I think here people would be surprised to find out that someone answering their question *is* a legal professional and not a 15-year-old. The only time I've seen a post that claimed to be from a lawyer, it started, "I am a lawyer, but this shouldn't be taken as professional legal advice."
Re:Preview of what's to come...
on
Internet2 Update
·
· Score: 2
Two reasons that won't happen:
The technology is going to migrate to the normal internet, so the normal internet will be able to do all of this. The main difference right now is that the people on Internet2 have really good connections to it, while the people on the Internet generally don't. But they'll get good connections to the regular Internet, not good connections to Internet2.
The Internet grew so much for largely cultural reasons: people who had been on it in research and academic contexts left those contexts, but wanted to maintain their ties. So they figured out ways to get internet connections from other sources. Soon, the people they knew from other contexts wanted to interact the same way, and everyone was getting online. This won't happen with internet2, because everyone on it is likely to be on the normal internet. They'll just use that once they leave, possibly bringing their applications over if they can get the bandwidth.
After all, Internet2 is essentially the same thing as the regular internet, except with only two backbones and few non-experimental programs.
The problem is not that the code didn't get peer review (how many people would actually search through this looking for bugs and report ones they found? Would anyone else actually want to use this?), but that it didn't get redesigned for clarity. Something like this should be reworked to the point where everything is obviously correct.
Adobe probably only wanted from the start to get Killustrator to change its name. In order to accomplish this in a way that counts as defending their trademark, they have to make it official (they can't just call the guy up and ask nicely). But the guy was in Germany, and, under German law, the lawyers take their pay out of the hide of the people they harass. Since this in an international project with significant visibility in the US, where that sort of thing is considered blatently unjust, they seem to be trying to get the lawyers in check.
Problem solved; chalk the whole mess up to cultural differences. It would be interesting to know, however, what Adobe told the lawyers to do in the first place; neither of them seems to have revealed whether Adobe actually asked the lawyers to pursue this particular case.
Linus used to code like that, but he's got better things to do. What makes him a really great programmer is that even more work on Linux gets done in a weekend now than when he did; a great programmer increases the productivity of everyone on the team by providing good examples of code to emulate, and by providing well-written, reusable code.
If one member of a team is 10 times as productive as another, then either the unproductive one isn't any good or the productive one isn't writing good clear code. This means that the code will probably have to be thrown out if the team changes. It also means that the prima donna is being unproductive overall, since they're wasting their coworkers' time. If the best programmer on a team can improve the code base such that the other members' code is easier to write, they can save themselves the trouble of writing the easy stuff. If the guru is only 4 times as productive as a normal member, but the other 8 people are twice as productive as they would otherwise be, the team gets more done overall.
Being 10 times as productive as an average person is a good reason to work alone. Which is, in fact, what Linus did initially when he was churning out tons of code.
It looks like something someone wrote without looking through an algorithms book. The encryption is just a random algorithm of the same general type as DES, but almost certainly weak (your average similar algorithm is not secure). The checksum thing is also error-prone; it doesn't notice if you transpose a 1 with an 8 or replace a 2 with a 9, for instance.
Allow small booksellers to compete with big ones
on
Books on Demand
·
· Score: 2
The way I see this being important is that it lets you not need a huge warehouse and a big store to offer a large selection. This means that a little storefront bookstore can offer as many books as a huge chain store. Of course, this doesn't let you browse the books that aren't on the shelves, but you should at least be able to look through a list of titles.
AOL has to do this sort of thing if they want to keep their trademark-- a competitor need only show that AOL had failed to pursue someone who used a similar name to AOL's.
On the other hand, AOL can only win if the names are determined to be different; the gaim people can say that AOL didn't pursue them for two years when AOL knew about gaim's use of the name. If the names are too similar, then AOL loses the trademark. It doesn't matter that AOL trademarked the name 2 years after gaim started using it, but it matters a whole lot that they didn't complain about gaim using the name for 3 years. That's failing to enforce your trademark, and the fact that they started enforcing it reasonably soon after they got it is not sufficient.
Presumably what AOL wants is to have gaim argue that the name isn't similar, and win the case in that way. Then AOL goes about its business, able to prevent other people from actually using the same name. This shouldn't be a hard case to win, especially because AOL simply wants to lose in the better way. If AOL won this, in fact, they'd lose to the first company with money, who could still argue that AOL didn't bug gaim for 3 years and thus really didn't defend their trademark.
I'm frankly somewhat worried about the titles of the parts: I could sort of see the Messiah one (and it might be interesting to be trying to save someone other than yourself), but the Children of Dune one is just too much. I don't want to know how that guy ends up with children...
I'm still seeing new boxes with '98. If OEMs aren't switching over to 2000 yet, it'll be a long time before they go to XP. The people who are getting free upgrades with their boxes will probably not take them, since upgrading a windows box is too difficult for your average user and XP seems to offer fewer features and an unfamiliar interface.
Kodak will have plenty of time to produce a version that works with XP if XP gets adopted. For now, it's probably better to tell customers not to get XP and ask for 2000 or before when they buy a computer. No point in making it easier for MicroSoft to get XP adopted if they're not going to be nice to developers.
So you think that Fuji will have an easier time getting XP to not use the MicroSoft version than Kodak has? It's not like MicroSoft has anything specifically against Kodak that they don't have against other makers.
Of course, the camera should work with XP-- it would probably be practically impossible to prevent that. But the software need not, and the software presumably comes with the camera.
You have higher total fixed costs, but you can cope better with relatively few people paying. Centralized power costs less per person, but a lot more total; if you don't have the installed base to amortize the costs over, US-style plants are just impossible. If you have to get power lines put in all the way from a central location instead of just a fork off of the neighbors', the cost of infrastructure will overwhelm the cost of electricity, no matter how it's produced. And if you're the only person on the power lines, they're going to get much less maintenance and service than something you can take care of yourself.
The thing that really makes solar better for them than anything else is that you can have your own solar panels, and the power doesn't have to be sent anywhere. That means that you don't rely on power lines which may go unmaintained or a power plant which may go offline. The first-world power grids work well because there's a lot of redundancy and utilities which can maintain the power lines.
If you don't have thousands of other paying customers on the same segment as you, you're not going to get great service, and you'd do better to just do it yourself with a solar system or a generator, and solar's just easier at that scale.
Pulling this kind of stunt makes Adobe look really bad. It may be acceptable by German law, but in most of Adobe's markets, they look anti-education and anti-individual. If US art schools see this and decide they don't want anything to do with Adobe, Adobe stands to lose much more than $1 million from the schools and students who would learn to use their products. So the logical thing for Adobe to do is pressure the lawyers into dropping the case, and maybe pay them for their efforts in getting Killustrator to change its name.
Kodak still controls their own packaging. They could probably simply say on it that the software and camera don't work with Windows XP. Considering how much trouble MS is going to have getting people to switch to XP, it might be best to simply discourage the use of XP if MS is going to do this sort of thing.
If people have the impression that, if they switch to XP, they'll have to go through a complicated process to get their computer to work with their camera, plus pay extra to get prints, it's only to MS's disadvantage.
Re:Make a decision, folks
on
ORBS Forks
·
· Score: 2
These lists don't actually list spammers. They list machines which are configured incorrectly such that people can misuse them. People aren't on the list because of what they are trying to say.
This assumes, of course, that the lists actually do what they're intended to do: prevent unresolved problems from hurting other people. Ideally, you would only end up on such a list if you were running an open relay, you would be notified, and you would be able to get off the list as soon as you fixed your mail server. The censorship issues have been with people getting on such lists when they were not, in fact, running an open relay, or not being removed from the list when they fixed it. That sort of thing, in addition to being generally bad for freedom, also reduces the effectiveness of the lists.
If you just write it on a slip of paper and stick it in your pocket, you're probably really safe. It's rather unlikely that someone will mug you for it, and you can make it really hard to guess while not leaving it around the computer. Given the number of people who pick good passwords and then leave them where it's easy for an attacker to get at and obvious what it's for, you'd think people would think to carry them.
I mean, they tell you not to carry your ATM PIN with you, but that's because you'd have the card and the PIN in the same place. You're probably not carrying your work computer with you...
Also, you'll probably have memorized your password after using it a bunch of times by looking at the slip of paper, at which point, you can destroy it.
I think what would be nicer is if everything used a common way of presenting information. Each driver that gets an IRQ or something has a different wording for telling you which one it has, and some drivers' messages sound like they might indicate errors even when the situation is pretty normal and supported.
I'd much rather see a list of:
driverx.c Driver for X on IRQ n DMA m
drivery.c Driver for Y device not detected
so you could see, at a glance, what stuff is in your kernel, what is actually active, and how it autoconfigured itself. Currently, you have to look through the messages for the one that applies to a misbehaving device, and then sometimes look at the driver source to figure out what the message means.
Most graphic design these days is either to be displayed on computers or is at least done on computers. In fact, most images made on computers are probably graphic design.
This doesn't mean that computer art can't be fine art; most images hung on walls are also graphic design products.
Of course, it's possible to define fine art to specify certain traditional media, such that, for instance, only Monet originals and not prints are fine art; similarly, Bach isn't fine art, but music. Oil painting and watercolor and acrylic are different skills from computer art, as they are from mosaic or sculpture. In this sense, computer art is, well, computer art, not anything else.
That doesn't mean it is graphic design, which is an entirely different thing whose principles are often in opposition. Nor is it somehow less pure or real than painting.
The way trademark law works, they have to bully everyone who uses their name to refer to something else, even if the name is only slightly related. As far as I can tell, if they end up in a real trademark dispute, they lose if the other people can produce anyone they didn't bully.
On the other hand, if you were using the name 2 years ago without permission and they didn't bully you, you should be able to win just by saying that they didn't bully you two years ago; they failed for two years to defend their trademark (nevermind that they didn't have it at the time), so they lose.
On the first hand again, they've done their job if you get a ruling that says the names are different, because they will not have failed to defend their trademark. I suspect that you would be able to reach a good settlement if told them you were willing to acknowledge their trademark of "AIM" (which seems to be currently unacknowledged on the sourceforge site), and thought that "gaim" was sufficiently different to not constitute an example of someone using their trademark for two years without them complaining.
If you're willing to argue that what you've done does not infringe their trademark (as opposed to arguing that they've failed to defend it), it's in their best interests to let you win, since they don't have a case, and they'd lose their trademark the other way. Of course, they can't very well send you a letter insisting that "AIM" and "gaim" are not confusingly similar, since that wouldn't get into the official record, even if that's the outcome they really want.
While it doesn't actually say it, I assume that, by posting to slashdot, you agree to allow slashdot to distribute your comment on the web, and probably assert that you actually hold the copyright on what you're posting.
I'm not entirely sure if slashdot is supposed to actually put this agreement in legalese somewhere, or if the actual behavior of the site is sufficient. For example, newspapers generally print letters to the editor, even if the letter doesn't explicitly give them the right to print it, because that's how letter to the editor just work normally.
The issue is actually about cases where someone has agreed to have their work used in a particular way (printed in the newspaper) for pay. The agreement didn't involve other media, so the newspaper doesn't have the right to put it on the web without negotiating a new contract. Everything is explicit in that case, and they were writing it for print.
Probably what this means is that, in the future, news sites will probably just put web distribution in their freelance contracts.
Reversible computers don't drain the charge to ground. One major goal is minimal power use, which means that, whenever possible, you don't discharge anything. In particular, you want to move the charge from one capacitor to another instead of to ground; which of the capacitors is determines whether the bit is a one or a zero.
Of course, there are only certain operations you can do without discharging a capacitor to ground or charging it to Vcc. And there's obviously some energy expenditure in the resistence of the wires, capacitor leakage and the effort of getting the computation to go forward.
The idea is that users can use any computer anywhere, and they access a central location, where their information is stored.
What this will really mean, though, is: people don't buy PCs to put their data on. They buy servers to put their data on. They've finally gotten their 24x7 connection, and they're going to stick a box in the basement next to the other utilities. Then they can go anywhere and get their desktop from their computer at home.
For the first iterations, the box will also be the PC they use, but it will act exactly the same as if it were storing things remotely. Of course, it will act as a thin client for other people who want to use other home machines, and it will be a server for when they're at their friends' houses.
Later, of course, they'll want more places to connect from than they want central stores, and they'll want to upgrade their client capabilities (graphics, raw processing, etc) at a different time from when they upgrade their server (long-term hard drive space), and the server will fall out of use as a client.
All this will be possible in... 1996. At least, that's when I started doing it, at least as far as ssh, and then screen, which covers most of the things I actually want to do with a centralized configuration. MicroSoft will probably end up with a similar model (using proprietary parts, of course, so MS client with MS server does things that either doesn't otherwise do). Being the server is probably a stopgap for people who aren't yet set up to run the servers themselves; after all, serving files for people isn't all that exciting, especially if they get people to use software subscriptions and pay to run the file servers on their own machines.
TUX does a good job of identifying inefficiencies in the network code, and of identifying at least one critical path that can be optimized. It doesn't actually do all that much which is particular to web serving; it's mainly a routine for efficiently getting a file from disk (or, most likely, cache) onto the network, with a bit of stuff that does HTTP. Doing the whole thing is the kernel is slightly useful for performance, but I think the main effect is that the whole thing gets written and optimized at once.
Normal web servers are two problems working together: the kernel and the web server. Having a single program doing it means that it's easier to track what's going on through the whole execution path, so end-to-end debugging is easier.
In any case, HTTP is significantly simpler than TCP and IP, which the kernel handle anyway. So most of your web server is really already in the kernel; it's just that, instead of responding to just pings and TCP control packets for you, it also responds to certain requests for you.
Slashdot Mirror
I think here people would be surprised to find out that someone answering their question *is* a legal professional and not a 15-year-old. The only time I've seen a post that claimed to be from a lawyer, it started, "I am a lawyer, but this shouldn't be taken as professional legal advice."
Two reasons that won't happen:
The technology is going to migrate to the normal internet, so the normal internet will be able to do all of this. The main difference right now is that the people on Internet2 have really good connections to it, while the people on the Internet generally don't. But they'll get good connections to the regular Internet, not good connections to Internet2.
The Internet grew so much for largely cultural reasons: people who had been on it in research and academic contexts left those contexts, but wanted to maintain their ties. So they figured out ways to get internet connections from other sources. Soon, the people they knew from other contexts wanted to interact the same way, and everyone was getting online. This won't happen with internet2, because everyone on it is likely to be on the normal internet. They'll just use that once they leave, possibly bringing their applications over if they can get the bandwidth.
After all, Internet2 is essentially the same thing as the regular internet, except with only two backbones and few non-experimental programs.
The problem is not that the code didn't get peer review (how many people would actually search through this looking for bugs and report ones they found? Would anyone else actually want to use this?), but that it didn't get redesigned for clarity. Something like this should be reworked to the point where everything is obviously correct.
Adobe probably only wanted from the start to get Killustrator to change its name. In order to accomplish this in a way that counts as defending their trademark, they have to make it official (they can't just call the guy up and ask nicely). But the guy was in Germany, and, under German law, the lawyers take their pay out of the hide of the people they harass. Since this in an international project with significant visibility in the US, where that sort of thing is considered blatently unjust, they seem to be trying to get the lawyers in check.
Problem solved; chalk the whole mess up to cultural differences. It would be interesting to know, however, what Adobe told the lawyers to do in the first place; neither of them seems to have revealed whether Adobe actually asked the lawyers to pursue this particular case.
Linus used to code like that, but he's got better things to do. What makes him a really great programmer is that even more work on Linux gets done in a weekend now than when he did; a great programmer increases the productivity of everyone on the team by providing good examples of code to emulate, and by providing well-written, reusable code.
If one member of a team is 10 times as productive as another, then either the unproductive one isn't any good or the productive one isn't writing good clear code. This means that the code will probably have to be thrown out if the team changes. It also means that the prima donna is being unproductive overall, since they're wasting their coworkers' time. If the best programmer on a team can improve the code base such that the other members' code is easier to write, they can save themselves the trouble of writing the easy stuff. If the guru is only 4 times as productive as a normal member, but the other 8 people are twice as productive as they would otherwise be, the team gets more done overall.
Being 10 times as productive as an average person is a good reason to work alone. Which is, in fact, what Linus did initially when he was churning out tons of code.
..."Information Superhighway"
Hey, it's a highway with a ton of information on it. Better go over to one with more free bandwidth...
It looks like something someone wrote without looking through an algorithms book. The encryption is just a random algorithm of the same general type as DES, but almost certainly weak (your average similar algorithm is not secure). The checksum thing is also error-prone; it doesn't notice if you transpose a 1 with an 8 or replace a 2 with a 9, for instance.
The way I see this being important is that it lets you not need a huge warehouse and a big store to offer a large selection. This means that a little storefront bookstore can offer as many books as a huge chain store. Of course, this doesn't let you browse the books that aren't on the shelves, but you should at least be able to look through a list of titles.
AOL has to do this sort of thing if they want to keep their trademark-- a competitor need only show that AOL had failed to pursue someone who used a similar name to AOL's.
On the other hand, AOL can only win if the names are determined to be different; the gaim people can say that AOL didn't pursue them for two years when AOL knew about gaim's use of the name. If the names are too similar, then AOL loses the trademark. It doesn't matter that AOL trademarked the name 2 years after gaim started using it, but it matters a whole lot that they didn't complain about gaim using the name for 3 years. That's failing to enforce your trademark, and the fact that they started enforcing it reasonably soon after they got it is not sufficient.
Presumably what AOL wants is to have gaim argue that the name isn't similar, and win the case in that way. Then AOL goes about its business, able to prevent other people from actually using the same name. This shouldn't be a hard case to win, especially because AOL simply wants to lose in the better way. If AOL won this, in fact, they'd lose to the first company with money, who could still argue that AOL didn't bug gaim for 3 years and thus really didn't defend their trademark.
I'm frankly somewhat worried about the titles of the parts: I could sort of see the Messiah one (and it might be interesting to be trying to save someone other than yourself), but the Children of Dune one is just too much. I don't want to know how that guy ends up with children...
I'm still seeing new boxes with '98. If OEMs aren't switching over to 2000 yet, it'll be a long time before they go to XP. The people who are getting free upgrades with their boxes will probably not take them, since upgrading a windows box is too difficult for your average user and XP seems to offer fewer features and an unfamiliar interface.
Kodak will have plenty of time to produce a version that works with XP if XP gets adopted. For now, it's probably better to tell customers not to get XP and ask for 2000 or before when they buy a computer. No point in making it easier for MicroSoft to get XP adopted if they're not going to be nice to developers.
So you think that Fuji will have an easier time getting XP to not use the MicroSoft version than Kodak has? It's not like MicroSoft has anything specifically against Kodak that they don't have against other makers.
Of course, the camera should work with XP-- it would probably be practically impossible to prevent that. But the software need not, and the software presumably comes with the camera.
You have higher total fixed costs, but you can cope better with relatively few people paying. Centralized power costs less per person, but a lot more total; if you don't have the installed base to amortize the costs over, US-style plants are just impossible. If you have to get power lines put in all the way from a central location instead of just a fork off of the neighbors', the cost of infrastructure will overwhelm the cost of electricity, no matter how it's produced. And if you're the only person on the power lines, they're going to get much less maintenance and service than something you can take care of yourself.
The thing that really makes solar better for them than anything else is that you can have your own solar panels, and the power doesn't have to be sent anywhere. That means that you don't rely on power lines which may go unmaintained or a power plant which may go offline. The first-world power grids work well because there's a lot of redundancy and utilities which can maintain the power lines.
If you don't have thousands of other paying customers on the same segment as you, you're not going to get great service, and you'd do better to just do it yourself with a solar system or a generator, and solar's just easier at that scale.
Pulling this kind of stunt makes Adobe look really bad. It may be acceptable by German law, but in most of Adobe's markets, they look anti-education and anti-individual. If US art schools see this and decide they don't want anything to do with Adobe, Adobe stands to lose much more than $1 million from the schools and students who would learn to use their products. So the logical thing for Adobe to do is pressure the lawyers into dropping the case, and maybe pay them for their efforts in getting Killustrator to change its name.
Kodak still controls their own packaging. They could probably simply say on it that the software and camera don't work with Windows XP. Considering how much trouble MS is going to have getting people to switch to XP, it might be best to simply discourage the use of XP if MS is going to do this sort of thing.
If people have the impression that, if they switch to XP, they'll have to go through a complicated process to get their computer to work with their camera, plus pay extra to get prints, it's only to MS's disadvantage.
These lists don't actually list spammers. They list machines which are configured incorrectly such that people can misuse them. People aren't on the list because of what they are trying to say.
This assumes, of course, that the lists actually do what they're intended to do: prevent unresolved problems from hurting other people. Ideally, you would only end up on such a list if you were running an open relay, you would be notified, and you would be able to get off the list as soon as you fixed your mail server. The censorship issues have been with people getting on such lists when they were not, in fact, running an open relay, or not being removed from the list when they fixed it. That sort of thing, in addition to being generally bad for freedom, also reduces the effectiveness of the lists.
If you just write it on a slip of paper and stick it in your pocket, you're probably really safe. It's rather unlikely that someone will mug you for it, and you can make it really hard to guess while not leaving it around the computer. Given the number of people who pick good passwords and then leave them where it's easy for an attacker to get at and obvious what it's for, you'd think people would think to carry them.
I mean, they tell you not to carry your ATM PIN with you, but that's because you'd have the card and the PIN in the same place. You're probably not carrying your work computer with you...
Also, you'll probably have memorized your password after using it a bunch of times by looking at the slip of paper, at which point, you can destroy it.
I think what would be nicer is if everything used a common way of presenting information. Each driver that gets an IRQ or something has a different wording for telling you which one it has, and some drivers' messages sound like they might indicate errors even when the situation is pretty normal and supported.
I'd much rather see a list of:
driverx.c Driver for X on IRQ n DMA m
drivery.c Driver for Y device not detected
so you could see, at a glance, what stuff is in your kernel, what is actually active, and how it autoconfigured itself. Currently, you have to look through the messages for the one that applies to a misbehaving device, and then sometimes look at the driver source to figure out what the message means.
Most graphic design these days is either to be displayed on computers or is at least done on computers. In fact, most images made on computers are probably graphic design.
This doesn't mean that computer art can't be fine art; most images hung on walls are also graphic design products.
Of course, it's possible to define fine art to specify certain traditional media, such that, for instance, only Monet originals and not prints are fine art; similarly, Bach isn't fine art, but music. Oil painting and watercolor and acrylic are different skills from computer art, as they are from mosaic or sculpture. In this sense, computer art is, well, computer art, not anything else.
That doesn't mean it is graphic design, which is an entirely different thing whose principles are often in opposition. Nor is it somehow less pure or real than painting.
The way trademark law works, they have to bully everyone who uses their name to refer to something else, even if the name is only slightly related. As far as I can tell, if they end up in a real trademark dispute, they lose if the other people can produce anyone they didn't bully.
On the other hand, if you were using the name 2 years ago without permission and they didn't bully you, you should be able to win just by saying that they didn't bully you two years ago; they failed for two years to defend their trademark (nevermind that they didn't have it at the time), so they lose.
On the first hand again, they've done their job if you get a ruling that says the names are different, because they will not have failed to defend their trademark. I suspect that you would be able to reach a good settlement if told them you were willing to acknowledge their trademark of "AIM" (which seems to be currently unacknowledged on the sourceforge site), and thought that "gaim" was sufficiently different to not constitute an example of someone using their trademark for two years without them complaining.
If you're willing to argue that what you've done does not infringe their trademark (as opposed to arguing that they've failed to defend it), it's in their best interests to let you win, since they don't have a case, and they'd lose their trademark the other way. Of course, they can't very well send you a letter insisting that "AIM" and "gaim" are not confusingly similar, since that wouldn't get into the official record, even if that's the outcome they really want.
While it doesn't actually say it, I assume that, by posting to slashdot, you agree to allow slashdot to distribute your comment on the web, and probably assert that you actually hold the copyright on what you're posting.
I'm not entirely sure if slashdot is supposed to actually put this agreement in legalese somewhere, or if the actual behavior of the site is sufficient. For example, newspapers generally print letters to the editor, even if the letter doesn't explicitly give them the right to print it, because that's how letter to the editor just work normally.
The issue is actually about cases where someone has agreed to have their work used in a particular way (printed in the newspaper) for pay. The agreement didn't involve other media, so the newspaper doesn't have the right to put it on the web without negotiating a new contract. Everything is explicit in that case, and they were writing it for print.
Probably what this means is that, in the future, news sites will probably just put web distribution in their freelance contracts.
Reversible computers don't drain the charge to ground. One major goal is minimal power use, which means that, whenever possible, you don't discharge anything. In particular, you want to move the charge from one capacitor to another instead of to ground; which of the capacitors is determines whether the bit is a one or a zero.
Of course, there are only certain operations you can do without discharging a capacitor to ground or charging it to Vcc. And there's obviously some energy expenditure in the resistence of the wires, capacitor leakage and the effort of getting the computation to go forward.
The idea is that users can use any computer anywhere, and they access a central location, where their information is stored.
What this will really mean, though, is: people don't buy PCs to put their data on. They buy servers to put their data on. They've finally gotten their 24x7 connection, and they're going to stick a box in the basement next to the other utilities. Then they can go anywhere and get their desktop from their computer at home.
For the first iterations, the box will also be the PC they use, but it will act exactly the same as if it were storing things remotely. Of course, it will act as a thin client for other people who want to use other home machines, and it will be a server for when they're at their friends' houses.
Later, of course, they'll want more places to connect from than they want central stores, and they'll want to upgrade their client capabilities (graphics, raw processing, etc) at a different time from when they upgrade their server (long-term hard drive space), and the server will fall out of use as a client.
All this will be possible in... 1996. At least, that's when I started doing it, at least as far as ssh, and then screen, which covers most of the things I actually want to do with a centralized configuration. MicroSoft will probably end up with a similar model (using proprietary parts, of course, so MS client with MS server does things that either doesn't otherwise do). Being the server is probably a stopgap for people who aren't yet set up to run the servers themselves; after all, serving files for people isn't all that exciting, especially if they get people to use software subscriptions and pay to run the file servers on their own machines.
TUX does a good job of identifying inefficiencies in the network code, and of identifying at least one critical path that can be optimized. It doesn't actually do all that much which is particular to web serving; it's mainly a routine for efficiently getting a file from disk (or, most likely, cache) onto the network, with a bit of stuff that does HTTP. Doing the whole thing is the kernel is slightly useful for performance, but I think the main effect is that the whole thing gets written and optimized at once.
Normal web servers are two problems working together: the kernel and the web server. Having a single program doing it means that it's easier to track what's going on through the whole execution path, so end-to-end debugging is easier.
In any case, HTTP is significantly simpler than TCP and IP, which the kernel handle anyway. So most of your web server is really already in the kernel; it's just that, instead of responding to just pings and TCP control packets for you, it also responds to certain requests for you.