The good news is that Lindows is built on Debian. And even better news is that the Lindows.com guys didn't rip out the APT tools. Lindows doesn't use them (they use their "Click-N-Run" stuff) but the tools are there.
It is actually possible to upgrade (or "side-grade" if you prefer the term) Lindows to just plain Debian.
Basically, you just edit sources.list to point to a Debian mirror near you. (Lindows has it pointing to the main Debian server; be a good net citizen and change that.) Then "apt-get update". Then blow away all packages that have "lindows" or "xandros" in the name, if you want that pure free-software feeling... or don't bother, if you don't mind a few Lindows packages floating around. "apt-get dist-upgrade", handle any conflicts APT can't suss on its own, and install anything you are missing. If you blow away the lindows* packages and xandros* packages, you will lose LILO and the kernel, so you will need to replace those.
Lindows by default sets up three partitions: a small/boot, a 256 MB swap partition, and the whole rest of the drive as a big ReiserFS partition, mounted as the root partition. I have not yet been able to build a kernel that can deal with the root ReiserFS; I keep getting the error "Unable to open initial console." I believe the problem is that it's trying to mount DevFS while the root partition is still mounted read-only, and I think the solution is to use an initrd (initial ramdisk). The 2.4.20 kernel that comes with Lindows 4 uses an initrd, and it of course works. I need to try building an initrd kernel soon.
There will be an article about this on the Linux Journal website sometime soon... I'm not sure exactly when. I took a Lindows MobilePC and upgraded it to full Debian unstable; it now boots with GRUB and has a GNOME desktop, because that's what I prefer.
It is possible to set up user accounts in Lindows. KUser, the KDE user manager tool, is available (renamed to "User Manager") and you can create users.
It doesn't work perfectly out of the box: you will need to manually add each user to the "dialout" and "dip" groups if you want Kppp to work, and the "Click-N-Run Installer" will ask for the root password each time a user logs in. (The solution to the latter problem is to disable the C-N-R Installer from auto-running).
Once you have created a non-root user, the KDE login manager will run and prompt for user name and password.
The above applies to Lindows 4.0 at least; I haven't really looked at other versions. (I wrote a review of Lindows 4 for Linux Journal.)
Looks pretty slick. I still think I'm happier with my Tungsten T (320x320 screen, 144 MHz processor) and a separate phone.
I'm looking forward to a StarTac-like phone with Bluetooth so I can use the Tungsten T without needing a wire. Meanwhile, I don't really need PDA web surfing that often, and I'm content with what I have now.
One potential problem with the new combination PDA and phones: they are digital-only.
Someday, in the bright happy cheerful future, digital cell phone service will be everywhere. Meanwhile, there are still places where there is only analog phone service, and I actually go to those places.
My battered old StarTac phone has two bands of CDMA, plus a fallback to analog. I can use that phone pretty much anywhere in the USA that has cell phone service at all. The same cannot be said of the new combination ones.
There is a cable I have to connect my PDA to the digital connector on the bottom of my StarTac, and then I can use the StarTac to call up my ISP. This does not require buying any special "data" features, only using my phone minutes. Depending on where I am, my ISP might be a long-distance call, of course (buy my ISP has a toll-free number I can call and use for ten cents per minute).
I think that the combination of a really nice PDA (mine is a Tungsten T) and a really nice phone (battered old StarTac) is better, for my purposes, than the new all-in-one gadgets. It might be different if I lived in a big city and spent all my time there.
I still carry an old StarTac phone. When the battery is low, I can pop it out and pop in a new one. That's nice.
I guess I can understand the new device not having an exposed, swappable battery; they would have had to make the thing bigger. But I hope they have some kind of auxiliary battery. Perhaps something you can connect to the HotSync cradle connector.
When I got a Treo 90, I also got an aux battery that plugged in to the HotSync connector, and uses AAA cells to power or recharge it. For my Tungsten T, I got one that uses AA cells. I'm wondering if there is a charger like this that can charge the new Treo, and whether AAA cells would have enough power to allow you to make phone calls when the main battery is dead.
I wish that someone would make a lithium ion or lithium polymer aux battery, with built-in folding prongs for a wall outlet so you could recharge it anywhere. I use rechargeable NiMH AA cells for my current emergency charger, and I think that a purpose-built aux battery would be more convenient. The NiMH AA cells only give 1.2 V each, rather than 1.5 V, and I think they still have a lot of power in them when the voltage drops off below useful and the Palm stops charging.
Right now, when you plug in an aux battery, the Palm device thinks it is in its cradle, connected to wall power. It would be nice if you could plug in a battery pack, and the Palm knew it was a battery pack and could tell you how much charge is left on that battery pack. (However, that feature is not by itself worth another redesign of the HotSync connector!)
Wow. Ford is buying new computers without floppy disk drives... and Dell is so proud they released a white paper about it!
"Realizing that it was purchasing obsolete technology that consumes space and costs money, Ford Motor Company began evaluating its alternatives with help from Dell."
They should have interviewed Steve Jobs. I'm sure he'd have given them some great quotes about floppy-less PCs.
Maybe Dell will next write a white paper about PCs without PS/2 keyboard or mouse ports!
GNOME doesn't have its own file selection dialog box for file open and file save. Instead, it just uses the standard one in GTK. GTK is developing a new release, GTK 2.4, and that will have a much improved file selection dialog box.
I don't think it makes sense for two big projects (GNOME and GTK) to try to release updates in lockstep. GNOME is ready to release now, and GTK isn't, so that's that.
Also, I'm glad that GNOME doesn't paste a layer over GTK for things like the file selection dialog box. Re-implementing basic features of GTK would just lead to bloat.
If you only get updates all at once, you might have to wait for GNOME 2.6 to get the improved file selection dialog box, but those of us who run Debian unstable or some other incrementally-updated distro will get the new dialog when GTK 2.4 is released.
Meanwhile, Debian unstable already has an improved file selection dialog box, but it isn't the same one that will ship with GTK 2.4. It's a bit nicer than the default GNOME one but I'm still waiting for the new one in GTK 2.4.
The biggest problem for train safety is that a train is hard to stop. It has so much mass that you can't just suddenly decide to stop it. In a perfect world, you would have some warning before you needed to stop it.
So make a much smaller vehicle that can stop quickly, and have that run out ahead of the train! Call it the "point car". Sensors on the point car would watch for an obstruction on the tracks (such as a stalled truck) and would halt the point car quickly; the train would stop more slowly, but it would have enough warning that it could stop before it reached the point car, let alone the obstruction. Also, you could mount a video camera on the front of the point car, and the engineer driving the train could watch a live video feed. A wireless radio link is probably the best way for the point car and the train to communicate.
I'm sure the biggest problem with my idea is that it would cost too much. The point car would need fuel of some sort, and would itself be an expensive piece of equipment, and you would need one for each train. It would be cool if the point car could be driven by electric motors that somehow parasite power off the train, but I don't think any sort of power extension cord would be very practical.
And of course, if India is only now spending the money to put cushions in for engineers to sit on, they won't be the first ones to try point cars.
I don't know much about train crashes -- what fraction of train crashes are preventable with just GPS, and what fraction are not? If the most common problem is a train hitting another train, then GPS on both trains would help a lot. But GPS won't do much good if a truck stalls across the tracks.
Perhaps that DOS extender was playing games with the segment registers?
Modern 32-bit operating systems (including Linux, *BSD, and Windows) all treat the x86 as a flat, 32-bit processor. The segment register features are not used. This discussion is about using the memory management unit to mark certain pages as non-executable.
That's as may be. But the big problem with the Shuttle was that they didn't do incremental development: no prototypes, no test versions, just build version 1.0 and hope it works. The Great Leap Forward theory of spacecraft design.
Before the Shuttle, everything NASA did was incremental, all the way back to the first rockets which were pretty much just tweaks to a proven missile design. What ever made them think they could crank out a perfect design without testing any part of it?
If there is any real enefit in space, then surely no government sponsored handout is needed.
True. But a government handout would provide an incentive that would get us to space sooner. Giving more money to NASA will not get us into space; giving a prize to the first people to build a real spacecraft would get us into space. Those of us in favor of the idea believe that the benefits would be worth a fairly big prize, and note that the prize would not be awarded for anything but working hardware. (Money to NASA will be spent on bureaucrats and studies.)
As a libertarian, I don't think government should be subsidising businesses. However, a real libertarian society would have many fewer taxes, regulations, and general red tape; given that we don't live in such a libertarian society and real companies do have to deal with the red tape, I'm willing to see government prize money for the space program. (And a few other generally beneficial innovations I can think of: a working hydrogen fusion power plant, a cure for cancer, etc.)
I'm only in favor of a prize for meeting a specific goal, with no other strings attached to the prize. For example, a big prize for the first ship to fly to orbit with a 1,000 kilogram payload, and then do it again within the next two weeks, landing safely both times. (And a smaller prize for second place, even, so it won't be "winner take all".)
First you take the tie you're going to dye and tie it. When the tie is tightly tied, you dip the tie in the tie-dye dye! Ooh! When the tie-dyed tie is dried, you take the dye you just applied and set it aside. Re-tie the tie-dyed tie, take another dye, dip the re-tied tie in this dye, too. Take it out, let it dry, untie the tie, and you've got a tie-dyed tie. And a tie's not all you can tie-dye. You can tie-dye a tutu, too! Take the tutu, tie the tutu, dip it in the dye, let it dry like the tie we dyed, now tear the tutu in two. Now you've got two tutus to tie-dye! Take the two ties you tie-dyed, and the two tutus you've torn in two and tied, and dip them in the dye!
If anyone knows where this came from, and where I can buy a CD or something with this, please let me know. (I believe I saw this performed on a TV show called Laugh Tracks or something like that, but it predates that show.)
The tricky part is that I don't think tests done with small rockets will necessarily give you a good idea of how the big rocket will perform.
The key here is that they are doing many cycles of prototyping and testing. No one seriously expects that they can build a tiny model, fly it, and then make a great leap forward to build an operational vehicle. The only organization that crazy is NASA (it's why the space shuttle sucks so bad).
They will build and fly tiny models. Then they will build and fly medium models. Then they will build and fly full-size models, and then they will go for the prize.
If they succeed, they will be able to charge whatever they want for Linux. $700, $1400, or more, per server. That would harm Linux in a very meaningful way. Not only would that be bad for me (cost me money) but it would be horrible for third-world countries and anyone else who likes Linux because it is cheap (cost them money they cannot afford).
You have "rough and free" versus "polished and expensive".
And the "rough" one is getting more and more polished as time goes by, while the polished one isn't getting noticeably more free.
I already like GNOME 2.x better than any other environment; all I need for complete happiness is for the apps to all get finished.
I have successfully moved my wife off a Windows computer and onto a Linux computer with a GNOME 2.x desktop. It's enough like Windows that she didn't require much training.
So, if I had to choose one, I choose GNOME. But there is no way to force such a choice on the free software community, so GNOME and KDE will have to continue to coexist.
There has been speculation that Darl and company wanted to make a big noise, and that IBM or someone would buy SCO to make the noise go away. This would avoid an Enron-like ending.
It quickly became clear that IBM didn't intend to buy SCO, but was (and is) willing to fight SCO forever in court. But perhaps by then they felt they were committed.
The reason for the "do as we say, not as we do" comment is probably because SCO originally described Linux as a bicycle next to the luxury car that is SCO UNIX. Of course, SCO changes what they are saying frequently, so that's old news at this point.
Agreed -- the 24x7 servers[1] are not going to be random old boxes. But not all servers have to be held to that standard. For example, a department-level print server could very well be a battered old desktop running Linux, with a couple of printers hooked up to it. (And for that application, it probably will run 24x7 with no worries until something fails... at which point the IT guys can clone another print server on another battered old computer and printing is back.)
More to the point, if you work in IT and you like Linux, you can set up one of these print servers and show it off to your boss. "It's rock solid and it's cheap." If you like the XServe, you have to convince your boss to drop $2000 on one before you can show it off. It's easier to do pilot projects with Linux.
[1] I used the phrase "enterprise servers" to refer to these, in the parent article. I suggested that the XServe might make a good enterprise server. Actually, thinking about it, the XServe might not be redundant enough, so I'm not sure what niche the XServe really fills.
And the hardware it runs on is essentially free. An IT department can take any extra PC and put it on the net as a Linux server. Let's say all the folks in Department E got shiny new computers to replace their old 300 MHz boxes. Those old 300 MHz boxes can have new life as a server.
So why is Linux also used for enterprise servers, where the XServe would work just fine? I suspect it is because most companies already have a preferred vendor. All the shiny new boxes in Department E came from Dell, and the enterprise servers did too. Unless the business is an Apple shop, already getting computers from Apple, buying an XServe means buying from multiple vendors. And maybe the desktops and the enterprise servers were all bought at once, in a package deal that saved some extra money.
All this is obvious. I moderate Cringely's latest column (-1, Troll).
Maybe it's overkill for you, but there are a few organizations out there who would like to be able to crank out large volumes of safely random number streams. Verisign probably could use this, for example.
But if you would actually read the article you would find that, despite the name, this project doesn't use a lava lamp. They use a CCD inside a light-proof can, so that all you can measure with the CCD is noise fluctuations. Then they hash the resulting data up big-time to remove any patterns.
I don't know how effective your idea will be, but I'm pretty sure it will take more time to get a chunk of chaotic data from a preamp than from a CCD. The CCD has a few million pixel sensors that can fluctuate and you can read them all in parallel. With the audio preamp you will probably get 16 bits per sample, and how many of those will be chaotic? You can just take more samples but that will take more time.
For true portability, just get a computer with a CPU that makes true random numbers by measuring heat. Again this will no doubt take more time than the CCD, but also no doubt less than the audio preamp. At least all the bits will be chaotic.
According to several articles I have read, such as the one on gizmodo.com, the Rio Karma will have USB 2.0 as its native interface; it will also come with a dock that will plug into an Ethernet network.
If you can just use standard file server protocols (NFS or SMB, I don't care) to put files on the Karma, I will buy one. If you have to run some modified jukebox app to move the files, so it can wrap your files in DRM junk, I won't buy one.
Slashdot Mirror
The good news is that Lindows is built on Debian. And even better news is that the Lindows.com guys didn't rip out the APT tools. Lindows doesn't use them (they use their "Click-N-Run" stuff) but the tools are there.
/boot, a 256 MB swap partition, and the whole rest of the drive as a big ReiserFS partition, mounted as the root partition. I have not yet been able to build a kernel that can deal with the root ReiserFS; I keep getting the error "Unable to open initial console." I believe the problem is that it's trying to mount DevFS while the root partition is still mounted read-only, and I think the solution is to use an initrd (initial ramdisk). The 2.4.20 kernel that comes with Lindows 4 uses an initrd, and it of course works. I need to try building an initrd kernel soon.
It is actually possible to upgrade (or "side-grade" if you prefer the term) Lindows to just plain Debian.
Basically, you just edit sources.list to point to a Debian mirror near you. (Lindows has it pointing to the main Debian server; be a good net citizen and change that.) Then "apt-get update". Then blow away all packages that have "lindows" or "xandros" in the name, if you want that pure free-software feeling... or don't bother, if you don't mind a few Lindows packages floating around. "apt-get dist-upgrade", handle any conflicts APT can't suss on its own, and install anything you are missing. If you blow away the lindows* packages and xandros* packages, you will lose LILO and the kernel, so you will need to replace those.
Lindows by default sets up three partitions: a small
There will be an article about this on the Linux Journal website sometime soon... I'm not sure exactly when. I took a Lindows MobilePC and upgraded it to full Debian unstable; it now boots with GRUB and has a GNOME desktop, because that's what I prefer.
steveha
the way lindows runs as root is just wrong
It is possible to set up user accounts in Lindows. KUser, the KDE user manager tool, is available (renamed to "User Manager") and you can create users.
It doesn't work perfectly out of the box: you will need to manually add each user to the "dialout" and "dip" groups if you want Kppp to work, and the "Click-N-Run Installer" will ask for the root password each time a user logs in. (The solution to the latter problem is to disable the C-N-R Installer from auto-running).
Once you have created a non-root user, the KDE login manager will run and prompt for user name and password.
The above applies to Lindows 4.0 at least; I haven't really looked at other versions. (I wrote a review of Lindows 4 for Linux Journal.)
steveha
Odd. My service plan does not charge differently for analog minutes or digital minutes. I'm on Verizon, by the way.
steveha
Looks pretty slick. I still think I'm happier with my Tungsten T (320x320 screen, 144 MHz processor) and a separate phone.
I'm looking forward to a StarTac-like phone with Bluetooth so I can use the Tungsten T without needing a wire. Meanwhile, I don't really need PDA web surfing that often, and I'm content with what I have now.
steveha
One potential problem with the new combination PDA and phones: they are digital-only.
Someday, in the bright happy cheerful future, digital cell phone service will be everywhere. Meanwhile, there are still places where there is only analog phone service, and I actually go to those places.
My battered old StarTac phone has two bands of CDMA, plus a fallback to analog. I can use that phone pretty much anywhere in the USA that has cell phone service at all. The same cannot be said of the new combination ones.
There is a cable I have to connect my PDA to the digital connector on the bottom of my StarTac, and then I can use the StarTac to call up my ISP. This does not require buying any special "data" features, only using my phone minutes. Depending on where I am, my ISP might be a long-distance call, of course (buy my ISP has a toll-free number I can call and use for ten cents per minute).
I think that the combination of a really nice PDA (mine is a Tungsten T) and a really nice phone (battered old StarTac) is better, for my purposes, than the new all-in-one gadgets. It might be different if I lived in a big city and spent all my time there.
steveha
I still carry an old StarTac phone. When the battery is low, I can pop it out and pop in a new one. That's nice.
I guess I can understand the new device not having an exposed, swappable battery; they would have had to make the thing bigger. But I hope they have some kind of auxiliary battery. Perhaps something you can connect to the HotSync cradle connector.
When I got a Treo 90, I also got an aux battery that plugged in to the HotSync connector, and uses AAA cells to power or recharge it. For my Tungsten T, I got one that uses AA cells. I'm wondering if there is a charger like this that can charge the new Treo, and whether AAA cells would have enough power to allow you to make phone calls when the main battery is dead.
I wish that someone would make a lithium ion or lithium polymer aux battery, with built-in folding prongs for a wall outlet so you could recharge it anywhere. I use rechargeable NiMH AA cells for my current emergency charger, and I think that a purpose-built aux battery would be more convenient. The NiMH AA cells only give 1.2 V each, rather than 1.5 V, and I think they still have a lot of power in them when the voltage drops off below useful and the Palm stops charging.
Right now, when you plug in an aux battery, the Palm device thinks it is in its cradle, connected to wall power. It would be nice if you could plug in a battery pack, and the Palm knew it was a battery pack and could tell you how much charge is left on that battery pack. (However, that feature is not by itself worth another redesign of the HotSync connector!)
steveha
Wow. Ford is buying new computers without floppy disk drives... and Dell is so proud they released a white paper about it!
"Realizing that it was purchasing obsolete technology that consumes space and costs money, Ford Motor Company began evaluating its alternatives with help from Dell."
They should have interviewed Steve Jobs. I'm sure he'd have given them some great quotes about floppy-less PCs.
Maybe Dell will next write a white paper about PCs without PS/2 keyboard or mouse ports!
steveha
GNOME doesn't have its own file selection dialog box for file open and file save. Instead, it just uses the standard one in GTK. GTK is developing a new release, GTK 2.4, and that will have a much improved file selection dialog box.
I don't think it makes sense for two big projects (GNOME and GTK) to try to release updates in lockstep. GNOME is ready to release now, and GTK isn't, so that's that.
Also, I'm glad that GNOME doesn't paste a layer over GTK for things like the file selection dialog box. Re-implementing basic features of GTK would just lead to bloat.
If you only get updates all at once, you might have to wait for GNOME 2.6 to get the improved file selection dialog box, but those of us who run Debian unstable or some other incrementally-updated distro will get the new dialog when GTK 2.4 is released.
Meanwhile, Debian unstable already has an improved file selection dialog box, but it isn't the same one that will ship with GTK 2.4. It's a bit nicer than the default GNOME one but I'm still waiting for the new one in GTK 2.4.
steveha
I have an idea for improving train safety.
The biggest problem for train safety is that a train is hard to stop. It has so much mass that you can't just suddenly decide to stop it. In a perfect world, you would have some warning before you needed to stop it.
So make a much smaller vehicle that can stop quickly, and have that run out ahead of the train! Call it the "point car". Sensors on the point car would watch for an obstruction on the tracks (such as a stalled truck) and would halt the point car quickly; the train would stop more slowly, but it would have enough warning that it could stop before it reached the point car, let alone the obstruction. Also, you could mount a video camera on the front of the point car, and the engineer driving the train could watch a live video feed. A wireless radio link is probably the best way for the point car and the train to communicate.
I'm sure the biggest problem with my idea is that it would cost too much. The point car would need fuel of some sort, and would itself be an expensive piece of equipment, and you would need one for each train. It would be cool if the point car could be driven by electric motors that somehow parasite power off the train, but I don't think any sort of power extension cord would be very practical.
And of course, if India is only now spending the money to put cushions in for engineers to sit on, they won't be the first ones to try point cars.
I don't know much about train crashes -- what fraction of train crashes are preventable with just GPS, and what fraction are not? If the most common problem is a train hitting another train, then GPS on both trains would help a lot. But GPS won't do much good if a truck stalls across the tracks.
steveha
Perhaps that DOS extender was playing games with the segment registers?
Modern 32-bit operating systems (including Linux, *BSD, and Windows) all treat the x86 as a flat, 32-bit processor. The segment register features are not used. This discussion is about using the memory management unit to mark certain pages as non-executable.
steveha
The article made me blink three times in confusion.
Apollo missions regularly landed within 2nm of the predicted point
2 nanometers?!?
I guess that's actually "nautical miles".
steveha
That's as may be. But the big problem with the Shuttle was that they didn't do incremental development: no prototypes, no test versions, just build version 1.0 and hope it works. The Great Leap Forward theory of spacecraft design.
Before the Shuttle, everything NASA did was incremental, all the way back to the first rockets which were pretty much just tweaks to a proven missile design. What ever made them think they could crank out a perfect design without testing any part of it?
steveha
If there is any real enefit in space, then surely no government sponsored handout is needed.
True. But a government handout would provide an incentive that would get us to space sooner. Giving more money to NASA will not get us into space; giving a prize to the first people to build a real spacecraft would get us into space. Those of us in favor of the idea believe that the benefits would be worth a fairly big prize, and note that the prize would not be awarded for anything but working hardware. (Money to NASA will be spent on bureaucrats and studies.)
As a libertarian, I don't think government should be subsidising businesses. However, a real libertarian society would have many fewer taxes, regulations, and general red tape; given that we don't live in such a libertarian society and real companies do have to deal with the red tape, I'm willing to see government prize money for the space program. (And a few other generally beneficial innovations I can think of: a working hydrogen fusion power plant, a cure for cancer, etc.)
I'm only in favor of a prize for meeting a specific goal, with no other strings attached to the prize. For example, a big prize for the first ship to fly to orbit with a 1,000 kilogram payload, and then do it again within the next two weeks, landing safely both times. (And a smaller prize for second place, even, so it won't be "winner take all".)
steveha
http://www.comedy-zone.net/cgi-bin/ikonboard/topi
If anyone knows where this came from, and where I can buy a CD or something with this, please let me know. (I believe I saw this performed on a TV show called Laugh Tracks or something like that, but it predates that show.)
steveha
The tricky part is that I don't think tests done with small rockets will necessarily give you a good idea of how the big rocket will perform.
The key here is that they are doing many cycles of prototyping and testing. No one seriously expects that they can build a tiny model, fly it, and then make a great leap forward to build an operational vehicle. The only organization that crazy is NASA (it's why the space shuttle sucks so bad).
They will build and fly tiny models. Then they will build and fly medium models. Then they will build and fly full-size models, and then they will go for the prize.
Build and fly. Over and over. That's the key.
steveha
SCO has not harmed Linux in any meaningful way!
Not for lack of trying.
If they succeed, they will be able to charge whatever they want for Linux. $700, $1400, or more, per server. That would harm Linux in a very meaningful way. Not only would that be bad for me (cost me money) but it would be horrible for third-world countries and anyone else who likes Linux because it is cheap (cost them money they cannot afford).
steveha
You have "rough and free" versus "polished and expensive".
And the "rough" one is getting more and more polished as time goes by, while the polished one isn't getting noticeably more free.
I already like GNOME 2.x better than any other environment; all I need for complete happiness is for the apps to all get finished.
I have successfully moved my wife off a Windows computer and onto a Linux computer with a GNOME 2.x desktop. It's enough like Windows that she didn't require much training.
So, if I had to choose one, I choose GNOME. But there is no way to force such a choice on the free software community, so GNOME and KDE will have to continue to coexist.
steveha
There has been speculation that Darl and company wanted to make a big noise, and that IBM or someone would buy SCO to make the noise go away. This would avoid an Enron-like ending.
It quickly became clear that IBM didn't intend to buy SCO, but was (and is) willing to fight SCO forever in court. But perhaps by then they felt they were committed.
steveha
The reason for the "do as we say, not as we do" comment is probably because SCO originally described Linux as a bicycle next to the luxury car that is SCO UNIX. Of course, SCO changes what they are saying frequently, so that's old news at this point.
84. Prior to IBM's involvement, Linux was the software equivalent of a bicycle. UNIX was the software equivalent of a luxury car.
steveha
Agreed -- the 24x7 servers[1] are not going to be random old boxes. But not all servers have to be held to that standard. For example, a department-level print server could very well be a battered old desktop running Linux, with a couple of printers hooked up to it. (And for that application, it probably will run 24x7 with no worries until something fails... at which point the IT guys can clone another print server on another battered old computer and printing is back.)
More to the point, if you work in IT and you like Linux, you can set up one of these print servers and show it off to your boss. "It's rock solid and it's cheap." If you like the XServe, you have to convince your boss to drop $2000 on one before you can show it off. It's easier to do pilot projects with Linux.
[1] I used the phrase "enterprise servers" to refer to these, in the parent article. I suggested that the XServe might make a good enterprise server. Actually, thinking about it, the XServe might not be redundant enough, so I'm not sure what niche the XServe really fills.
steveha
Linux is far cheaper
And the hardware it runs on is essentially free. An IT department can take any extra PC and put it on the net as a Linux server. Let's say all the folks in Department E got shiny new computers to replace their old 300 MHz boxes. Those old 300 MHz boxes can have new life as a server.
So why is Linux also used for enterprise servers, where the XServe would work just fine? I suspect it is because most companies already have a preferred vendor. All the shiny new boxes in Department E came from Dell, and the enterprise servers did too. Unless the business is an Apple shop, already getting computers from Apple, buying an XServe means buying from multiple vendors. And maybe the desktops and the enterprise servers were all bought at once, in a package deal that saved some extra money.
All this is obvious. I moderate Cringely's latest column (-1, Troll).
steveha
Maybe it's overkill for you, but there are a few organizations out there who would like to be able to crank out large volumes of safely random number streams. Verisign probably could use this, for example.
But if you would actually read the article you would find that, despite the name, this project doesn't use a lava lamp. They use a CCD inside a light-proof can, so that all you can measure with the CCD is noise fluctuations. Then they hash the resulting data up big-time to remove any patterns.
I don't know how effective your idea will be, but I'm pretty sure it will take more time to get a chunk of chaotic data from a preamp than from a CCD. The CCD has a few million pixel sensors that can fluctuate and you can read them all in parallel. With the audio preamp you will probably get 16 bits per sample, and how many of those will be chaotic? You can just take more samples but that will take more time.
For true portability, just get a computer with a CPU that makes true random numbers by measuring heat. Again this will no doubt take more time than the CCD, but also no doubt less than the audio preamp. At least all the bits will be chaotic.
steveha
Does anybody actually have any WMA files?
Sure. Lots of people. But why do you even care?
Big ROMs are cheap these days. How does it hurt you if they add WMA support, in addition to MP3, Vorbis, and maybe AAC, Audible, and XYZ and PDQ?
steveha
According to several articles I have read, such as the one on gizmodo.com, the Rio Karma will have USB 2.0 as its native interface; it will also come with a dock that will plug into an Ethernet network.
If you can just use standard file server protocols (NFS or SMB, I don't care) to put files on the Karma, I will buy one. If you have to run some modified jukebox app to move the files, so it can wrap your files in DRM junk, I won't buy one.
steveha
Can't you just put a 2.2 kernel on your servers? SCO hasn't (yet) claimed they own anything in 2.2.
steveha