This can't be about contractual obligations anymore: what Sun is shipping is completely different from what Microsoft and Sun made a contract over.
It is bad that Microsoft has been shipping a broken version of Java, and they should be ordered to stop. Beyond that, it's Sun's obligation to get their software onto people's desktops. AOL, which is also in competition with Microsoft, shows that it can be done.
It's only other big corporations with a lousy product and a large legal staff that get to force Microsoft to bundle their products.
In any case, it would have been entirely appropriate for the judge to order Microsoft to stop shipping MS "Java"--Microsoft was violating Sun's trademark and engaging in unfair business practices. But it was inappropriate for him to order Microsoft to ship Sun's software: Sun has to figure out how to do that just like other companies. And AOL shows that companies can successfully do that (AOL software is pretty much everywhere, despite the AOL/MSN competition).
Yes, what Microsoft is doing is bad. But the remedy is wrong.
First of all, unlike IE,.NET really, really is not a separate solution. Microsoft is going to be basing almost its entire operating system on it. That's actually a reasonable thing for them to do. It's not a competitor with Java in that sense.
Second, Java is just another proprietary solution, one that happens to come from Sun. Why should Sun get special treatment?
Third, both Sun and Netscape's products were more hype than reality. Netscape's browser lost in the market because it increasingly sucked compared to IE: it was slow, buggy, and failed to be standards compliant. It wasn't until Mozilla that it became competitive again. And what Sun has been doing with Java isn't exactly pretty either: Java has become bloated, and Sun has failed to deliver on numerous important promises. I used to be a strong supporter of Java, but Sun has been lying and failing to deliver for so long that I just have to say: don't touch Java. And I also have to say: don't touch.NET--it's still shrouded in legal uncertainty.
The correct remedy is to require Microsoft to stop shipping their broken version of Java and to stop exclusive distribution arrangements with PC vendors. But, ultimately, to actually get Java pre-installed on end user PCs, it is still Sun's responsibility to do the hard work of negotiating with PC vendors and creating attractive distributions for end users.
Then Microsoft should be required to unbundle their middleware. But forcing people to include Sun's middleware is not the right solution because it just increases the club of companies that can use coercion to get market share by one.
A partial analogy would be Microsoft owning the default Yellow Pages distributed to everyone's door and selecting who can be in it -- say, virtually everyone but "Sun."
Lots of people aren't included in Microsoft Windows distributions, and the ones that are presumably have committed to some serious contractual obligations. Why should Sun have it easier than everybody else?
The specific harm to Sun in this case is that Microsoft ships a broken version of Java, not that they fail to ship Sun's product for them.
If you don't think MS should have been sued in the first place, you will not believe any of these rationales, and probably not that antitrust is necessary in the first place.
Oh, I think Microsoft should have been sued, and I think there should be a remedy in this case and others. But the remedy is to require Microsoft to stop shipping their broken version of Java, not to force them to include Sun's. If Sun wants to get their version of Java into Windows, they can negotiate with Microsoft and computer manufacturers, and create compelling Java applications that end users will actually want to download.
In general, the best remedy for anti-trust problems is to create the conditions under which market forces can operate again. And that may mean breaking up the company or forcing it to unbundle its products. Forcing Microsoft to bundle Sun Java just replaces one company that forces software people with another company that forces software on people.
I see lots of lawsuits coming on. Now, gun manufacturers will be sued when the thing doesn't fire when the legitimate owner tries to defend himself, and when the thing does fire when some child points it the wrong way.
Apple's lowest price is $799 for a 600MHz G3 with 128M SDRAM, a 40G Ultra ATA, and a CD-ROM. Is that too much? For $629, you get a Dell Dimension 2350 with a 2GHz P4 (much faster than the 600MHz G3), 256M of RAM, 60G of disk, CD-ROM, and a 15" monitor.
I have a classic iMac, but I was willing to pay a premium for a fanless machine in a pretty case for a special application. But for general desktop use, the Dell hardware beats the iMac hardware in every way.
Except when that technology revolves around how to keep people alive, warm, oxygenated, etc.
What use is any such space technology to most people? The risks and dangers most people face on earth, for the most part, are completely different from the technology developed for space.
including in the long term the idea that man may someday be able to live in outer space.
We have a perfectly good planet here that's really comfortable for living on. We can do all our exploration by robotic probes for now. In a century or two, technology will have advanced so far that then manned space travel will be much easier. There is no need to hurry this. Our resources are better allocated elsewhere for now.
but heavy reliance on computers in the aircraft was a pretty foreign concept in the late 60s
That's my point. The integrated circuit was invented in 1958. If we had pushed on unmanned flight and unmanned space probes, VLSI and microprocessors would have advanced much earlier and much more quickly. The use of human pilots was a crutch that cost us dearly in terms of technology.
Put a man on Mars. Let's see where it goes from there.
It goes nowhere from there. Some very self-important person will walk around there, contaminate everything with earth microbes, return, and that will be that. In the process, we'll be wasting trillions of dollars and real exploration of the solar system will be held back by several decades. What a damnable waste.
To me, Flash is one of the most annoying web technologies around. It is distracting and hard to get rid of. Anything that makes it more proprietary and less cross-platform is good, as far as I'm concerned. If Microsoft acquires Macromedia, Flash may degenerate into something akin to ActiveX--used by die-hard Microsoft fanatics on their web pages but largely ignored by the mainstream.
There is a need for vector graphics on the web, but it is being filled by SVG. SVG is more standards-based, easier to generate, integrates better with the rest of the browser, and is easier to build tools for. And, hopefully, one can disable the "dynamic" bits of SVG.
The technology bleed-through from the moon explorations has paved the way for things of wide variety, like Oakley glassed, Tang, ceramic insulators, new engine fuels and designs, etc.
If we had aggressively pushed an unmanned space program using the same resources, artificial intelligence, robotics, power supplies, and integrated labs would be much further along than they are now. The PC revolution might have happened a decade earlier, and biotechnology, driven by work on exobiology and automation, might be further along as well. We might even have propulsion systems that would make a manned mission to Mars feasible. That's "real bleed-through".
Orange drinks and fashion accessories for astronauts are really trinkets in comparison, and would have been better developed by consumer companies.
Presenting the space missions with "heroes" also has advantages in maintaining public support, so that the massive technology thrust funded by NASA can continue to contribute to US technology commerce.
NASA could contribute more to US technology and commerce if they didn't waste as much money trying to figure out how to lift people's carcasses into orbit. And, these days, the US population, to the degree that they notice space at all, seems to be much more vowed by pretty pictures from other planets than by some Joe floating around in the space station having a glass of Tang. A color camera on a robotic lander is a much cheaper concession to good PR than manned flight.
Asking whether NASA should refute crackpots is the wrong question. Questions of whether the moon landing actually took place are symptomatic of a deeper problem. If NASA spends many billions of dollars on a project and all the average person gets out of it is a photo op that could have been staged at a Hollywood studio, it's no wonder that these questions come up. Refuting them at such a late point is too late.
NASA got itself into this problem by presenting itself as a frontier organization, a group of heroic explorers. And to maintain that image, they are wasting lots of money on useless projects like the space shuttle and the space station.
What should NASA do? They should present themselves as a scientific organization and forego the wild-west mentality. They should stop presenting astronauts as "heroes", reduce manned space travel to next-to-nothing, and instead go mostly with comparatively low-cost, unmanned probes. As you may have noticed, people don't generally ask whether unmanned probes are fake or not, and even if they did, nobody would really care very much.
And, of course, the other problem is that the US population isn't exactly up to speed on science, on average. Refuting a single crackpot is too little too late, but NASA should take its educational role in the sciences more seriously and they should get the funding to do it--they are trying, but they aren't making a dent.
If we had a scientifically literate population, and NASA stuck to doing science and didn't create a heroic mysticism around manned exploration, crackpots wouldn't stand a chance. The way it is, NASA is merely reaping what they sowed.
The deployment never happened, and that was reported pretty widely (e.g., here).
This wasn't a problem with Linux, it was a problem with politics and funding. And, frankly, in that kind of situation, I think Linux is better deployed through an incremental grass-roots effort anyway. The Danish approach seems better.
On the one hand, I'd really like to see an open source alternative to Outlook--something that fulfills the same functions and is easy for Outlook users to pick up.
On the other hand, I think Outlook-like programs are prime candidates for breaking with the straight-jacket of Windows-like GUIs. With sustained funding and free from the shackles of backwards compatibility with outmoded paradigms, an open source project, together with some HCI and information retrieval researchers, could really do something ground-breakingly better than anything Microsoft, or anybody else, is delivering.
Microsoft's PR department would probably let you know if some big institution that switched away from Windows or Office came back to it. And there are plenty of computer columnists and reporters who would have a field day with that as well.
While individuals and small installations may go back to Windows, I suspect that for most large installations, the cost advantages and reliability of Linux are so compelling that they tend to stay with it.
Stem cell rsearch and the recent discoveries about the function of small RNA molecules are great advances and will let us address some very serious medical problems.
However, I wouldn't count on a "world free from disease". First of all, we can't even deliver many medical advances to the people of the US, despite spending enormous amounts of money on medical care (and even with a more efficient medical insurance system, there would have to be limits). But the use of stem cells and manipulations of RNA are likely to remain labor intensive and require careful personal attention by skilled labor (we haven't even been able to automate much simpler body-related services like cutting hair, at least not well).
Second, even if we get complete control over biological processes, we aren't built to last. After how many brain stem cell transplants and regenerations has your personality changed dramatically? How many times can you carry out gene therapy until you spend more time patching up its errors than actual disease? Think of it this way: even software projects, where we have complete control over everything, eventually fall apart.
So, these techniques will give more medical options to the reasonably well-off (which I won't complain about), and they may have some spin-offs for the poor. They may increase out life span significantly. But I wouldn't get my hopes up that they will eliminate disease, even for people able to pay for them.
I'm sorry, but that's no way to run an ISP. An ISP should provide some degree of insulation of customers from one another. There are a lot worse things that people can do by accident, or deliberately, than "plug in a NAT backwards".
The iBook is more than twice as heavy, about twice as large, and has less than half the battery life. The iBook is a decent, small laptop, if you are willing to live with OS X or PPC Linux, but it's in a completely different class from the ultra-portable machines.
Direct search using various phonetic representations has been around for many years. All things being equal, it's known to be somewhat better than searching the output of speech recognizer using approximate string matching. But you have to weigh against that that both speech recognition and approximate string matching are being pushed much harder than this kind of search, so you may end up getting better performance using speech recognition and string searching anyway.
Regardless of HOW MUCH time and resources it uses, its still use of resources which are neither infinite or cheap. Due to the nature of SMTP, the vast majority of these resources are consumed on the receiver's end.
All the spammers in the world don't have the bandwidth to even make a dent in the aggregate bandwidth available to spam recipients. The only reason spam is a resource problem is that it is funneled through bottlenecks at the ISP.
I call bullshit. Our company does not provide internet access to most employees for example. Out of our monthly bandwidth usage, 80% would be e-mail. If 50% of that is unsolicited (ie, spam) thats 40% of our bandwidth costs directly attributed to spam. 40% of our bill that is UNWANTED traffic.
So what? 50% of not much bandwidth is still not much bandwidth.
You end up with a mass of poorly configured open relays, happily flooding everybody else with even more spam.
Why would those machines run mailers that can even be configured to relay? The enormous numbers of misfeatures in our current mail servers is just another expression of an infrastructure that is fundamentally misdesigned.
I suggest you leave arguments regarding things you have no clue about to those who do...
What's broken with our mail system is that it is based on a small number of centralized servers and bottlenecks. Of course, spam is going to be a problem for those servers. The solution is to get rid of the centralization and bottlenecks, not fight the inevitable.
Or, to use your own language, it's idiots like you who can't see beyond their own navel that perpetuate this problem.
The C-700 gives you a very restricted Linux environment with applications designed for a handheld running on a non-x86 processor and a keyboard that's probably murder to type on.
Although I haven't used it myself, from the specs, the Fujitsu Lifebook P1000 looks like a much better alternative, and at $1200 it isn't all that much more expensive. With an extended battery, it runs for 9 hours. Sony's Picturebook series is another ultra-portable choice.
Linux is nothing like Windows: its kernel, its system administration, its core software, its graphics and GUI, are all of a completely different design from Windows.
The Windows kernel is an all-singing all-dancing behemoth, while the Linux kernel APIs still mostly stick to the minimalism of the original UNIX design. Windows uses object-orientation extensively throughout its kernel, while Linux sticks with the relative paucity of APIs, again from the traditional UNIX kernel design.
In userland, things look rather different, too. The core set of applications on Linux are text-based, command-line oriented programs that are combined via pipes and files and store their data in text files. Few programs use threads. Servers can be run from the command line. Alsmot can be scripted from the shell. This is in sharp contrast to Windows, where the core sets of applications are GUI-based, component-oriented programs talking to each other via various shared memory and object embeddings, use threads extensively, and use databases for a lot of their data. The Windows design a CS major's wet dream, implementing every software feature and ad-hoc idea under the sun, while core Linux programs stick, again, to the simple principles of the traditional UNIX design.
Only when it comes to desktop software, like Gnome, KDE, OpenOffice, and Mozilla, has the Linux world copied liberally from Windows. The resulting desktop software is very non-UNIX like and has many of the same limitations as its Windows counterparts. Still, the strategy of cloning Windows features is a necessity, because people coming to Linux from Windows want a more familiar environment. Fortunately, no matter how many kludgy ideas something like Mozilla inherits from Windows, the underlying modularity and simplicity of Linux and X11 mean that those Windows-applications-clones can live peacefully side-by-side with implementations of new ideas in GUI software and user interaction.
Now, let's get to the meat of it:
So what needs to happen? First of all, the desktop-window metaphor has had a good run and has its place, but can't we try something different?
Yes, and Linux is the best place for this to happen right now. Because, unlike Windows and Macintosh, where assumptions about the GUI are coded throughout the system, Linux and X11 are highly layered: you can build an entirely different user environment on top of Linux and X11 and still take advantage of a vast amount of existing Linux and UNIX software out there. People on UNIX and Linux have made "software components" and "software reuse" work in a way none of the mega-platform-projects at Microsoft, Apple, IBM, or NeXT ever have (remember Pink?).
Linux can and does successfully imitate Windows in some of its distributions. But it also makes it easy to build completely different systems. And that is why I think Linux will be the platform where the next true innovations in human computer interaction will be delivered.
Slashdot Mirror
It is bad that Microsoft has been shipping a broken version of Java, and they should be ordered to stop. Beyond that, it's Sun's obligation to get their software onto people's desktops. AOL, which is also in competition with Microsoft, shows that it can be done.
In any case, it would have been entirely appropriate for the judge to order Microsoft to stop shipping MS "Java"--Microsoft was violating Sun's trademark and engaging in unfair business practices. But it was inappropriate for him to order Microsoft to ship Sun's software: Sun has to figure out how to do that just like other companies. And AOL shows that companies can successfully do that (AOL software is pretty much everywhere, despite the AOL/MSN competition).
First of all, unlike IE, .NET really, really is not a separate solution. Microsoft is going to be basing almost its entire operating system on it. That's actually a reasonable thing for them to do. It's not a competitor with Java in that sense.
Second, Java is just another proprietary solution, one that happens to come from Sun. Why should Sun get special treatment?
Third, both Sun and Netscape's products were more hype than reality. Netscape's browser lost in the market because it increasingly sucked compared to IE: it was slow, buggy, and failed to be standards compliant. It wasn't until Mozilla that it became competitive again. And what Sun has been doing with Java isn't exactly pretty either: Java has become bloated, and Sun has failed to deliver on numerous important promises. I used to be a strong supporter of Java, but Sun has been lying and failing to deliver for so long that I just have to say: don't touch Java. And I also have to say: don't touch .NET--it's still shrouded in legal uncertainty.
The correct remedy is to require Microsoft to stop shipping their broken version of Java and to stop exclusive distribution arrangements with PC vendors. But, ultimately, to actually get Java pre-installed on end user PCs, it is still Sun's responsibility to do the hard work of negotiating with PC vendors and creating attractive distributions for end users.
Then Microsoft should be required to unbundle their middleware. But forcing people to include Sun's middleware is not the right solution because it just increases the club of companies that can use coercion to get market share by one.
Lots of people aren't included in Microsoft Windows distributions, and the ones that are presumably have committed to some serious contractual obligations. Why should Sun have it easier than everybody else?
The specific harm to Sun in this case is that Microsoft ships a broken version of Java, not that they fail to ship Sun's product for them.
If you don't think MS should have been sued in the first place, you will not believe any of these rationales, and probably not that antitrust is necessary in the first place.
Oh, I think Microsoft should have been sued, and I think there should be a remedy in this case and others. But the remedy is to require Microsoft to stop shipping their broken version of Java, not to force them to include Sun's. If Sun wants to get their version of Java into Windows, they can negotiate with Microsoft and computer manufacturers, and create compelling Java applications that end users will actually want to download.
In general, the best remedy for anti-trust problems is to create the conditions under which market forces can operate again. And that may mean breaking up the company or forcing it to unbundle its products. Forcing Microsoft to bundle Sun Java just replaces one company that forces software people with another company that forces software on people.
I see lots of lawsuits coming on. Now, gun manufacturers will be sued when the thing doesn't fire when the legitimate owner tries to defend himself, and when the thing does fire when some child points it the wrong way.
I can't get particularly excited about the iBook. With a 600MHz G3, it's rather slow, and at about 5 pounds, it's pretty heavy.
A Sony VAIO SRX99 makes a much better ultra-portable laptop: longer battery life, more memory, much lighter and thinner at 2.76 pounds.
And something like the HP ze4101 gives you a much better general purpose laptop at less money and little extra weight.
I have a classic iMac, but I was willing to pay a premium for a fanless machine in a pretty case for a special application. But for general desktop use, the Dell hardware beats the iMac hardware in every way.
What use is any such space technology to most people? The risks and dangers most people face on earth, for the most part, are completely different from the technology developed for space.
including in the long term the idea that man may someday be able to live in outer space.
We have a perfectly good planet here that's really comfortable for living on. We can do all our exploration by robotic probes for now. In a century or two, technology will have advanced so far that then manned space travel will be much easier. There is no need to hurry this. Our resources are better allocated elsewhere for now.
but heavy reliance on computers in the aircraft was a pretty foreign concept in the late 60s
That's my point. The integrated circuit was invented in 1958. If we had pushed on unmanned flight and unmanned space probes, VLSI and microprocessors would have advanced much earlier and much more quickly. The use of human pilots was a crutch that cost us dearly in terms of technology.
Put a man on Mars. Let's see where it goes from there.
It goes nowhere from there. Some very self-important person will walk around there, contaminate everything with earth microbes, return, and that will be that. In the process, we'll be wasting trillions of dollars and real exploration of the solar system will be held back by several decades. What a damnable waste.
Those are good attributes, as far as I'm concerned. I don't want "animation", I want vector graphics and simple interaction.
There is a need for vector graphics on the web, but it is being filled by SVG. SVG is more standards-based, easier to generate, integrates better with the rest of the browser, and is easier to build tools for. And, hopefully, one can disable the "dynamic" bits of SVG.
If we had aggressively pushed an unmanned space program using the same resources, artificial intelligence, robotics, power supplies, and integrated labs would be much further along than they are now. The PC revolution might have happened a decade earlier, and biotechnology, driven by work on exobiology and automation, might be further along as well. We might even have propulsion systems that would make a manned mission to Mars feasible. That's "real bleed-through".
Orange drinks and fashion accessories for astronauts are really trinkets in comparison, and would have been better developed by consumer companies.
Presenting the space missions with "heroes" also has advantages in maintaining public support, so that the massive technology thrust funded by NASA can continue to contribute to US technology commerce.
NASA could contribute more to US technology and commerce if they didn't waste as much money trying to figure out how to lift people's carcasses into orbit. And, these days, the US population, to the degree that they notice space at all, seems to be much more vowed by pretty pictures from other planets than by some Joe floating around in the space station having a glass of Tang. A color camera on a robotic lander is a much cheaper concession to good PR than manned flight.
NASA got itself into this problem by presenting itself as a frontier organization, a group of heroic explorers. And to maintain that image, they are wasting lots of money on useless projects like the space shuttle and the space station.
What should NASA do? They should present themselves as a scientific organization and forego the wild-west mentality. They should stop presenting astronauts as "heroes", reduce manned space travel to next-to-nothing, and instead go mostly with comparatively low-cost, unmanned probes. As you may have noticed, people don't generally ask whether unmanned probes are fake or not, and even if they did, nobody would really care very much.
And, of course, the other problem is that the US population isn't exactly up to speed on science, on average. Refuting a single crackpot is too little too late, but NASA should take its educational role in the sciences more seriously and they should get the funding to do it--they are trying, but they aren't making a dent.
If we had a scientifically literate population, and NASA stuck to doing science and didn't create a heroic mysticism around manned exploration, crackpots wouldn't stand a chance. The way it is, NASA is merely reaping what they sowed.
The deployment never happened, and that was reported pretty widely (e.g., here). This wasn't a problem with Linux, it was a problem with politics and funding. And, frankly, in that kind of situation, I think Linux is better deployed through an incremental grass-roots effort anyway. The Danish approach seems better.
On the other hand, I think Outlook-like programs are prime candidates for breaking with the straight-jacket of Windows-like GUIs. With sustained funding and free from the shackles of backwards compatibility with outmoded paradigms, an open source project, together with some HCI and information retrieval researchers, could really do something ground-breakingly better than anything Microsoft, or anybody else, is delivering.
While individuals and small installations may go back to Windows, I suspect that for most large installations, the cost advantages and reliability of Linux are so compelling that they tend to stay with it.
However, I wouldn't count on a "world free from disease". First of all, we can't even deliver many medical advances to the people of the US, despite spending enormous amounts of money on medical care (and even with a more efficient medical insurance system, there would have to be limits). But the use of stem cells and manipulations of RNA are likely to remain labor intensive and require careful personal attention by skilled labor (we haven't even been able to automate much simpler body-related services like cutting hair, at least not well).
Second, even if we get complete control over biological processes, we aren't built to last. After how many brain stem cell transplants and regenerations has your personality changed dramatically? How many times can you carry out gene therapy until you spend more time patching up its errors than actual disease? Think of it this way: even software projects, where we have complete control over everything, eventually fall apart.
So, these techniques will give more medical options to the reasonably well-off (which I won't complain about), and they may have some spin-offs for the poor. They may increase out life span significantly. But I wouldn't get my hopes up that they will eliminate disease, even for people able to pay for them.
I'm sorry, but that's no way to run an ISP. An ISP should provide some degree of insulation of customers from one another. There are a lot worse things that people can do by accident, or deliberately, than "plug in a NAT backwards".
The iBook is more than twice as heavy, about twice as large, and has less than half the battery life. The iBook is a decent, small laptop, if you are willing to live with OS X or PPC Linux, but it's in a completely different class from the ultra-portable machines.
The Libretto isn't officially sold or supported in the US, and it's $600 more. I don't think it's a real alternative to either.
Direct search using various phonetic representations has been around for many years. All things being equal, it's known to be somewhat better than searching the output of speech recognizer using approximate string matching. But you have to weigh against that that both speech recognition and approximate string matching are being pushed much harder than this kind of search, so you may end up getting better performance using speech recognition and string searching anyway.
NeXT used GNU C, and Objective-C was much more of a competitive feature of NeXT than a bunch of modifications to a kernel that was open source anyway.
And it's not like NeXT really had much choice anyway: there weren't a lot of other kernels they could have used around back then.
All the spammers in the world don't have the bandwidth to even make a dent in the aggregate bandwidth available to spam recipients. The only reason spam is a resource problem is that it is funneled through bottlenecks at the ISP.
I call bullshit. Our company does not provide internet access to most employees for example. Out of our monthly bandwidth usage, 80% would be e-mail. If 50% of that is unsolicited (ie, spam) thats 40% of our bandwidth costs directly attributed to spam. 40% of our bill that is UNWANTED traffic.
So what? 50% of not much bandwidth is still not much bandwidth.
You end up with a mass of poorly configured open relays, happily flooding everybody else with even more spam.
Why would those machines run mailers that can even be configured to relay? The enormous numbers of misfeatures in our current mail servers is just another expression of an infrastructure that is fundamentally misdesigned.
I suggest you leave arguments regarding things you have no clue about to those who do...
What's broken with our mail system is that it is based on a small number of centralized servers and bottlenecks. Of course, spam is going to be a problem for those servers. The solution is to get rid of the centralization and bottlenecks, not fight the inevitable.
Or, to use your own language, it's idiots like you who can't see beyond their own navel that perpetuate this problem.
Although I haven't used it myself, from the specs, the Fujitsu Lifebook P1000 looks like a much better alternative, and at $1200 it isn't all that much more expensive. With an extended battery, it runs for 9 hours. Sony's Picturebook series is another ultra-portable choice.
Linux is nothing like Windows: its kernel, its system administration, its core software, its graphics and GUI, are all of a completely different design from Windows.
The Windows kernel is an all-singing all-dancing behemoth, while the Linux kernel APIs still mostly stick to the minimalism of the original UNIX design. Windows uses object-orientation extensively throughout its kernel, while Linux sticks with the relative paucity of APIs, again from the traditional UNIX kernel design.
In userland, things look rather different, too. The core set of applications on Linux are text-based, command-line oriented programs that are combined via pipes and files and store their data in text files. Few programs use threads. Servers can be run from the command line. Alsmot can be scripted from the shell. This is in sharp contrast to Windows, where the core sets of applications are GUI-based, component-oriented programs talking to each other via various shared memory and object embeddings, use threads extensively, and use databases for a lot of their data. The Windows design a CS major's wet dream, implementing every software feature and ad-hoc idea under the sun, while core Linux programs stick, again, to the simple principles of the traditional UNIX design.
Only when it comes to desktop software, like Gnome, KDE, OpenOffice, and Mozilla, has the Linux world copied liberally from Windows. The resulting desktop software is very non-UNIX like and has many of the same limitations as its Windows counterparts. Still, the strategy of cloning Windows features is a necessity, because people coming to Linux from Windows want a more familiar environment. Fortunately, no matter how many kludgy ideas something like Mozilla inherits from Windows, the underlying modularity and simplicity of Linux and X11 mean that those Windows-applications-clones can live peacefully side-by-side with implementations of new ideas in GUI software and user interaction.
Now, let's get to the meat of it:
So what needs to happen? First of all, the desktop-window metaphor has had a good run and has its place, but can't we try something different?
Yes, and Linux is the best place for this to happen right now. Because, unlike Windows and Macintosh, where assumptions about the GUI are coded throughout the system, Linux and X11 are highly layered: you can build an entirely different user environment on top of Linux and X11 and still take advantage of a vast amount of existing Linux and UNIX software out there. People on UNIX and Linux have made "software components" and "software reuse" work in a way none of the mega-platform-projects at Microsoft, Apple, IBM, or NeXT ever have (remember Pink?).
Linux can and does successfully imitate Windows in some of its distributions. But it also makes it easy to build completely different systems. And that is why I think Linux will be the platform where the next true innovations in human computer interaction will be delivered.