacklists and I don't know how much overhead the kernel does. However when I "made" this line, I was getting around 5 attempts per second from one IP.
PS: I have a rule which accepts RELATED and ESTABLISHED packets, so the line above applies to new packets only.
Actually, every computer has lithium batteries.....remember those CMOS batteries? CR2032? Those ARE lithium bateries. CR = Lithium/manganese dioxide, at 3V Gratned, it's not lithium-ion (like the laptop batteries), where the lithium is used also in the electrolyte, not only in the cathode.
Actually, your boss has the right to tell you what you can or can't listen to at work.
I'm don't agree with you. While I don't know the law, I do know that forbidding using headphones (except for noise-reducing ones in loud enviroments) is in safety-related work. While programming can be FOR a safety functionality, programming itself does not relate to split-second reactions. Even in car and bicicle riding, headphones are "recommended" not to be used. In my case, when I drive I listen to music but not loud enough to not hear an ambulance. When I ride my bicycle I also listen to music, but I adjust it's volume to know when a car is close to me. Listening to music on motorcycle is stopped just because I cannot use regular heaphones with my helmet, but there are helmets with speakers. So again, I don't agree on forcing people to stop listening to music while programming. The only reason I can see it enforced is for pirated music (where they "obey the law"), but then there is a matter of how to check for legal music. However, I doubt a company has 100% licenced software. Remember MS having win98 theme sounds done with pirated goldwave?
As my point of view related to quality of code, background music (at least music that you like) do help keep the mind relazxed. So while stopping music whould probably raise the quality for the 1st month, I doubt it can keep it there. As I analyze my behavior, I realize that when I'm concentrated on programming, I barely hear the music. But when that concentration is interrupted by failiure to find a solution (things not working as I expect), music brings me back to a relaxed state. If I would be stopped from using music, I would just start yelling at co-workers with "shut up".
Oh yeah, and quality of code is a stupid metric in my opinion. Whenever a boss talk about quality code, it talks about documents, and a stupid code can get good quality because of a bad (choice of ) reviewer. So think of quality metrics like an XOR between author and reviewer. And I know what I'm talking about: I work for automotive (embedded) industry with 10+ year old code. And when I see a bad code and ask permission to change it, I always get the response "it's already well tested and it works". I even had once to make a workaround instead of fixing the main problem (checking for an end-of-list marker only from the 2nd element, causing traps if the list is empty). And one more thing, before you say "maintenance", I'm talking about code which is still in development (having already many freezes entered in maintenance). In my view, quality metrics is just a way to justify keeping bad (and cheap) programmers.
Old timers...... I have the following server/router: Intel Core2Duo E5300 (I actually wanted the E5200) Intel DG45FC WD Scorpio Black 320GB 200W Foxconn power suply (my 500W Antec Phantom is not that efficient at this load....I get +10W)
Now, my APC UPS measures around 30-40W in idle and I think 75W in full load (kernel recompile). I think the only way power could go down is with Atom and/or PicoPSU.
When copying files, I can easily do 30MB/s (mainly HDD is the bottleneck). However, My DNS-323 is limited to 17MB/s. But the DNS is limited by CPU, whereas my server is limited by HDD. Oh yeah.. my server is a 24-7 torrent client (the reason why I opted for a linux PC router and not embedded).
I wish someone would have told me this before..... Now seriously, I knew this would be a risk, I just ignored it (bad! bad! bad!).
Last week I was hacked (my home system/router) and I noticed because my ssh server refused my authentification. That's how I knew I was hijacked. I had 1 user with dictionary-word password and full sudo access. The reason it stayed so long is because I saw long time ago a ssh brute-force attempt (100s of tries/second), so using iptables I limited the incomming ssh connection attempts to 5/min. This would be enough for my daily connections from work, but very slow for brute-force. But of course, slow is not the same as impossible.
So my server migration plans (gentoo -> ubuntu) were finally executed. I wanted to do the change for a year but always delayed it because of "if it works, don't fix it" and the time required. This weekend I had no choice.
PS: Am I just paranoid or there is too big of a coincidence of this article appearing just after I was hijacked? I keep hearing the phrase: "nothing is coincidence".
By default, on a 4GB system, 2GB is for userland and 2GB is for the kernel.
Another virtual vs physical confusion. 2+2GB limitation is per process/task/application. This is the SW. If there is not enough RAM, swap is used. 4GB system is a system with 4GB of RAM. This is HW. No matter how much RAM you have, ONE 32-bit application will NOT use more than 4GB (2+2 or 3+1 if/3GB is used) even on 64-bit OS. Let's just say that PAE on 32-bit OS allows you to use the above-4GB RAM as super-fast swap space. And this is the XP (as opposed to 2K3 server) limitation: XP just does not let you use it. I'm also thinking it's because of drivers and DMA devices. PAE is a very old hack.
In fact, bleating that it's not fair because Linux developers are volunteers may make things worse. You're basically saying Linux is amateur. People want their software to be professional.
I would rather use a SW made by volunteers than by cheap pay-by-hour "professionals". Were I work are many times rantings about how they can't keep their good programmers. I doubt MS is different. On the other hand, if a "professional" SW breakes, you can sue the company, as for GNU/Linux, who do you sue? Everybody will say "you get what you payed for".
Slashdot Mirror
I use TPLink's TL-WR1043ND, however with openwrt, not dd-wrt (although it is supported by ddwrt since feb. 2011 build).
My main reason was gigabit ethernet followed by flash (8MB) and price (I think it's one of the cheapest gigabit routers)
I'm using it as a main router for over 2-3 months, with wifi for android phone and 2 laptops (only 1 laptop with 802.11n).
This is all about W8 logo. W8 will run on BIOS/unsecure UEFI (just like W7), but MS will not allow the OEM to use the W8 logo. If the OEM will have enough complaints, they will abandon MS requirements or offer firmware with the option. Maybe unsupported/beta firmware. I'm sure the OEMs will "create" a W8 "ready" logo.
I second the TP-WR1043ND. It was the cheapest GB LAN router I could find in my country from the OpenWRT list. I'm using it on a 20/2.5Mbits/s VDSL line and it works great (although I do limit my torrent speeds to 800/150KB/s down/up, served from another PC).
I read just a few line in the article, but apart from history, it seems the article is about navigation SW (with GPS HW), not pure GPS.
I remember I just used 2-3 times the "navigation" part of the SW, the rest was just birds-eye view of my current location.
My ubuntu server did this by default (I don't remember which 8.x was used for installing). I remember when I replaced my mini-itx with a laptop (HDD went inside laptop) to continue as my router and had to shuffle MACs (2nd ethernet was a USB bridge which was unchanged since it was present on both HWs).
You said it yourself..."without warning them". That is not what racing drivers do.
But what you are talking about is driver preference, not manufacturer designs. Usually the driver prefers a quieter engine, and that means low rpm. And he usually drives it at low rpm too. For gasoline engines, that means 80% of times below peak torque rpm. My engine has peak torque at 4K rpm, and that is usually the latest I shift. So to answer your question, I don't prefer high-revving engines for every-day driving. But the topic is about car manufacturers and their reliability.
And about your horsepower and torque, remember that what the manufacturers specify are maximums. You will always have some horsepower and torque, which are affected by pedal position and rpm. And if you calculate the HP at peak torque and rpm, you will see that you have somewhere between 50-70% of your peak HP, depending on engine. And for best acceleration, forget about engine torque, it's torque at wheels you want. And you will always maximize that when you are around peak engine power.
And a high-rpm engine will always keep you on edge, as you need to downshift prior to making a high-acceleration manoeuvre.
While I don't disagree completely, there is one more detail: engine maintenance. A 10-year old engine NEEDS some parts replaced. In 10 years of running (let's say 7000 miles/year) you will have to change the seals, otherwise you will have lots of exhaust gases in the oil and lots of burnt oil. Keeping on running that is not healthy even on the short term. And considering todays 15-20 year old cars, I would say you argument is not valid.
That must be an old Garmin. Because most (is not all) GPS devices in the last 10 years are ARM based.
But 6000rpm is not uncommon. My car has a 1.6L engine with a 6400rpm cut-off. That means they need to guarantee the engine works flawlessly up to 6400rpm, with any fuel from the specified range (95 or 98 in Europe). It does not matter how I drive it. The manufacturer specifies peak HP at 6000rpm, so it has to deliver it. Failiures are allowed only from wear, not electronic.
And for motorcycles.....my city/touring Yamaha has a 12000rpm cut-off. But the 1L Fazer has peak power at 12500rpm.
Actually there is...at least I know about the Infineon Tricore 1782...it's rated for automotive use (-40C to 150C), is build in 90nm technology and works at 133MHz or 180MHz. I don't really know why they have 2 speed ratings...possibly the reccomended values for power envelope, so the project decides what to use. But you probably won't see anything based on it until mid-2011, as it's a pretty new processor.
Please read the article again.....It talks about the microcontroller of the engine, not the suspension. The suspension ECU does not need the 150C grade. And 1ms control for engine is way too slow. At 5000rpm you have 12ms/rotation, that means the engine turns 30 degrees/ms. When you control just spark advance, 30 degrees is difference between knocking (too high advance, very dangerous to the engine) and burning the fuel in the catalyzer (too low advance, dangerous to the catalyzer). And now also add the fuel injection for the cylinder that is in intake phase. Even more, you can have an airflow prediction algorithm so you don't keep the same fuel flow for the intake phase, so you get as close to the perfect mix in the cylinder.
In other words, the suspension is something that you notice immediately, but the engine output is very hard to notice. You cannot notice a 10nm difference in engine output, unless you are a race driver and comparing lap times. But even the suspension can affect the lap times.
Engine Control Units ARE close to the engine. If you've seen some in the passenger compartment, it was either an old car, or it was not and engine ECU (ECU is used as "electronical control unit" in the automotive industry).
This automotive grade comes for those units that are directly clamped to the engine. While many are just in the engine compartment (shorter cables for high-current control), some are actually clamped to the engine block. That means it should work normally at 90C, even at 110-120C. This temperature is linked to the maximum temperature a coolant can achieve (which is water-based) before turning to vapor. But the water is constantly running, so heat buildup cannot be achieved for the coolant, but that does not exclude the engine block reaching 100+C.
It may be, but that is mechanical and hidraulic only. Do remember that around 90s and 2K they started targeting "consumerism" (or what you call that). About the cars, I can tell you that all SW needs to be available 15 years from the launch date. So any bugs found after 5 years, you can still repair them. But after 15 years, you do what you want with the archives. But that means that the 1st cars with ECUs (pre 1995) just started to "expire". I doesn't mean they won't work. The car manufacturer is responsible of having enough ECUs to cover those 15 years (or more) for replacements, but the idea is to make the people drop the old cars and drive the new, less polutant cars. The pollution only came around 2000 (at elast in Europe).
Excuse me, but x86 is used only on desktop. Only where compatibility is a gain (like running 20-year-old SW without changing...don't laugh....it's still done). Car algorithms are evolving each year, so the SW is constantly recompiled. So they are not burdened by x86. Even your GPS device is NOT x86. You cell phone is NOT x86. It seems you are one of the people who buy the "big number" instead of the performance. And by performance, I don't refer solely to number-crunching. I mean what you need, like lowest power for the "enough" speed.
Compatibility is being able to take a 2008 Camaro engine ECU and putting it in a 1998 Camaro. You just cannot do that. So when talking about car computers, forget everything you know about Windows....it's useless.
I'm sure those were not for the engine. There are cars with 10 or more ECUs, from which only 1 is for the engine.
I think I just bit a flamebait. But the only part you are right is with the placement under the seat. While I don't know for older models (pre-90s ???), that was never true for engine ECUs. I know a specific model of gearbox that has the ECU in passenger compartment, but all post-2000 engine ECUs are clamped (or close) to the engine block. That is because what is controlled by the ECU has a high current, and the wires can be cut. So the shorter the wires, the less chance of cutting them.
Also, I really don't see how you can say then engine compartment stays cooler...... This is only true for a car in the sun which was NOT started that day. The engine water will get to 80-90C within minutes of driving. So the engine compartment can have 50C very easily even in winter.
Since he is quoting 16-year-olds who start playing games like GTA, I think they should consider the teenager psichology: they always want to show-off. It's a constant competition, and that is the #1 reason of teenage car accidents. I know, I tipped-over a car after 6 month of driving license (18 years). Locky for me my passengers were only bruised.
PS: I play simulators now. At that time I was a big NFS fan.
Do you mean the source code for the Apache web server itself? Hasn't that always been available? Since when has it been a closed source product like IIS?
The problem is they could do changes and reviews posing as other long-time trusted developers. This is a vulnerability in the chain of trust. If such an attack would be discovered too late, all the latest updates could be left vulnerable. For example, if the attack is discovered 1 month later, some back-door patches could get in the stable (production) release.
So this is like cancer: discovered in time, the consequences can be controlled, but discovered too late, all hell breaks loose.
I'm personally using:
iptables with "-A INPUT -p tcp -m tcp --dport 22 -m hashlimit --hashlimit-upto 5/min --hashlimit-mode srcip --hashlimit-name sshd -j ACCEPT" and a default DROP rule.
The only problem with this is that it does not make permanent blacklists and I don't know how much overhead the kernel does. However when I "made" this line, I was getting around 5 attempts per second from one IP.
PS: I have a rule which accepts RELATED and ESTABLISHED packets, so the line above applies to new packets only.
Actually, every computer has lithium batteries.....remember those CMOS batteries? CR2032? Those ARE lithium bateries.
CR = Lithium/manganese dioxide, at 3V
Gratned, it's not lithium-ion (like the laptop batteries), where the lithium is used also in the electrolyte, not only in the cathode.
Actually, your boss has the right to tell you what you can or can't listen to at work.
I'm don't agree with you. While I don't know the law, I do know that forbidding using headphones (except for noise-reducing ones in loud enviroments) is in safety-related work. While programming can be FOR a safety functionality, programming itself does not relate to split-second reactions. Even in car and bicicle riding, headphones are "recommended" not to be used. In my case, when I drive I listen to music but not loud enough to not hear an ambulance. When I ride my bicycle I also listen to music, but I adjust it's volume to know when a car is close to me. Listening to music on motorcycle is stopped just because I cannot use regular heaphones with my helmet, but there are helmets with speakers.
So again, I don't agree on forcing people to stop listening to music while programming. The only reason I can see it enforced is for pirated music (where they "obey the law"), but then there is a matter of how to check for legal music. However, I doubt a company has 100% licenced software. Remember MS having win98 theme sounds done with pirated goldwave?
As my point of view related to quality of code, background music (at least music that you like) do help keep the mind relazxed. So while stopping music whould probably raise the quality for the 1st month, I doubt it can keep it there.
As I analyze my behavior, I realize that when I'm concentrated on programming, I barely hear the music. But when that concentration is interrupted by failiure to find a solution (things not working as I expect), music brings me back to a relaxed state. If I would be stopped from using music, I would just start yelling at co-workers with "shut up".
Oh yeah, and quality of code is a stupid metric in my opinion. Whenever a boss talk about quality code, it talks about documents, and a stupid code can get good quality because of a bad (choice of ) reviewer. So think of quality metrics like an XOR between author and reviewer. And I know what I'm talking about: I work for automotive (embedded) industry with 10+ year old code. And when I see a bad code and ask permission to change it, I always get the response "it's already well tested and it works". I even had once to make a workaround instead of fixing the main problem (checking for an end-of-list marker only from the 2nd element, causing traps if the list is empty). And one more thing, before you say "maintenance", I'm talking about code which is still in development (having already many freezes entered in maintenance).
In my view, quality metrics is just a way to justify keeping bad (and cheap) programmers.
Old timers......
I have the following server/router:
Intel Core2Duo E5300 (I actually wanted the E5200)
Intel DG45FC
WD Scorpio Black 320GB
200W Foxconn power suply (my 500W Antec Phantom is not that efficient at this load....I get +10W)
Now, my APC UPS measures around 30-40W in idle and I think 75W in full load (kernel recompile).
I think the only way power could go down is with Atom and/or PicoPSU.
When copying files, I can easily do 30MB/s (mainly HDD is the bottleneck). However, My DNS-323 is limited to 17MB/s. But the DNS is limited by CPU, whereas my server is limited by HDD.
Oh yeah.. my server is a 24-7 torrent client (the reason why I opted for a linux PC router and not embedded).
Or you could just not use weak passwords.
I wish someone would have told me this before.....
Now seriously, I knew this would be a risk, I just ignored it (bad! bad! bad!).
Last week I was hacked (my home system/router) and I noticed because my ssh server refused my authentification. That's how I knew I was hijacked. I had 1 user with dictionary-word password and full sudo access. The reason it stayed so long is because I saw long time ago a ssh brute-force attempt (100s of tries/second), so using iptables I limited the incomming ssh connection attempts to 5/min. This would be enough for my daily connections from work, but very slow for brute-force. But of course, slow is not the same as impossible.
So my server migration plans (gentoo -> ubuntu) were finally executed. I wanted to do the change for a year but always delayed it because of "if it works, don't fix it" and the time required. This weekend I had no choice.
PS: Am I just paranoid or there is too big of a coincidence of this article appearing just after I was hijacked? I keep hearing the phrase: "nothing is coincidence".
By default, on a 4GB system, 2GB is for userland and 2GB is for the kernel.
Another virtual vs physical confusion. /3GB is used) even on 64-bit OS.
2+2GB limitation is per process/task/application. This is the SW. If there is not enough RAM, swap is used.
4GB system is a system with 4GB of RAM. This is HW.
No matter how much RAM you have, ONE 32-bit application will NOT use more than 4GB (2+2 or 3+1 if
Let's just say that PAE on 32-bit OS allows you to use the above-4GB RAM as super-fast swap space. And this is the XP (as opposed to 2K3 server) limitation: XP just does not let you use it. I'm also thinking it's because of drivers and DMA devices.
PAE is a very old hack.
In fact, bleating that it's not fair because Linux developers are volunteers may make things worse. You're basically saying Linux is amateur. People want their software to be professional.
I would rather use a SW made by volunteers than by cheap pay-by-hour "professionals". Were I work are many times rantings about how they can't keep their good programmers. I doubt MS is different. On the other hand, if a "professional" SW breakes, you can sue the company, as for GNU/Linux, who do you sue? Everybody will say "you get what you payed for".