Bootup time is a concern for the Celeron based Eees because sleep sucks so much power, but for anything else (including the Atom-based Eees) I just can't see any need to boot regularly enough for 30 seconds to matter. Just use sleep. I've never heard of a distribution which can boot to any desktop on any machine faster than Ubuntu wakes from sleep on my Eee (6 seconds, which feels glacial if your last laptop was a Mac).
I'm the same, I've got Xubuntu and compiz running and it's perfect, get some minimal window decorations and you're sorted.
That strikes me as an odd combination. I thought always thought XFCE and Compiz were aimed at totally different sets of people. Xfce is a desktop environment for people who want something light and fast, whereas Compiz is a window manager for people who are like pretty visual effects and are willing to trade some performance. Is Compiz faster and/or more efficient than Xfwm, or do you just happen to prefer Xfce for reasons other than efficiency?
I find Gnome runs fine on my Eee 1000 (running stock Ubuntu with the array.org kernel), except the panel doesn't always hide when it should.
I'm always amazed at people who can get seemingly endless enjoyment from beating people up and crashing cars in games like GTA. I'm not saying it's no fun at all, I've played three GTAs and completed two of them, and all were great fun and of course the first thing I did was check out the ragdoll physics by driving down the pavement (sidewalk) and crash into things to wreck my car. But just doing that that gets old, fast. There's no skill or challenge, so there is no satisfaction. There is a lot more to GTA than mowing down old ladies.
Which part of "flight of several helicopters" did you not understand? And there are two planes they use as Air Force One (plus an old one they keep as a spare) and they both have "I'm the fucking president, feel my power bitches" written on the side.
They might know where he is when he's home, but it's all about subtlety when he's out and about. What about when he sneaks out in a secret motorcade with only 923 vehicles with flashing lights followed by dozens of members of the press with cellphones up the wazoo who track his every move 24/7? Or when he sneaks out in a flight of several helicopters with highly obvious markings? Or a Boeing 747 with an obvious paint-job and "I'm the fucking president, feel my power bitches" written on the side? Nobody can track him then.
I don't even see what's so coincidental about the timing, what's so special that now, over 2 years after China did it would be a good time for Russia to have a pop at it again too? Is there something special about around 2 years and 3 weeks later that allows it to be defined as coincidental?
You conveniently forgot the USA's test. The Chinese ASAT test was January 2007. The USA's ASAT test was Feburary 2008. This happened February 2009. Give or take a few weeks, that's a one year gap between each. It most certainly could be coincidence, but the one year gap between the Chinese and American tests demonstrates that those are the timescales on which these kinds of international dick swinging contests take place.
If you don't mind the really crappy bandwidth and the requirement for huge antennas with powerful transmitters SW does indeed work for long-range communications. You could saturate the entire SW spectrum with the data from one 100Mbps NIC. If you want high bandwidth for lots of users with compact antennas and more modest power needs you need to use much higher frequencies, a couple of orders of magnitude higher. Those higher frequencies don't propagate well around the globe or through/around structures or terrain. You either put lots of masts everywhere, which isn't viable in remote places, or you put your "masts" in space where it's easy to get get LOS.
It just begs a question, for WHOM REALLY is Windows and its software a limiting factor? It looks now as if the only people who just would die from not having access to Windows are those who put their whole commerce on it and people who sit and play the latest games.
People who don't understand computers. They get a Windows-only CD with their new broadband provider; we know you don't need it, but they don't know how to enter the settings manually. They get a CD with their new digital camera; we know you just need to stick the memory card in your card reader and it'll just work, but they don't. They get a new printer with a Windows- and Mac-only CD; we know if you just try it there's a good chance it'll work, but they don't. They put the CD in, it doesn't work and they think that means their new toy doesn't work with this crappy non-standard OS. They phone their broadband provider/widget manufacturer for help and they are told Linux isn't supported. Their local geek tells them to use the Ubuntu forums to ask for help, where people start babbling on about typing in commands and their eyes glaze over and they take their "broken" computer back for a refund.
It is a chicken-and-egg situation. Peripheral, software and service suppliers won't support Linux till there are enough people using it to make it worth supporting. Normal people won't use it till they can go to a store and at least see a Linux logo on something and know it will work with their system.
I run (non-default) Linux on my NSLU2. In the early days, it was a lot of hassle tweaking Makefiles and fiddling with code. But now, I use Debian ARM on an NSLU2 and my experiences don't match your at all. It's totally painless. I don't need to recompile stuff myself and the repositories have everything I need. Even the kernel gets updated and flashed automagically. Perhaps you should look again, I've not had to touch gcc on my NSLU2 or go through the pain of getting a cross-compiler working for a long time. The NSLU2 is an officially supported platform for Debian ARM, which has come a long way in the past couple of years. As of today, there are 17354 normal binary packages in the repositories for Etch. There are probably even more for Lenny.
True, most filesystems these days don't become corrupt if you just pull the plug, but that doesn't mean they won't lose your data. Most filesystems these days by default only journal filesystem metadata, they don't journal your data. Pull the plug half-way through re-writing a file and your data is garbage whether you're using NTFS, ext3 (with default settings) or FAT16.
Being dead doesn't sound too bad to me. The process of dying almost always sucks and I don't want to be dead, but once I am dead I can guarantee you I won't give a shit about it.
This was bound to happen and will happen again. The interesting question is how come they didn't maneuver one of them out of the way. I don't know if 22675 is an active payload that still has power but Iridium33 certainly has the capability of moving. This one was avoidable.
I find it a little suspicious myself. China shows off satellite takedown capability. Not long after that the USA shows off satellite takedown capability. Not long after that a "dead" Russian satellite has an "accident" and takes down a satellite. It may just be that Iridium got a warning and played it too close to save fuel or because an insurance payout would make for a handy cash injection, but the timing and the fact that it involves a satellite from the only superpower not to have demonstrated a satellite takedown capability is quite some coincidence.
If Firefox hit 95% market share the proportion of net users using AdBlock would undoubtedly be high enough that a lot of sites would move to some plugin or other for their content - Flash, Silverlight or something new, most likely one with DRM. Just as software pirates brought about DRM for games, music and movies, ad-blockers ("web pirates" seems an appropriate term - they take stuff without paying for it) will bring about DRM for the web. And it'll suck. And the people who bitch the loudest about it will be the ones who caused it. So I really hope Firefox doesn't hit 95% market share or the web as we know it will be dead.
Ha ha ha. So, tell me how many applications for the most popular PC OS on the planet are built with gcc. I'll give you a clue: it's not very many compared to those built with $visual_studio's_compiler or $the_.net_compiler.
I imagine Mozilla now have lots of superfluous executives, they need them to blow all the cash they got from Google. I just hope they don't fuck up the browser in the process.
SLC NAND works like that, but MLC NAND (or rather the controller) does have to detect and correct multi-bit errors. There are 2 (or more) bits per cell so errors tend to affect more than one bit. Read errors are frequent enough that failing to retrieve data on single-cell errors isn't generally considered acceptable.
The Cray and Doom approximations weren't probabilistic, they were just not as accurate as they could be. Given the same input they would always give the same output.
Doom's inverse square root function gives me the horn - the way it manipulates the bits in a float to get the initial approximation is just beautiful.
Detecting and correcting single bit errors isn't too computationally expensive and you can do it easily in hardware, but to detect and correct multi-bit errors you need a lot more horsepower.
You already use this technology. In fact I remember encountering this for the first time when writing an AppleScript program in sixth grade to help a room full of students add up Giant Eagle receipts. I was very embarrassed to have no explanation for someone suddenly getting $54.2300000001 on the screen (looking back, our computer teacher probably should have been able to offer us an explanation.)
That's not probabilistic computing, that's an artifact of computing in base-2 with limited precision and displaying the output in decimal. It's entirely deterministic and reproducible.
For working with money use integer cents/pence/whatever and have rules for who gets the remainder. If you do use floating point binary representations, at least round the output for display.
The UNIX date is a signed integer - you can have negative values and go back to 1901 with it. The reason it only covers 130 years or so is that when it was invented memory was scarce. The 64-bit version of UNIX time with the same 00:00 1st Jan 1970 epoch goes back around 20 times the age of the universe and forward to a couple of hundred billions years after the Sun has died. Is that good enough for you?
Anyway, the fact that the operating system's internal time representation has a limited range doesn't mean you can't compute using dates outside the range, you just have to write a little more code (and I really do mean a little - probably well under 100 lines, even with all the changing calendar crap).
MS may not have got it wrong in Windows, but their dates in Excel are wrong. IIRC that was for compatibility with Lotus 1-2-3. That anomaly may even have made it into OOXML. Old bugs never die in MS land.
Slashdot Mirror
Bootup time is a concern for the Celeron based Eees because sleep sucks so much power, but for anything else (including the Atom-based Eees) I just can't see any need to boot regularly enough for 30 seconds to matter. Just use sleep. I've never heard of a distribution which can boot to any desktop on any machine faster than Ubuntu wakes from sleep on my Eee (6 seconds, which feels glacial if your last laptop was a Mac).
I'm the same, I've got Xubuntu and compiz running and it's perfect, get some minimal window decorations and you're sorted.
That strikes me as an odd combination. I thought always thought XFCE and Compiz were aimed at totally different sets of people. Xfce is a desktop environment for people who want something light and fast, whereas Compiz is a window manager for people who are like pretty visual effects and are willing to trade some performance. Is Compiz faster and/or more efficient than Xfwm, or do you just happen to prefer Xfce for reasons other than efficiency?
I find Gnome runs fine on my Eee 1000 (running stock Ubuntu with the array.org kernel), except the panel doesn't always hide when it should.
I'm always amazed at people who can get seemingly endless enjoyment from beating people up and crashing cars in games like GTA. I'm not saying it's no fun at all, I've played three GTAs and completed two of them, and all were great fun and of course the first thing I did was check out the ragdoll physics by driving down the pavement (sidewalk) and crash into things to wreck my car. But just doing that that gets old, fast. There's no skill or challenge, so there is no satisfaction. There is a lot more to GTA than mowing down old ladies.
Which part of "flight of several helicopters" did you not understand? And there are two planes they use as Air Force One (plus an old one they keep as a spare) and they both have "I'm the fucking president, feel my power bitches" written on the side.
The tracking devices emit frequencies of a few tens of THz and are called "secret service agents".
They might know where he is when he's home, but it's all about subtlety when he's out and about. What about when he sneaks out in a secret motorcade with only 923 vehicles with flashing lights followed by dozens of members of the press with cellphones up the wazoo who track his every move 24/7? Or when he sneaks out in a flight of several helicopters with highly obvious markings? Or a Boeing 747 with an obvious paint-job and "I'm the fucking president, feel my power bitches" written on the side? Nobody can track him then.
I don't even see what's so coincidental about the timing, what's so special that now, over 2 years after China did it would be a good time for Russia to have a pop at it again too? Is there something special about around 2 years and 3 weeks later that allows it to be defined as coincidental?
You conveniently forgot the USA's test. The Chinese ASAT test was January 2007. The USA's ASAT test was Feburary 2008. This happened February 2009. Give or take a few weeks, that's a one year gap between each. It most certainly could be coincidence, but the one year gap between the Chinese and American tests demonstrates that those are the timescales on which these kinds of international dick swinging contests take place.
If you don't mind the really crappy bandwidth and the requirement for huge antennas with powerful transmitters SW does indeed work for long-range communications. You could saturate the entire SW spectrum with the data from one 100Mbps NIC. If you want high bandwidth for lots of users with compact antennas and more modest power needs you need to use much higher frequencies, a couple of orders of magnitude higher. Those higher frequencies don't propagate well around the globe or through/around structures or terrain. You either put lots of masts everywhere, which isn't viable in remote places, or you put your "masts" in space where it's easy to get get LOS.
It just begs a question, for WHOM REALLY is Windows and its software a limiting factor? It looks now as if the only people who just would die from not having access to Windows are those who put their whole commerce on it and people who sit and play the latest games.
People who don't understand computers. They get a Windows-only CD with their new broadband provider; we know you don't need it, but they don't know how to enter the settings manually. They get a CD with their new digital camera; we know you just need to stick the memory card in your card reader and it'll just work, but they don't. They get a new printer with a Windows- and Mac-only CD; we know if you just try it there's a good chance it'll work, but they don't. They put the CD in, it doesn't work and they think that means their new toy doesn't work with this crappy non-standard OS. They phone their broadband provider/widget manufacturer for help and they are told Linux isn't supported. Their local geek tells them to use the Ubuntu forums to ask for help, where people start babbling on about typing in commands and their eyes glaze over and they take their "broken" computer back for a refund.
It is a chicken-and-egg situation. Peripheral, software and service suppliers won't support Linux till there are enough people using it to make it worth supporting. Normal people won't use it till they can go to a store and at least see a Linux logo on something and know it will work with their system.
I run (non-default) Linux on my NSLU2. In the early days, it was a lot of hassle tweaking Makefiles and fiddling with code. But now, I use Debian ARM on an NSLU2 and my experiences don't match your at all. It's totally painless. I don't need to recompile stuff myself and the repositories have everything I need. Even the kernel gets updated and flashed automagically. Perhaps you should look again, I've not had to touch gcc on my NSLU2 or go through the pain of getting a cross-compiler working for a long time. The NSLU2 is an officially supported platform for Debian ARM, which has come a long way in the past couple of years. As of today, there are 17354 normal binary packages in the repositories for Etch. There are probably even more for Lenny.
True, most filesystems these days don't become corrupt if you just pull the plug, but that doesn't mean they won't lose your data. Most filesystems these days by default only journal filesystem metadata, they don't journal your data. Pull the plug half-way through re-writing a file and your data is garbage whether you're using NTFS, ext3 (with default settings) or FAT16.
Being dead doesn't sound too bad to me. The process of dying almost always sucks and I don't want to be dead, but once I am dead I can guarantee you I won't give a shit about it.
Good luck using your terrestrial packet switched networks in the middle of the Atlantic or in a remote valley in the Himalayas.
This was bound to happen and will happen again. The interesting question is how come they didn't maneuver one of them out of the way. I don't know if 22675 is an active payload that still has power but Iridium33 certainly has the capability of moving. This one was avoidable.
I find it a little suspicious myself. China shows off satellite takedown capability. Not long after that the USA shows off satellite takedown capability. Not long after that a "dead" Russian satellite has an "accident" and takes down a satellite. It may just be that Iridium got a warning and played it too close to save fuel or because an insurance payout would make for a handy cash injection, but the timing and the fact that it involves a satellite from the only superpower not to have demonstrated a satellite takedown capability is quite some coincidence.
If Firefox hit 95% market share the proportion of net users using AdBlock would undoubtedly be high enough that a lot of sites would move to some plugin or other for their content - Flash, Silverlight or something new, most likely one with DRM. Just as software pirates brought about DRM for games, music and movies, ad-blockers ("web pirates" seems an appropriate term - they take stuff without paying for it) will bring about DRM for the web. And it'll suck. And the people who bitch the loudest about it will be the ones who caused it. So I really hope Firefox doesn't hit 95% market share or the web as we know it will be dead.
Cue downmods from freetards in 3, 2, 1...
Those are stats from a site for web developers, they're hardly representative of the population as a whole.
Ha ha ha. So, tell me how many applications for the most popular PC OS on the planet are built with gcc. I'll give you a clue: it's not very many compared to those built with $visual_studio's_compiler or $the_.net_compiler.
I imagine Mozilla now have lots of superfluous executives, they need them to blow all the cash they got from Google. I just hope they don't fuck up the browser in the process.
The cost of error correcting logic is low compared to acres of silicon for storing data, but not so low compared to the silicon used in, say, an ALU.
SLC NAND works like that, but MLC NAND (or rather the controller) does have to detect and correct multi-bit errors. There are 2 (or more) bits per cell so errors tend to affect more than one bit. Read errors are frequent enough that failing to retrieve data on single-cell errors isn't generally considered acceptable.
The Cray and Doom approximations weren't probabilistic, they were just not as accurate as they could be. Given the same input they would always give the same output.
Doom's inverse square root function gives me the horn - the way it manipulates the bits in a float to get the initial approximation is just beautiful.
Detecting and correcting single bit errors isn't too computationally expensive and you can do it easily in hardware, but to detect and correct multi-bit errors you need a lot more horsepower.
You already use this technology. In fact I remember encountering this for the first time when writing an AppleScript program in sixth grade to help a room full of students add up Giant Eagle receipts. I was very embarrassed to have no explanation for someone suddenly getting $54.2300000001 on the screen (looking back, our computer teacher probably should have been able to offer us an explanation.)
That's not probabilistic computing, that's an artifact of computing in base-2 with limited precision and displaying the output in decimal. It's entirely deterministic and reproducible.
For working with money use integer cents/pence/whatever and have rules for who gets the remainder. If you do use floating point binary representations, at least round the output for display.
The UNIX date is a signed integer - you can have negative values and go back to 1901 with it. The reason it only covers 130 years or so is that when it was invented memory was scarce. The 64-bit version of UNIX time with the same 00:00 1st Jan 1970 epoch goes back around 20 times the age of the universe and forward to a couple of hundred billions years after the Sun has died. Is that good enough for you?
Anyway, the fact that the operating system's internal time representation has a limited range doesn't mean you can't compute using dates outside the range, you just have to write a little more code (and I really do mean a little - probably well under 100 lines, even with all the changing calendar crap).
MS may not have got it wrong in Windows, but their dates in Excel are wrong. IIRC that was for compatibility with Lotus 1-2-3. That anomaly may even have made it into OOXML. Old bugs never die in MS land.