You don't have to blame Ubuntu to figure out why UserLinux has gone nowhere fast. The concept was simply doomed to failure: "Lets take a bunch of people too busy to work on this project and give them the opportunity to collaborate!" It's an interesting idea, but its near impossible for a single person to simultaneously develop a distribution of software and OS AND offer contracts to support it. Even Perens admits that he couldn't meet both obligations. You need a team of people, such that some can serve the current contracts, some can add the new features, and some can find the people willing to pay for support contracts with those new features. The critical part here is that the revenue is shared among the group. UserLinux was (is?) a fend for yourself volunteer effort. Sure, it's entrepeneurial, and its egalitarian, but its also inefficient to have everyone do everything.
Additionally, you need every chance you can get to get people interested in your idea. The first step would be a pretty website that suggests you're not a collection of amateurs. Shuttleworth also went out of his way to create ShipIt, which not only operated as a marketing tool of itself, but encouraged others to order more than one and share with their peers, AND made a few headlines. You'll notice that UserLinux still uses Debian repos, and their FAQ suggests that UserLinux *IS* Debian, that's somehow been "streamlined" (I'm guessing this means the installer installs everything they want to support by default). Of course, they're still attempting to brand themselves as UserLinux rather than some Debian ISV project. If UserLinux is really Debian, why not just embrace the fact?
I donno what's wrong with your dorm, but putting up a CS server in a dorm is like setting up a 24 hour crack store your favorite dilapidated inner city intersection.
It's obvious he was writing using his background as a specialist in software as a basis for authority on the subject. It's pretty much the only way to write a successful persuasive argument. You don't prepend "this is just my opinion" when you intended to persuade someone to listen to you. You drop how lots of important people go to you for advice, and you always tell them "blah." Clearly, he is trying to project, as he probably should be, the image of someone authoritative.
Does he have to say something like "Look, brokedick, I'm a genius and you're a dumbass. I'm right and you're wrong, and I know more than you could ever possibily know about the subject" to write as an authority on the subject? Fuck no; that's just being an asshole. But as I attempted to put forth in my post above, there are ways of subtly including a basis for authority without being a dick about it. He doesn't have to say "I'm an authority on the subject" for his writing to claim he's an authority on the subject.
Look at the value added things distros do. Apt-get, RPM, documentation, network-configuration tools. Window managers, menu editors, etc. A lot of the work distros do is to make configuring and manipulating free software simple, and the price can be expensive diskwise. If he had said he planned to eliminate bloat from applications instead of this sillyness about 1/3, 2/3s, would you have been more likely to believe he'd succeed?
Well, did he preface his article with, "I'm no expert," or "Maybe I'm wrong, but this is what I think:" or "you don't have any reason to listen to me, but..."; or did he drop the fact that he advises several tech startups on patent and software problems, and claim to understand hackers and proceed to speak on their behalf?
Negroponte has several key people working for him. One of them is Jim Gettys, who's labored for many years to improve X, used to sit on the GNOME foundation board and now sits on the X.org foundation board. I don't think he's the sort to let his boss wander with stars in his eyes, uninformed of what his system can and can't accomplish. Rather, I think that much more is feasible than you think, and that perhaps you're reading too much into his statement. Pretty much every big standard OS's standard desktop is too damn big to fit into their system. Yes, they're using a compression scheme to save space, but there still needs to be some work done to whip it into shape for his needs. I don't think he's trying to give the impression that he looked into it and said "Oh shit! this wont do!" but rather a small dissapointment with how RAM hungry (and storage hungry) today's apps are.
I guess he should blame the same phenomenon that makes his project feasible: rapidly declining prices of technology.
In fact, that BlueMug article quotes one of the main technical / software developers involved with OLPC, Jim Gettys. I believe the OLPC has the right people to accomplish much of the task before them.
500 mhz isn't too bad for oo.o, actually. The startup times can be painful, but that's not the real problem. You've got a 500 mhz laptop and 128M of RAM. Now throw away all your hard drive space, or create a new 512 meg partition on the leftover space on your drive. That swap partition should do nicely, and it's suitable because you wont be using a swap partition at all. dpigs puts oo.o in at around 300 megs between openoffice.org-bin, openoffice.org-core, open-office.org and openoffice.org-common. My distro default linux image clocks in at 50 megs on disk, but there's a lot of driver code for devices I don't own in there, and probably some other useless stuff that an integrated vendor can ditch.
This is why they're looking at Abiword and Gnumeric, which are a bit more lenient on the storage space. And that 512 disk space really is a barrier that has to be overcome. It's not impossible, but it will be an uphill battle. Modern software and distros typically assume the ludicrous amounts of disk space available today. Hell, people put 40 gig drives in THEIR MP3 player! But this is something that has to be sacrificed to make the system work.
To me, the biggest question will be whether Linux and computing in general will recieve any lasting benefit from their experiences and efforts, or if it will all be lost in time like a tear shed in the rain.
Obviously this device has been designed for distribution in countries where people have no expectations of using "computers these days" in their lifetime, by Negroponte's own admission and enthusiasm. But it's not the case that getting a full featured desktop is impossible. 128 is a bit light on RAM, but 192 plus swap wasn't awful on my computer running GNOME or KDE a few years back. I'd expect they could chop together something a bit lighter than either of those without sacrificing much. A couple of small tweaks (perhaps a persistant button to access the applications menu) to XFCE could fill the role nicely.
The specs the article mentions is 500 mhz and 128 megs of RAM. Obviously with 512M of CF, you won't want swap. But its enough to do many of the basic functions an office has need of, if done cleverly. You probably won't have the disk space to be decoding MP3s with it, but it's not outside the specs of the CPU. The trouble is going to be Office Applications. Abiword might fit the bill, but I had a bad experience years ago and haven't revisited it since. But gedit or some other simple text editor should be light weight enough to operate within constraints, and comes with the benefit that it can likely be stored on CF space.
But a much bigger challenge than disk or RAM space is power. Disk and RAM grow bigger and cheaper daily, and the OLPC project has a couple of very bright people capable of handling (or finding someone else who can) those problems. But power requirements aren't as well defined in the market, and that's gonna lead to some problems. Negroponte claims he'll power it with a foot crank or some bizairre no grid required system. This is going to be a real piece of engineering, because unlike the rest of the project, it's not about taking someone else's work and gluing it together. He keeps saying that it's gonna run on something ridiculus like 1W of power, but there's no way you can do that and transmit anything over wifi long distances.
Much of the reason people are starving in Africa has less to do with world production levels and more to do with the political needs of 3rd world dictators. I recall one or two leaders rejecting a significant amount of grain shipments because they were genetically modified, and thus dangerous. Anyone with half a wit would notice that GM food is at least slightly safer than starving to death, and brings the leader's motivation into suspicion. One of the missions of the Bosnian peacekeeping force was to distribute food directly to the people, rather than hand it over to the local warlord for distribution at his whim.
I'd argue that an intelligent and well informed populace is more likely to resist and solve its social ills, and that a computer that networks intelligently could aid these people in communicating with one another about their poverty. This reminds me of a computer ethics final I saw (but have since lost) in which the final question asked how a poor, uneducated mother with a dying child can use a laptop to "solve her problem." The problem being implied that the plight of the poor is caused by an oppressive regime. If anyone has a copy or a link to share on the subject, I'd much appreciate it.
Yes, partnering with Redhat to make a distro is going to make things difficult. But if you give it a brief amount of love, it can easily work. Until two years ago, I ran Debian on a machine with very similar specs, although it did have more disk space. The base install of Debian takes up over 500 megabytes. Localization for a langauge other than english may require up to another 200. The "desktop task" on Debian takes up almost 2 gigs. Although this installs both KDE and GNOME, clearly this system isn't designed for a 100 dollar laptop market that doesn't exist. If rather than install you cook up a single image for all these devices, then KDE+GNOME is down to 1.3 Gigabytes. This is still too damn big. I don't know offhand what eats up a lot of debian's base, but its clear the big guys aren't after this non-existant market at the moment.
However, its not impossible, moreover, its been done before. Fuck, theres even a HOWTO on the subject. There's also several distros and projects on the subject, but many of them have died out as the need for them has waned. In short, you cant just put fedora core on your 512M CF card, and if you expected this, you're much further from done than you think.
Yea, its a shame really. The ratings on the Watergate were so bad, they had to invent a new suffix for any political scandal that gets sufficient attention by viewers. Clearly our fourth branch of the govt is broken. Or, you could stop watching Fox news. If spin is what you're looking to avoid, I know of no better outlet than C-SPAN. It might not get high ratings, but I don't think it's going broke either.
Because if they settle there, then not only do they open themselves up to the Helvetica and Palatino, but they'd have to stop the practice for the one they steal AFTER this one.
If you're going about setting a policy for shareholders to trade their shares privately, you don't hold the meeting behind closed doors with one of your holders presumably dying of leukemia. Not only is it bad form, but its simply too damn late to get any meaningful negotiaions. Yes, small companies have rules in place, primarily so that the stock isn't deemed as a publicly traded company and thus subject to reporting rules. A 36 percent share is big, but it's not nessecarily a controlling interest. You'd have to convince at least 15 percent of the shareholders that your way is the right way. I don't know the full context of which this meeting was occurring, but as I understand it Microsoft didn't have many, if any, outside investors. I don't know whether Allen's condition was a wake up call for the planning procedure or if they were planning to wrest control from whoever might take his place, but doing it ostensibly behind his back like that is going to cause friction.
On the other hand, he didn't seem to be so disquieted with their practices that he felt like selling off his stock entirely(causing a huge drop in value, and likely a downward spiral as others sold). It's hard to fault him though; mentally he probably treats the benefits of all that money as a reward for time building the company and dealing with those assholes, and I don't think that's a bad outlook. Frankley, if he's dissatisfied with management and unable to replace them, I'm not sure why he doesn't simply divest himself of the stock entirely at this point. Perhaps he's worried about SEC punishments for stock manipulation? Yahoo's page doesn't list him as a major holder, so I doubt he holds enough at this point to warrant much reporting.
It helps that Target bothers to spend the money to remove the things in their store that remind them they're standing in a Giant Fucking Warehouse, unlike Wal-Mart's exposed structures and ventilation (and omnipresent eye). FWIW, they pay about the same for either place.
"The way I look at it, just like in warfare, in capitalism there are Patriots, and there are Traitors- and Wal-Mart is effectively the economic version of the Chinese Secret Police."
So my neighbor who went to work as an engineer outside the US for a few years is some sort of gurellia operative or terrorist?
The article implies actually, that Apple intends to use BAPCo to benchmark windows emulation. With Vista slowly retreating, it appears that Apple is moving their army men towards the Microsoftian borders and readying a brutal assault, but windows compatibility is a difficult task to accomplish. Although, if anyone knows how to polish a pig, it's Steve Jobs and company.
I wonder how such a compatibility layer would affect developer's drive to port to OSX. We've seen similar arguments that Wine and ndiswrapper reduce the likelyhood of native ports, although typically the arguments go one step further and claim it's hurting the progression of Free Software.
Ship to ship combat isn't entirely over. In a state of nation vs nation war, ship to ship combat isn't expected, but piracy remains even with America as a rouge superpower. Policing the shipping lanes helps keep the consumer goods the world values safe.
Of course, an Aircraft Carrier isn't suitable for this sort of escort / patrolling mission. The US mainly keeps their carriers in operation globally to maintian a high state of readiness to respond, as you alluded to. Someone starts some shit, the fact that we've got aircraft response 16 hours away will make em think twice. Air superiority is, as I'm sure you know, tantamount to success.
BeOS was 90 percent hot air. To achieve any seriously noticable improvement you'd have to own a SMP system, which were incredibly rare for desktop machines. Yes, it meant you could decode DVDs while doing something else, back when computers of the day were ill prepared to accomplish the task.
As for Linux, and the improvements they can do, most have already been done. There are programs like readahead that do the same thing as OSX and XP's prefetching. That particular readahead was first packaged in late 04. Prelinking ("prebinding")is supported on gcc, but it can be dangerous so most distributions leave it as an option rather than a default. And obviously any of the other gcc supported "speedups" are available to linux.
About the only filesystem feature I'm aware of that Linux has is journalling. ext3, like HFS+, has journalling tacked on to it, with impressive results. ext3 also supports full data journalling, unlike HFS and ext3. I'm not sure hot file clustering is worth the troubles -- you've already got a cache in RAM of files, and it's much faster. The fact that header files find their way into the hot file cache suggests to me that it's not very useful. ext and HFS are both extent based which makes them fairly resistant to fragmentation on large drives. Someone more familiar with ext might chime in with a tool to measure this on ext3.
Kernel extension caches I'm not sure about. As far as I know, Linux modules load quickly; if you wanted them cached I presume you could compile them into the kernel;). I don't think it would be impossible to accomplish but I doubt anyone's tried.
An interesting theory, but most of the actual BSD software has been released from that clause. It's also not very specific; I could for example, bury the fact that my software uses a tweaked version of your software (perhaps altering some identifying strings) very deep in some obscure documentation, and you'd never even know that I used it unless someone takes the time to find and dig through that list of credits.
Surprisingly, I just discovered that it is installed by default on Ubuntu as part of ubuntu-base. And that it's got an HTTP interface that no longer needs the MTA. Not found any website yet but I bet it's there somewhere.
I know. That's why I said popcon does this. popcon = popularity contest. The trick is getting something like popcon that's trustworthy to install by default. I forget where debian places the results, but I've seen the raw "most popular stuff" before.
Ubuntu already approaches this. What Automatix does is allow you to take your habits with you. Say you need opera. It's a somewhat complicated process to do correctly; you add the repo to sources.list, you apt-get install opera, and from now on every update includes opera in the mix from that repo. It lets people jumpstart the process of tweaking their computer. The default system is pretty usable, but it doesn't do things like distribute patented software they don't have a liscence for (ie mp3 players). Automatix is a point and click tool that lets you "fix" percieved problems with the default install.
If you consider the need to seemlessly run Windows apps part of the "1-2-3 I'm done" process, automatix attempts to fix this. Amusingly, you have to enable a repo in a text file etc to get it running;)
Slashdot Mirror
Now I can lose my cell phone and my credit card at the same time!
You don't have to blame Ubuntu to figure out why UserLinux has gone nowhere fast. The concept was simply doomed to failure: "Lets take a bunch of people too busy to work on this project and give them the opportunity to collaborate!" It's an interesting idea, but its near impossible for a single person to simultaneously develop a distribution of software and OS AND offer contracts to support it. Even Perens admits that he couldn't meet both obligations. You need a team of people, such that some can serve the current contracts, some can add the new features, and some can find the people willing to pay for support contracts with those new features. The critical part here is that the revenue is shared among the group. UserLinux was (is?) a fend for yourself volunteer effort. Sure, it's entrepeneurial, and its egalitarian, but its also inefficient to have everyone do everything.
Additionally, you need every chance you can get to get people interested in your idea. The first step would be a pretty website that suggests you're not a collection of amateurs. Shuttleworth also went out of his way to create ShipIt, which not only operated as a marketing tool of itself, but encouraged others to order more than one and share with their peers, AND made a few headlines. You'll notice that UserLinux still uses Debian repos, and their FAQ suggests that UserLinux *IS* Debian, that's somehow been "streamlined" (I'm guessing this means the installer installs everything they want to support by default). Of course, they're still attempting to brand themselves as UserLinux rather than some Debian ISV project. If UserLinux is really Debian, why not just embrace the fact?
I donno what's wrong with your dorm, but putting up a CS server in a dorm is like setting up a 24 hour crack store your favorite dilapidated inner city intersection.
....
It's obvious he was writing using his background as a specialist in software as a basis for authority on the subject. It's pretty much the only way to write a successful persuasive argument. You don't prepend "this is just my opinion" when you intended to persuade someone to listen to you. You drop how lots of important people go to you for advice, and you always tell them "blah." Clearly, he is trying to project, as he probably should be, the image of someone authoritative.
Does he have to say something like "Look, brokedick, I'm a genius and you're a dumbass. I'm right and you're wrong, and I know more than you could ever possibily know about the subject" to write as an authority on the subject? Fuck no; that's just being an asshole. But as I attempted to put forth in my post above, there are ways of subtly including a basis for authority without being a dick about it. He doesn't have to say "I'm an authority on the subject" for his writing to claim he's an authority on the subject.
Look at the value added things distros do. Apt-get, RPM, documentation, network-configuration tools. Window managers, menu editors, etc. A lot of the work distros do is to make configuring and manipulating free software simple, and the price can be expensive diskwise. If he had said he planned to eliminate bloat from applications instead of this sillyness about 1/3, 2/3s, would you have been more likely to believe he'd succeed?
Well, did he preface his article with, "I'm no expert," or "Maybe I'm wrong, but this is what I think:" or "you don't have any reason to listen to me, but..."; or did he drop the fact that he advises several tech startups on patent and software problems, and claim to understand hackers and proceed to speak on their behalf?
Negroponte has several key people working for him. One of them is Jim Gettys, who's labored for many years to improve X, used to sit on the GNOME foundation board and now sits on the X.org foundation board. I don't think he's the sort to let his boss wander with stars in his eyes, uninformed of what his system can and can't accomplish. Rather, I think that much more is feasible than you think, and that perhaps you're reading too much into his statement. Pretty much every big standard OS's standard desktop is too damn big to fit into their system. Yes, they're using a compression scheme to save space, but there still needs to be some work done to whip it into shape for his needs. I don't think he's trying to give the impression that he looked into it and said "Oh shit! this wont do!" but rather a small dissapointment with how RAM hungry (and storage hungry) today's apps are.
I guess he should blame the same phenomenon that makes his project feasible: rapidly declining prices of technology.
In fact, that BlueMug article quotes one of the main technical / software developers involved with OLPC, Jim Gettys. I believe the OLPC has the right people to accomplish much of the task before them.
500 mhz isn't too bad for oo.o, actually. The startup times can be painful, but that's not the real problem. You've got a 500 mhz laptop and 128M of RAM. Now throw away all your hard drive space, or create a new 512 meg partition on the leftover space on your drive. That swap partition should do nicely, and it's suitable because you wont be using a swap partition at all. dpigs puts oo.o in at around 300 megs between openoffice.org-bin, openoffice.org-core, open-office.org and openoffice.org-common. My distro default linux image clocks in at 50 megs on disk, but there's a lot of driver code for devices I don't own in there, and probably some other useless stuff that an integrated vendor can ditch.
This is why they're looking at Abiword and Gnumeric, which are a bit more lenient on the storage space. And that 512 disk space really is a barrier that has to be overcome. It's not impossible, but it will be an uphill battle. Modern software and distros typically assume the ludicrous amounts of disk space available today. Hell, people put 40 gig drives in THEIR MP3 player! But this is something that has to be sacrificed to make the system work.
To me, the biggest question will be whether Linux and computing in general will recieve any lasting benefit from their experiences and efforts, or if it will all be lost in time like a tear shed in the rain.
Obviously this device has been designed for distribution in countries where people have no expectations of using "computers these days" in their lifetime, by Negroponte's own admission and enthusiasm. But it's not the case that getting a full featured desktop is impossible. 128 is a bit light on RAM, but 192 plus swap wasn't awful on my computer running GNOME or KDE a few years back. I'd expect they could chop together something a bit lighter than either of those without sacrificing much. A couple of small tweaks (perhaps a persistant button to access the applications menu) to XFCE could fill the role nicely.
The specs the article mentions is 500 mhz and 128 megs of RAM. Obviously with 512M of CF, you won't want swap. But its enough to do many of the basic functions an office has need of, if done cleverly. You probably won't have the disk space to be decoding MP3s with it, but it's not outside the specs of the CPU. The trouble is going to be Office Applications. Abiword might fit the bill, but I had a bad experience years ago and haven't revisited it since. But gedit or some other simple text editor should be light weight enough to operate within constraints, and comes with the benefit that it can likely be stored on CF space.
But a much bigger challenge than disk or RAM space is power. Disk and RAM grow bigger and cheaper daily, and the OLPC project has a couple of very bright people capable of handling (or finding someone else who can) those problems. But power requirements aren't as well defined in the market, and that's gonna lead to some problems. Negroponte claims he'll power it with a foot crank or some bizairre no grid required system. This is going to be a real piece of engineering, because unlike the rest of the project, it's not about taking someone else's work and gluing it together. He keeps saying that it's gonna run on something ridiculus like 1W of power, but there's no way you can do that and transmit anything over wifi long distances.
Much of the reason people are starving in Africa has less to do with world production levels and more to do with the political needs of 3rd world dictators. I recall one or two leaders rejecting a significant amount of grain shipments because they were genetically modified, and thus dangerous. Anyone with half a wit would notice that GM food is at least slightly safer than starving to death, and brings the leader's motivation into suspicion. One of the missions of the Bosnian peacekeeping force was to distribute food directly to the people, rather than hand it over to the local warlord for distribution at his whim.
I'd argue that an intelligent and well informed populace is more likely to resist and solve its social ills, and that a computer that networks intelligently could aid these people in communicating with one another about their poverty. This reminds me of a computer ethics final I saw (but have since lost) in which the final question asked how a poor, uneducated mother with a dying child can use a laptop to "solve her problem." The problem being implied that the plight of the poor is caused by an oppressive regime. If anyone has a copy or a link to share on the subject, I'd much appreciate it.
Yes, partnering with Redhat to make a distro is going to make things difficult. But if you give it a brief amount of love, it can easily work. Until two years ago, I ran Debian on a machine with very similar specs, although it did have more disk space. The base install of Debian takes up over 500 megabytes. Localization for a langauge other than english may require up to another 200. The "desktop task" on Debian takes up almost 2 gigs. Although this installs both KDE and GNOME, clearly this system isn't designed for a 100 dollar laptop market that doesn't exist. If rather than install you cook up a single image for all these devices, then KDE+GNOME is down to 1.3 Gigabytes. This is still too damn big. I don't know offhand what eats up a lot of debian's base, but its clear the big guys aren't after this non-existant market at the moment.
However, its not impossible, moreover, its been done before. Fuck, theres even a HOWTO on the subject. There's also several distros and projects on the subject, but many of them have died out as the need for them has waned. In short, you cant just put fedora core on your 512M CF card, and if you expected this, you're much further from done than you think.
Yea, its a shame really. The ratings on the Watergate were so bad, they had to invent a new suffix for any political scandal that gets sufficient attention by viewers. Clearly our fourth branch of the govt is broken. Or, you could stop watching Fox news. If spin is what you're looking to avoid, I know of no better outlet than C-SPAN. It might not get high ratings, but I don't think it's going broke either.
Because if they settle there, then not only do they open themselves up to the Helvetica and Palatino, but they'd have to stop the practice for the one they steal AFTER this one.
If you're going about setting a policy for shareholders to trade their shares privately, you don't hold the meeting behind closed doors with one of your holders presumably dying of leukemia. Not only is it bad form, but its simply too damn late to get any meaningful negotiaions. Yes, small companies have rules in place, primarily so that the stock isn't deemed as a publicly traded company and thus subject to reporting rules. A 36 percent share is big, but it's not nessecarily a controlling interest. You'd have to convince at least 15 percent of the shareholders that your way is the right way. I don't know the full context of which this meeting was occurring, but as I understand it Microsoft didn't have many, if any, outside investors. I don't know whether Allen's condition was a wake up call for the planning procedure or if they were planning to wrest control from whoever might take his place, but doing it ostensibly behind his back like that is going to cause friction.
On the other hand, he didn't seem to be so disquieted with their practices that he felt like selling off his stock entirely(causing a huge drop in value, and likely a downward spiral as others sold). It's hard to fault him though; mentally he probably treats the benefits of all that money as a reward for time building the company and dealing with those assholes, and I don't think that's a bad outlook. Frankley, if he's dissatisfied with management and unable to replace them, I'm not sure why he doesn't simply divest himself of the stock entirely at this point. Perhaps he's worried about SEC punishments for stock manipulation? Yahoo's page doesn't list him as a major holder, so I doubt he holds enough at this point to warrant much reporting.
It helps that Target bothers to spend the money to remove the things in their store that remind them they're standing in a Giant Fucking Warehouse, unlike Wal-Mart's exposed structures and ventilation (and omnipresent eye). FWIW, they pay about the same for either place.
"The way I look at it, just like in warfare, in capitalism there are Patriots, and there are Traitors- and Wal-Mart is effectively the economic version of the Chinese Secret Police."
So my neighbor who went to work as an engineer outside the US for a few years is some sort of gurellia operative or terrorist?
The article implies actually, that Apple intends to use BAPCo to benchmark windows emulation. With Vista slowly retreating, it appears that Apple is moving their army men towards the Microsoftian borders and readying a brutal assault, but windows compatibility is a difficult task to accomplish. Although, if anyone knows how to polish a pig, it's Steve Jobs and company.
I wonder how such a compatibility layer would affect developer's drive to port to OSX. We've seen similar arguments that Wine and ndiswrapper reduce the likelyhood of native ports, although typically the arguments go one step further and claim it's hurting the progression of Free Software.
BSDs made plenty of headway without much support from GNU/GPL/FSF. Make of it what you will.
Ship to ship combat isn't entirely over. In a state of nation vs nation war, ship to ship combat isn't expected, but piracy remains even with America as a rouge superpower. Policing the shipping lanes helps keep the consumer goods the world values safe.
Of course, an Aircraft Carrier isn't suitable for this sort of escort / patrolling mission. The US mainly keeps their carriers in operation globally to maintian a high state of readiness to respond, as you alluded to. Someone starts some shit, the fact that we've got aircraft response 16 hours away will make em think twice. Air superiority is, as I'm sure you know, tantamount to success.
BeOS was 90 percent hot air. To achieve any seriously noticable improvement you'd have to own a SMP system, which were incredibly rare for desktop machines. Yes, it meant you could decode DVDs while doing something else, back when computers of the day were ill prepared to accomplish the task.
;). I don't think it would be impossible to accomplish but I doubt anyone's tried.
As for Linux, and the improvements they can do, most have already been done. There are programs like readahead that do the same thing as OSX and XP's prefetching. That particular readahead was first packaged in late 04. Prelinking ("prebinding")is supported on gcc, but it can be dangerous so most distributions leave it as an option rather than a default. And obviously any of the other gcc supported "speedups" are available to linux.
About the only filesystem feature I'm aware of that Linux has is journalling. ext3, like HFS+, has journalling tacked on to it, with impressive results. ext3 also supports full data journalling, unlike HFS and ext3. I'm not sure hot file clustering is worth the troubles -- you've already got a cache in RAM of files, and it's much faster. The fact that header files find their way into the hot file cache suggests to me that it's not very useful. ext and HFS are both extent based which makes them fairly resistant to fragmentation on large drives. Someone more familiar with ext might chime in with a tool to measure this on ext3.
Kernel extension caches I'm not sure about. As far as I know, Linux modules load quickly; if you wanted them cached I presume you could compile them into the kernel
An interesting theory, but most of the actual BSD software has been released from that clause. It's also not very specific; I could for example, bury the fact that my software uses a tweaked version of your software (perhaps altering some identifying strings) very deep in some obscure documentation, and you'd never even know that I used it unless someone takes the time to find and dig through that list of credits.
Surprisingly, I just discovered that it is installed by default on Ubuntu as part of ubuntu-base. And that it's got an HTTP interface that no longer needs the MTA. Not found any website yet but I bet it's there somewhere.
I know. That's why I said popcon does this. popcon = popularity contest. The trick is getting something like popcon that's trustworthy to install by default. I forget where debian places the results, but I've seen the raw "most popular stuff" before.
Ubuntu already approaches this. What Automatix does is allow you to take your habits with you. Say you need opera. It's a somewhat complicated process to do correctly; you add the repo to sources.list, you apt-get install opera, and from now on every update includes opera in the mix from that repo. It lets people jumpstart the process of tweaking their computer. The default system is pretty usable, but it doesn't do things like distribute patented software they don't have a liscence for (ie mp3 players). Automatix is a point and click tool that lets you "fix" percieved problems with the default install.
;)
If you consider the need to seemlessly run Windows apps part of the "1-2-3 I'm done" process, automatix attempts to fix this. Amusingly, you have to enable a repo in a text file etc to get it running