This is thinking along the right lines. For years I thought it would be cool to have some sort of tablet PC so I could surf the net on my couch. After getting a 42" plasma screen a couple of weeks ago, I now realise the solution I always needed was to have an old PC hooked up to a huge screen and just have a remote keyboard & mouse. Works great - you can watch tv, movies etc and also surf the net at a lovely 1920x1280 resolution. Total cost: About $AU 1700 for a 42" plasma, + $100 bucks for a logitech wireless keyboard & mouse with 10M range (about 30 ft) + one old PC I already had lying around. Well worth it, and way better than any hand held solution. It's just not feasible to surf today's web content on a pocket PC.
I think you hit the nail exactly on the head. The amount of network traffic needed by n number of players in the same area goes up at a squared rate ie. on average, if you have 10 people that are all able to see each other, the servers have to send out 10 movement messages (one to each player) for every player that moves. so if every player does something once a second, this is 10x10 messages per second if there are 10 nearby players. if there are 20, it would be 20x20 = 400,
Imagine how much more traffic the server would have to deal with if you could get half of a realm's population turning up all at once in some location from wherever they were, when something interesting happens there suddenly? If there were 2000 people on the server, and half of them all suddenly decided there was something cool going on at point X, first the server would have to tell anyone in the immediate region where you were teleporting from that you had vanished. Then there would be a massive amount of updates for each player as they teleport to the new region, because the player would have to get a complete set of info about all the other players in that area. Finally, the server would then have to be sending Next, the server would have to and would then have to be dishing out 1000 messages per person in the area, ie, 1000000 messages alltogether, every time everyone took a step or changed direction.
By having slower travel times, it means that at any given moment, a fairly large chunk of the world's population will be distributed in sparsely populated areas, so there are much fewer messages to send to nearby players as they move to their destinations, because each player would only be near say, 3 or 4 others - and often not near anyone at all.
Secondly, because it is a bit of a hastle to go somewhere, there is much less tendancy for as many people to suddenly congregate at a particular point when something happens there.
This is also a good reason to have cities or regions that are specialising in different things - if you can get everything you possibly need from one place, then this will also tend to make players bunch up in htat cities instead of going to different cities, and cause high load. This was done to a certain extent in UO, but I think WoW deals with the problem better with the taxis because it also gives you a chance to go grab a cup of coffee or a bio break and mabey actually wrench yourself away from your chair for a quick leg stretch while in transit.
Good thing that that we don't have to use fossil fuels to supply the power needed to keep that big ball of plasma and nuclear goodness hot then isn't it?
I am not a physicist, but I am pretty sure the Carnot efficiency cycle is the theoretical limit of a heat engine - ie. one that works by extracting work from expanding gases.
Last time I checked, Solar was a means of converting electromagnetic radiation into electricity, so the Carnot cycle does not apply to it - even if you are driving it from say, a burning chunk of magnesium or something. Instead, it is the band-gap of the materials used in the P-N junction that limits the wavelengths that are absorbed. If you had a light source perfectly matched to the band-gaps of the materials in your solar panel, the efficiency would be much higher - and a lot of the current solar research is to do with layering different materials so that there is a wider range of band-gaps in the materials so that more wavelengths are absorbed and converted to electricity.
I would be interested to know what sort of effect applying this sort of laser surface treatment would have to the efficiency of both LED's and solar panels, actually. Any physicists like to comment?
vi or emacs is great if you are writing a small application, or making a few small changes, but time is money, and programmers time is not cheap.
A good IDE can save heaps of time, and in my opinion, automatic code completion is essential in any good programming environment, as it substantially reduces the time it takes looking up exactly what that object's field name was, or what the parameters are for a particular function. It is also extremely useful if there is a way to jump straight to the line of code where a particular variable or class is defined.
Code::blocks is a fairly good IDE that can do this, and has a nice clean interface, and is my IDE of choice when I am working on C++ projects on Linux ( I tried kdevelop, eclipse, and a few others before settling on code::blocks) , but unfortunately it is still not quite as good as Visual Studio on windows, and even though you have to shell out about 1200 bucks for the professional edition here in Aus, if you are getting paid to write software, then that cost is well worth the investment considering the thousands of hours a large project takes to write. I write programs for a living, and really don't want to be spending all my time just trying to navigate around my code, or trying to use a command line based build environment and debugger which is not tightly integrated with the IDE.
If you are just writing stuff on the side, and it is not your day job, then check out code::blocks.
If you are getting paid for your time, then do your client a favour and get the right tools for the job, and spend the money on a tool which allows you to make the most efficient use of your time. Claim the cost of the tool on your tax.
If you use linux you might want to check out http://ktechlab.org/. It is a circuit simulator for PIC micro-controllers, where you can write your PIC code in either C, a flowcharting language, or in assembler. You can attach circuit elements to the PIC like potentiometers, resistors, capacitors, and other components to build a circuit around the PIC to test the software you have written even before you go anywhere near hardware. It also supports several PIC burners too, aparently - for when its time to produce real hardware from your simulation.
Ktechlab does have a few bugs here an there, but still complete enough to be useful. Apparently the project has recently changed hands and is getting under way again after a couple of years of stagnation. It's a fantastic tool - way easier to use than any of the other circuit simulators I have used, although the version in the Ubuntu 8.10 repository is broken (crashes when placing a component), and I had to download a version directly from sourceforge.
I have spent quite a bit of time searching for the right display - If you re-read my original post, you will note that I am looking for something that has at least a passing resemblance to a pair of sunglasses - slim shades or even (but preferably not) a Geordie style visor. All the high resolution ones you have posted look like a brick-in-face, and I wouldn't be caught dead wearing them on a train or out from behind locked doors. Would you? The lower resolution one is getting there, but still only 800x600, and one eye only. I think it would be maddening looking at a display with only one eye for any extended period of time.
If it needs a strap to hold it on my head, it's too damn big.
There are now displays that have an acceptable appearance and are semi-transparent, much like the example product I linked in my original post, which seems great, except it's too low resolution, being only 640x480, and I think its even stereo.
I have pretty much given up hope of finding them and was sort of hoping that if there were enough people that wanted something like I specified, and enough people asked the manufacturers for what we want (wearables suitable for use as computer displays) instead of what they seem to be producing (ie. wearables that are really aimed at playing movies) then we might eventually get what we want.
Thanks for taking the time to add all those links, but unfortunately I have seen them all before.
You nailed it exactly. I had a PDA for a while, but the teeny screen was always disappointing. I am sure when headphones first came out they looked like a couple of speakers duct taped on your head and sounded awful. They progressed to smaller and smaller units until you get the nice slim ubiquitous ear bud headphones of today, with acceptable (if not perfect) audio quality too. Now it's time the same thing was done for our visual field.
The first HMD I saw was a sony glasstron while I was in Japan , much like you describe. It was good, but too bulky and too low rez. That was about 12 years ago. I have been looking for a decent display ever since, and can't believe that despite the fact there has been two orders of magnitude change in CPU power and transistor density, the LCD's for wearable displays have hardly even doubled in resolution.
The reason I would like them semi-transparent is so that you do have local environment awarenes still, and can flick your eyes into the distance every now and then to keep the muscles working through multiple focal depths. (I code all day, and when I can,I sit next to a window so I can focus in the distance every 20 min, or when thinking. I am convinced it's the only reason I still have better than 20/20 vision after decades of programming)
The reason I want them high res is because I want to be able to use them for textual based apps ( think suring the web, doing a bit of coding, all the stuff you usually do on a PC, but having the flexibility of being able to do it anywhere.) The reason I want them to be capable of at least 30fps ( and preferably 60fps as LCD monitors typically are) is because of that motion lag you so correctly pointed out. The reason I want them to not look ludicrous is because I actually want to wear them, sometimes out doors too!
Wow, they must be etching the glass substrate by hand with the Hope diamond for that price. It still looks like a brick tied to your face, too. Nice rez though.
You have pretty much nailed my list of priorities - though wouldn't you count high resolution as part of image quality? I know it is practically impossible to achieve a wide fov with an unobtrusive system, and actually wide fov is not that critical to me - given a choice between an unobtrusive high resolution field of view that was equivalent to whatever a 24" monitor on a desk 2 feet away is, (20 to 30 degrees?) and a low resolution 120 degree field of view, I would definitely be choosing the former. Expense wise, well I would spend up to $1000 or so I guess - perhaps $1500, if the product was usable for long periods of time just as a regular monitor is without eye strain.
Seems like there are a lot of problems to solve still - I just find it frustrating that there seem to be products out there that seem so close to my requirements, only lacking in resolution, and probably refresh rate.
I don't wear glasses, despite years at a screen since the Apple//e appeared on the scene, and years of reading books in the dark before that, so I have never had to worry about whether you could wear glasses with some particular hmd in my quest for the perfect one, but I suppose that also adds another whole level of complexity to trying to bring a product to as wide a market as possible.
Is there really such a passive price difference between the LCD's they use in these systems that do 640x480 compared to one that could do 1024x768 though? I could live with it still looking about the same apparent size and distance as a 24" monitor sitting on my desk, but I just don't want to wear a device that looks too out of place on a train or in public, hence the requirement for them to also look passingly like a regular pair of sunglasses.
I'd happily wear something that looks like these, but it's the resolution problem. Its really gotta have true 1024x768 (not downsampled to 640x480). Thats whats so annoying - there's a few companies out there making stuff I could possibly see myself wearing, but its just too low res, and although the form factors seem to be getting better, the resolutions arent. Seems like manufacturers only envision people using wearable displays for watching video - not actually doing text oriented or high resolution graphics work.
I have the money to spend on a good display. The problem is there are no solutions in the market that I see that meet the specs I posted in the original question, that don't look like some kind of helmet.
Every single head mounted display I have seen has been ludicrous in size, or pitiful in resolution. I want an HMD that looks passably like a normal pair of sunglasses, with decent resolution - 1280x1024 minimum, and a watchable refresh rate - ie. greater than 30fps, or ideally 60fps. if you know of any products out there that meet these specs, please post a link. I absolutely want to get into the research side of augmented reality - amongst many other uses for a good display. Your phone may do many things, but I bet you have never written a letter on it or worked on a spreadsheet with any reasonable sort of efficiency. Plug in a decent resolution HMD, and your phone suddenly becomes a far more useful platform. That's why I want a decent HMD - its the missing link. There's plenty of pocket sized pc power out there - PDA's and mobile phones have plenty of CPU for a lot of tasks, but a good portable display is the big missing piece that stops them ever being usable for anything other than hey-look-at-what-my-phone-can-do type applications.
That's a great example of brick-tied-to-face technology. I wouldn't be caught dead wearing them out doors, and my girlfriend would kill herself laughing if I wore them round the house.
They only do 800x600 too unfortunately.
The cost doesn't worry me - Id happily drop a k or more on something that met my desired specs.
I think it was the sodium chromate reference in the salt shaker a few posts back.
sodium chromate vs salt (sodium chloride),in the salt shaker analogy leads to a chlorine reference in the car analogy. ( chrome bumper vs chlorine or something - yep Its a bit tenuous, but thats the only way I can see it working.
In the case of most hardware under Linux, it's the Linux community supporting the harder, making the community responsible for crappy drivers.
You neglect to take into consideration crappy drivers that are crappy because the manufacturer has not released any information OR drivers for their product (for a particular platform), and the community has had to make do with whatever information they could glean from the hardware. eg. as used to be the case for many video drivers under Linux - the video card makers would provide minimal, if any, information about various low level hardware details for fear of providing dome edge to their competition. This meant the community had to basically reverse engineer a lot of the workings of the card instead of being able to just write a driver according to proper hardware and firmware specs.
I would argue that in this case, although the poorly performing driver may have been written by the community, it is still largely the fault of the hardware company that the drivers do not work as well could be wished, because the hardware company refused to release data about the hardware that would make it possible to write more stable drivers.
Better yet - if the water consumption of all those server rooms is water that is evaporated and lost, why not re-condense it to pure water and sell the stuff back to the utilities? So the waste heat from all those server rooms is actually being used to turn (possibly quite literally) crappy water into stuff that's as pure as rain water?
The additional advantage is that with that system, it's actually good if the server rooms are burning throuhg more water - it's producing even more pure water, and the airconditioners dont have to re-cool the hot pure water all the way back down again - they can just take in more new crappy water.
The reason they are hiring is because the conditions suck so badly. If they actually paid a decent wage, then they would have no problems getting more people. This is actually very similar to how you get companies whining here that they need more immigrant workers. It's not a shortage of people able to do the job - its a shortage of people willing to do the job for the amount they want to pay.
I only lived there a few years but I did do a lot of driving around Europe and the UK. I'm pretty sure there's nowhere in Europe where fuel is sold by the gallon - it's all in litres, even in the UK where they still sign the roads in M/h rather than Km/h. I'd say its a pretty safe bet that the parent poster already did that calc using US gallons.
I dont know about boost, but I have been using the SDL libraries for a stupid 3d multiplayer asteroid game I have been hacking around with. It's using opengl, threadding and networking, all through the SDL libraries.
There's exactly one #ifdef used so far, to use getcwd() which is defined in windows instead of .
The exact same source files compile and run on Windows and Linux. I have yet to try it on a mac, but from what I have seen so far it should be no problem.
As for IDE, I have been compling it using GCC and the code::blocks gui on Linux, and visual studio on windows, with a separate project file/solution for each.
Slashdot Mirror
This is thinking along the right lines. For years I thought it would be cool to have some sort of tablet PC so I could surf the net on my couch.
After getting a 42" plasma screen a couple of weeks ago, I now realise the solution I always needed was to have an old PC hooked up to a huge screen and just have a remote keyboard & mouse. Works great - you can watch tv, movies etc and also surf the net at a lovely 1920x1280 resolution. Total cost: About $AU 1700 for a 42" plasma, + $100 bucks for a logitech wireless keyboard & mouse with 10M range (about 30 ft) + one old PC I already had lying around. Well worth it, and way better than any hand held solution. It's just not feasible to surf today's web content on a pocket PC.
I think you hit the nail exactly on the head.
The amount of network traffic needed by n number of players in the same area goes up at a squared rate ie. on average, if you have 10 people that are all able to see each other, the servers have to send out 10 movement messages (one to each player) for every player that moves.
so if every player does something once a second, this is 10x10 messages per second if there are 10 nearby players.
if there are 20, it would be 20x20 = 400,
Imagine how much more traffic the server would have to deal with if you could get half of a realm's population turning up all at once in some location from wherever they were, when something interesting happens there suddenly? If there were 2000 people on the server, and half of them all suddenly decided there was something cool going on at point X, first the server would have to tell anyone in the immediate region where you were teleporting from that you had vanished.
Then there would be a massive amount of updates for each player as they teleport to the new region, because the player would have to get a complete set of info about all the other players in that area. Finally, the server would then have to be sending Next, the server would have to and would then have to be dishing out 1000 messages per person in the area, ie, 1000000 messages alltogether, every time everyone took a step or changed direction.
By having slower travel times, it means that at any given moment, a fairly large chunk of the world's population will be distributed in sparsely populated areas, so there are much fewer messages to send to nearby players as they move to their destinations, because each player would only be near say, 3 or 4 others - and often not near anyone at all.
Secondly, because it is a bit of a hastle to go somewhere, there is much less tendancy for as many people to suddenly congregate at a particular point when something happens there.
This is also a good reason to have cities or regions that are specialising in different things - if you can get everything you possibly need from one place, then this will also tend to make players bunch up in htat cities instead of going to different cities, and cause high load. This was done to a certain extent in UO, but I think WoW deals with the problem better with the taxis because it also gives you a chance to go grab a cup of coffee or a bio break and mabey actually wrench yourself away from your chair for a quick leg stretch while in transit.
Good thing that that we don't have to use fossil fuels to supply the power needed to keep that big ball of plasma and nuclear goodness hot then isn't it?
I am not a physicist, but I am pretty sure the Carnot efficiency cycle is the theoretical limit of a heat engine - ie. one that works by extracting work from expanding gases.
Last time I checked, Solar was a means of converting electromagnetic radiation into electricity, so the Carnot cycle does not apply to it - even if you are driving it from say, a burning chunk of magnesium or something. Instead, it is the band-gap of the materials used in the P-N junction that limits the wavelengths that are absorbed. If you had a light source perfectly matched to the band-gaps of the materials in your solar panel, the efficiency would be much higher - and a lot of the current solar research is to do with layering different materials so that there is a wider range of band-gaps in the materials so that more wavelengths are absorbed and converted to electricity.
I would be interested to know what sort of effect applying this sort of laser surface treatment would have to the efficiency of both LED's and solar panels, actually. Any physicists like to comment?
vi or emacs is great if you are writing a small application, or making a few small changes, but time is money, and programmers time is not cheap.
A good IDE can save heaps of time, and in my opinion, automatic code completion is essential in any good programming environment, as it substantially reduces the time it takes looking up exactly what that object's field name was, or what the parameters are for a particular function. It is also extremely useful if there is a way to jump straight to the line of code where a particular variable or class is defined.
Code::blocks is a fairly good IDE that can do this, and has a nice clean interface, and is my IDE of choice when I am working on C++ projects on Linux ( I tried kdevelop, eclipse, and a few others before settling on code::blocks) , but unfortunately it is still not quite as good as Visual Studio on windows, and even though you have to shell out about 1200 bucks for the professional edition here in Aus, if you are getting paid to write software, then that cost is well worth the investment considering the thousands of hours a large project takes to write. I write programs for a living, and really don't want to be spending all my time just trying to navigate around my code, or trying to use a command line based build environment and debugger which is not tightly integrated with the IDE.
If you are just writing stuff on the side, and it is not your day job, then check out code::blocks.
If you are getting paid for your time, then do your client a favour and get the right tools for the job, and spend the money on a tool which allows you to make the most efficient use of your time. Claim the cost of the tool on your tax.
If you use linux you might want to check out http://ktechlab.org/. It is a circuit simulator for PIC micro-controllers, where you can write your PIC code in either C, a flowcharting language, or in assembler. You can attach circuit elements to the PIC like potentiometers, resistors, capacitors, and other components to build a circuit around the PIC to test the software you have written even before you go anywhere near hardware. It also supports several PIC burners too, aparently - for when its time to produce real hardware from your simulation.
Ktechlab does have a few bugs here an there, but still complete enough to be useful. Apparently the project has recently changed hands and is getting under way again after a couple of years of stagnation.
It's a fantastic tool - way easier to use than any of the other circuit simulators I have used, although the version in the Ubuntu 8.10 repository is broken (crashes when placing a component), and I had to download a version directly from sourceforge.
I have spent quite a bit of time searching for the right display - If you re-read my original post, you will note that I am looking for something that has at least a passing resemblance to a pair of sunglasses - slim shades or even (but preferably not) a Geordie style visor. All the high resolution ones you have posted look like a brick-in-face, and I wouldn't be caught dead wearing them on a train or out from behind locked doors. Would you? The lower resolution one is getting there, but still only 800x600, and one eye only. I think it would be maddening looking at a display with only one eye for any extended period of time.
If it needs a strap to hold it on my head, it's too damn big.
There are now displays that have an acceptable appearance and are semi-transparent, much like the example product I linked in my original post, which seems great, except it's too low resolution, being only 640x480, and I think its even stereo.
I have pretty much given up hope of finding them and was sort of hoping that if there were enough people that wanted something like I specified, and enough people asked the manufacturers for what we want (wearables suitable for use as computer displays) instead of what they seem to be producing (ie. wearables that are really aimed at playing movies) then we might eventually get what we want.
Thanks for taking the time to add all those links, but unfortunately I have seen them all before.
You nailed it exactly.
I had a PDA for a while, but the teeny screen was always disappointing.
I am sure when headphones first came out they looked like a couple of speakers duct taped on your head and sounded awful. They progressed to smaller and smaller units until you get the nice slim ubiquitous ear bud headphones of today, with acceptable (if not perfect) audio quality too. Now it's time the same thing was done for our visual field.
Find me those glasses, and I'll write you one.
The first HMD I saw was a sony glasstron while I was in Japan , much like you describe. It was good, but too bulky and too low rez. That was about 12 years ago. I have been looking for a decent display ever since, and can't believe that despite the fact there has been two orders of magnitude change in CPU power and transistor density, the LCD's for wearable displays have hardly even doubled in resolution.
The reason I would like them semi-transparent is so that you do have local environment awarenes still, and can flick your eyes into the distance every now and then to keep the muscles working through multiple focal depths. (I code all day, and when I can,I sit next to a window so I can focus in the distance every 20 min, or when thinking. I am convinced it's the only reason I still have better than 20/20 vision after decades of programming)
The reason I want them high res is because I want to be able to use them for textual based apps ( think suring the web, doing a bit of coding, all the stuff you usually do on a PC, but having the flexibility of being able to do it anywhere.)
The reason I want them to be capable of at least 30fps ( and preferably 60fps as LCD monitors typically are) is because of that motion lag you so correctly pointed out.
The reason I want them to not look ludicrous is because I actually want to wear them, sometimes out doors too!
Wow, they must be etching the glass substrate by hand with the Hope diamond for that price.
It still looks like a brick tied to your face, too.
Nice rez though.
You have pretty much nailed my list of priorities - though wouldn't you count high resolution as part of image quality?
I know it is practically impossible to achieve a wide fov with an unobtrusive system, and actually wide fov is not that critical to me - given a choice between an unobtrusive high resolution field of view that was equivalent to whatever a 24" monitor on a desk 2 feet away is, (20 to 30 degrees?) and a low resolution 120 degree field of view, I would definitely be choosing the former.
Expense wise, well I would spend up to $1000 or so I guess - perhaps $1500, if the product was usable for long periods of time just as a regular monitor is without eye strain.
Seems like there are a lot of problems to solve still - I just find it frustrating that there seem to be products out there that seem so close to my requirements, only lacking in resolution, and probably refresh rate.
I don't wear glasses, despite years at a screen since the Apple //e appeared on the scene, and years of reading books in the dark before that, so I have never had to worry about whether you could wear glasses with some particular hmd in my quest for the perfect one, but I suppose that also adds another whole level of complexity to trying to bring a product to as wide a market as possible.
Is there really such a passive price difference between the LCD's they use in these systems that do 640x480 compared to one that could do 1024x768 though?
I could live with it still looking about the same apparent size and distance as a 24" monitor sitting on my desk, but I just don't want to wear a device that looks too out of place on a train or in public, hence the requirement for them to also look passingly like a regular pair of sunglasses.
I'd happily wear something that looks like these, but it's the resolution problem.
Its really gotta have true 1024x768 (not downsampled to 640x480). Thats whats so annoying - there's a few companies out there making stuff I could possibly see myself wearing, but its just too low res, and although the form factors seem to be getting better, the resolutions arent. Seems like manufacturers only envision people using wearable displays for watching video - not actually doing text oriented or high resolution graphics work.
Good overviews are here: http://www.stereo3d.com/hmd.htm http://www.est-kl.com/aufbau_general/index_hard_hmd.html?http://www.est-kl.com/hardware/hmd/kaiser/pv4050stm.html
Nice try, but sorry,
All the above fail under the Looks-like-a-brick-tied-to-my-face criteria.
I have the money to spend on a good display. The problem is there are no solutions in the market that I see that meet the specs I posted in the original question, that don't look like some kind of helmet.
Every single head mounted display I have seen has been ludicrous in size, or pitiful in resolution.
I want an HMD that looks passably like a normal pair of sunglasses, with decent resolution - 1280x1024 minimum, and a watchable refresh rate - ie. greater than 30fps, or ideally 60fps. if you know of any products out there that meet these specs, please post a link.
I absolutely want to get into the research side of augmented reality - amongst many other uses for a good display. Your phone may do many things, but I bet you have never written a letter on it or worked on a spreadsheet with any reasonable sort of efficiency. Plug in a decent resolution HMD, and your phone suddenly becomes a far more useful platform. That's why I want a decent HMD - its the missing link. There's plenty of pocket sized pc power out there - PDA's and mobile phones have plenty of CPU for a lot of tasks, but a good portable display is the big missing piece that stops them ever being usable for anything other than hey-look-at-what-my-phone-can-do type applications.
That's a great example of brick-tied-to-face technology. I wouldn't be caught dead wearing them out doors, and my girlfriend would kill herself laughing if I wore them round the house.
They only do 800x600 too unfortunately.
The cost doesn't worry me - Id happily drop a k or more on something that met my desired specs.
Crap is a byproduct of animal life too - and plants love that too - just like carbon dioxide.
I think you can safely argue that it is also a pollutant though, and not something we should just be dumping into the environment all over the place.
I think it was the sodium chromate reference in the salt shaker a few posts back.
sodium chromate vs salt (sodium chloride),in the salt shaker analogy leads to a chlorine reference in the car analogy. ( chrome bumper vs chlorine or something - yep Its a bit tenuous, but thats the only way I can see it working.
In the case of most hardware under Linux, it's the Linux community supporting the harder, making the community responsible for crappy drivers.
You neglect to take into consideration crappy drivers that are crappy because the manufacturer has not released any information OR drivers for their product (for a particular platform), and the community has had to make do with whatever information they could glean from the hardware. eg. as used to be the case for many video drivers under Linux - the video card makers would provide minimal, if any, information about various low level hardware details for fear of providing dome edge to their competition. This meant the community had to basically reverse engineer a lot of the workings of the card instead of being able to just write a driver according to proper hardware and firmware specs.
I would argue that in this case, although the poorly performing driver may have been written by the community, it is still largely the fault of the hardware company that the drivers do not work as well could be wished, because the hardware company refused to release data about the hardware that would make it possible to write more stable drivers.
They should use closed circuit cooling system.
Better yet - if the water consumption of all those server rooms is water that is evaporated and lost, why not re-condense it to pure water and sell the stuff back to the utilities? So the waste heat from all those server rooms is actually being used to turn (possibly quite literally) crappy water into stuff that's as pure as rain water?
The additional advantage is that with that system, it's actually good if the server rooms are burning throuhg more water - it's producing even more pure water, and the airconditioners dont have to re-cool the hot pure water all the way back down again - they can just take in more new crappy water.
What's wrong with a train? It will get you there efficiently and more cheaply than driving.
The reason they are hiring is because the conditions suck so badly. If they actually paid a decent wage, then they would have no problems getting more people. This is actually very similar to how you get companies whining here that they need more immigrant workers. It's not a shortage of people able to do the job - its a shortage of people willing to do the job for the amount they want to pay.
I only lived there a few years but I did do a lot of driving around Europe and the UK. I'm pretty sure there's nowhere in Europe where fuel is sold by the gallon - it's all in litres, even in the UK where they still sign the roads in M/h rather than Km/h. I'd say its a pretty safe bet that the parent poster already did that calc using US gallons.
I dont know about boost, but I have been using the SDL libraries for a stupid 3d multiplayer asteroid game I have been hacking around with. It's using opengl, threadding and networking, all through the SDL libraries.
There's exactly one #ifdef used so far, to use getcwd() which is defined in windows instead of .
The exact same source files compile and run on Windows and Linux. I have yet to try it on a mac, but from what I have seen so far it should be no problem.
As for IDE, I have been compling it using GCC and the code::blocks gui on Linux, and visual studio on windows, with a separate project file/solution for each.
Works great so far.