Universities are there to train people to do research and invent cool new things. Unfortunately, many institutes claiming to be a university are really technical schools whose only purpose is to learn students some very basic job skills. Indeed this is something that can be done individually by a disciplined, intelligent person. However, many people who claim to have done so are in fact idiots with a stack of VB for newbies books on their desks. They are the people being kicked out by bankrupt.coms (probably there is some correlation with whether these people actually had a degree).
Now, MIT is known for having produced some very bright people. Going there means you are educated by some very bright teachers. Getting a degree there means you are quite something. This is an experience you can't replicate by buying any book. Claiming that you can says more about you (dropped out of college?) than about universities.
I work at a university. I have some colleagues that came back to university after having worked in industry for a couple of years. All state as the reason that working in industry as a programmer did not satisfy their intellectual curiosity. To me working here is worth the lower salary. I pitty wage slaves coding 80 hours a week on stupid ecommerce sites, endlessly reinventing the wheel.
No idea, but since they are in the business of selling prepackaged gnome installations on CD, I can't imagine it will take them long to update their site. Rumor has it on debianplanet that they already have some.debs for nautilus. My guess is that they are finalizing red-carpet and will make the new Gnome available as a channel on that. Red-carpet is already very stable on my machine and I have no idea why they are keeping it back, other than perhaps waiting for gnome 1.4 to be released (which is happening as we speak probably).
Both problems can fixed easily. For KDE, point your browser to kde.org, find the page with the installation instructions, find the deb sources. Do a apt-get install task-kde. For Gnome, go to ximian.com, find the installation instructions (involves calling lynx and piping its contents to sh). It doesn't get easier. And mind you these instructions even work for potato. All this assumes you have a network connection of course. If you are on a modem, I can imagine you could just put the debs somewhere local.
So after ximian updates their site you (and I) will be able to upgrade. I love debian for this ease of use. A few days ago apt-get informed me that kde 2.1.1 was available, before that news reached me through the regualar news sites. 20 minutes later it was up and running (required no manual intervention at all, no questions asked).
I agree that it would be nice if debian was a little faster in adopting these things but as long as I can get them from third parties this way it is fine with me.
It is in their interest, it would be a reason for users to upgrade to XP. And considering there are very few other reasons to do so, it would be a good idea to include this feature. On the other hand, bluetooth seems to be a moving target right now and I agree with MS that it would be a bad idea to include a beta product with XP (assuming they have no intention of further delaying XP). IN addition, vendors of bluetooth enabled devices might include their own drivers for bluetooth, there are even open source versions of bluetooth drivers (ok they are for linux, not for win32).
You linux people begin to sound like government officials: not my job! go fuck yourself!
Just kidding:-), but seriously. Binaries are acceptable for just about anything but the kernel. It is good enough for me that the source code is available. I have no wish to have it on my hard drive though. I (and I'm pretty sure the majority of linux users) have no intention of editing a single line of it (and again, if we do we are capable of finding the ftp site). The compilation process is time consuming and not something for newbies. In addition it seems to consist mainly of choosing what to compile and what not to compile. Yet it is the only way to get important security updates and other important stuff.
What is it that puts the kernel aside from other software packages?
To answer a few concerns you point out:
"Should it support athlon? P-III? 386? All?"
Who owns a 386 these days anyway? The majority of users would be happy to use a 586 build (Mandrake gets their whole revenue from 586 builds!).
"Compiling a binary for each separate hardware is not the task of the kernel.org-people"
Assuming they run tests before they release, they should have some builds readily available (at least for commonly used platforms like i586). If not, I'm sure some linux distributers would be quite capable of compiling (I hope).
"Which is why you probably use mandrake, redhat or suse or similar and are happy with it."
I run debian. I must admit I like apt-get. But the install procedure is naive at best (debian developers are completely unaware of the concept of usability if you ask me).
I understand that for some people it is handy to have source code around, but why are kernels always distributed as source:
- they're huge downloads
- all you do with them is compiling (and hope you don't screw up somewhere)
- most configuration takes place after compilation (inserting modules and that sort of stuff).
I'd love to play with this new kernel release, however I don't have the time nor the patience to build the thing and figure out which modules I need.
You're not listening. If I drop icq I have to convince about 20 or so people to drop it too AND adopt the same im client as me . Since the whole point of having an icq client is to communicate with my friends I have no other option than to keep using icq (or compatible clients).
It's the same reason windows is still popular and it's the same reason linux on the desktop is going nowhere.
Unfortunately all my friends use icq, so I'm locked into icq. Unless of course I no longer want to communicate with my friends. The problem is not that there are no alternatives because there are and some are quite good actually. However, if you want to use an IM tool to actually communicate with somebody, the person on the other side will need a compatible client. And that's where AOL has us locked in since a great deal of people are using ICQ.
yes, I just had the deja vu feeling of getting a soundcard working in linux. In Beos and MS Windows it just works. It's a soundblaster 64, what's the fucking problem?
It turns out I have to fiddle a bit with the module settings, run red-hats sndconfig utility and since I'm on Debian, manually chmod the/dev audio devices in such a way that ordinary users (i.e. non root) can also enjoy sound. The only thing is, it took me two full days to get to the last part due to a gross lack of feedback as to what was actually going wrong. That is user unfriendly. That is linux. Some distributions handle this sort of thing much better these days but the bottom line is that most Linux distributions require you to do this type of fiddling to get the most trivial tasks done. Want the wheel on your mouse to work? Oh just change this and that obscure line in an even more obscure text file. Oh you want readable fonts in konqueror (ironically even the KDE website is unreadable without any tweaking)? Here's the enormously tedious procedure to get truetype fonts working.
Now don't get me wrong, I actually enjoy this kind of fiddling and was able to resolve all of the above issues (the reason I'm running Debian:). But to ordinary users who just expect their mp3 player to play their mp3 when they click play this is annoying.
Now you can say what you want about apple but the fact remains that while the UNIX community was talking about making UNIX user friendly, Apple got it done. I'm almost sorry I don't own a mac so I can't play with MacOS X.
This will mean less home users will use red-hat. This is a problem, because this is also the type of user that convinces their employers to adopt linux. And which distribution will they recommend? Exactly, the one they use at home. Red-hat won't get rich from this small fee. However, they will loose a lot of customers who will simply use something else since they don't like to cough up the money. It seems like a stupid thing to do.
The main source of revenue for Red Hat should be support contracts. A secondary source is retail sales of Red Hat CDs. Asking money for updates (a potential third revenue source)will harm those two revenue sources so it is a bad idea.
And the people who do realize don't realize they are not addressing any performance bottlenecks by using it. Get over it, native compilation does not solve many Java performance problems. There's a reason for native compilers such as towerJ and gcj not being used that often: the performance gain is not that big and sometimes not even there!
But I do Know. I've been working with various UNIX versions over the past few years. I tried BeOs, heck I even had a commodore 64 once. Trust me I've seen quite a few GUIs and UIs. While things have improved over the years, Unix never really caught up. Both Gnome, KDE and even enlightenment can be visually very nice. However, it's just a facade over a commandline interface. In fact the latest Gnome and KDE incarnations can hardly be considered as moderately more than a facade. Windows dropped the commandline interface long time ago (well it is still there but you rarely need it), apple probably never had one (I wouldn't know really since I don't use apples). Unix however has one and even more, requires one. The whole UNIX paradigm is based on pipes & filters, small tools, powerfull scripting and so on. In the hands of a trained professional like a programmer or system administrator it is a powerfull tool, specifically designed to make their job easier. However, making such a system usable for non pros requires more than a facade. I have to admit I like both Gnome and KDE. I see a lot of potential, but it's not finished yet.
Now I have seen a lot of Unix people like you. So stuck in the Unix paradigm that they are no longer even capable of using a windows UI. A colleague of mine has programmed C on solaris for years. You should see the poor bastard struggling with windows to accomplish even the most simplest tasks. However it is not windows that is the problem, it is is strong Unix bias. The problem Linux currently has is that the people designing the GUI are this type of persons. They try to understand but they just don't get it.
Whatever the qualities are of windows 9x/NT/2000, one thing is for sure: The windows GUI has been pretty much the same since 95 and most applications written for it integrate nicely with it. Some projects under linux have gone to great length to duplicate and even improve on the features offered by the windows GUI. However, none have come to the point yet where most linux apps integrate nicely. Up until KDE 2.0, the linux desktop lacked a decent filebrowser (which is why everybody chooses to use the commandline). Now with Nautilus, there are two, reasonably fully featured file browsers (the other is konqueror).
So what makes Windows better? Consistency and standardized behavior (industry standard of course, but then, are there any other?). Theming is a nice gimmick, and yes there are ways to do it on win32. However, more advanced users usually refrain from using such features. Better of course is entirely subjective. However to me it is clear that under windows people actually use the GUI whereas under linux you almost have to use the commandline to get anything non trivial done. As long as this is the case windows/apple will be the superior desktop for many users (including me).
Don't get me wrong, I have the latest Gnome and KDE running on the Debian PC next to me and I love them both. But surfing, reading mail and doing serious stuff like writing a presentation or a paper is much more comfortable under windows 2000 at this point in time.
PDF is about spreading information. Some information is under copyright. This copyright should be respected. If printing can be disabled, it may convince authors to distribute a non printable version for free (very usefull if you want to preview a book for instance). If this feature is bypassed by default by popular viewers, authors may choose not to distribute their material at all. That would be our loss, not theirs.
Now of course adobe's feature is implemented rather clumsily if it is that easy to bypass. No doubt, cracked versions of the viewer will be distributed anyway, so in a way it is already too late. But since debian hackers are self proclaimed moralists (all that bullshit about free speech rather than beer and so on), they should behave in a consistent way and respect other peoples copyright regardless of the technical possibilities. After all, they expect other people to do the same with their stuff.
But then, why would they? I should point out that there are java versions of the java compiler. I think jdk 1.1 used to ship with a java compiler written in java. So, if it fits in the memory of a cellphone, it can run there.
Sure it's cool but:
- most of the time involved with booting goes into starting operating services. My kernel takes about 1/2 seconds to get to the part where it starts doing stuff with modules and inetd, those live on my HD.
- it seems to be linux exclusively
- should you be booting that often?
I think all this is a start of what might be a good solution to an existing problem: the current BIOS is slow, does stuff that could be considered redundant and does so in a suboptimal way.
I would like to see sing a bit speedier, however my PC spends only 5 seconds or so in the BIOS. If this project can save me about 2 seconds (which is what they claim), that sure is great but I still have to sit through the rest of the booting process. Can't we speed that up? I would love to see a hibernate function as in W2k appear in linux for instance.
You're forgetting about all those little GUI glitches. No show stoppers, but they do become annoying if you use Mozilla long enough. I think it is good they inserted an additional development cycle. This will allow them to address some of the more annoying bugs. That, and they need to speed up news and mail. As it is now it is not very useful.
Qt is indeed platform independent, however this wrapper also wraps KDE stuff and as far as I know KDE only runs on UNIX (though it might be possible to port it).
I don't understand your argument about binding code into the JVM. Of course you can create 2 GUIs for your app and choose dynamically which one to use. However, in practice this limits the way you can design your code and will cause you to write a lot of non trivial gluecode. And why write a second GUI if you have a crossplatform swing based GUI anyway?
>You're confusing terms. Commonly: heavy==slow,large. light==fast,small. However, in terms of Java GUI components, heavyweight components are native (a la AWT), whilst lightweight components are rendered entriely by Java.
Lets elaborate a little here. Java leaves the rendering to Java2D. On windows this means that your widgets are actually rendered using directdraw (i.e. hardware accelerated) since the Java2D rendering is just a layer on top of platform specific rendering. What AWT did was much more ugly, it tried to wrap native components in such a way that it would be platform independent, it was ugly, buggy and never really became a good solution for crossplatform GUIs. That, by the way is also the trouble with this KDE wrapper, it will look nice on linux, but what about other platforms? It sounds like AWT2 to me, I don't see how the problems AWT had are addressed in any way by the KDE wrapping. But then it probably never was intended as a crossplatform solution. So in short, the KDE2 wrapper has the same inherent problems the original AWT had, yet it is posed as an alternative to the much better (than AWT) swing solution.
"things like server-class stability are largely irrelevant"
You couldn't be more wrong. Nobody wants to reset their palm PC and lose all their data. Typically the software on this type of machines is running continually. The machine is suspended rather than turned off. Because of this, stability is a very strong requirement.
Another important requirement that is also associated with BSD is security. Consider the following scenario: I walk into the office of my companies largest competitor. I'm carrying a PDA containing very sensitive data. The PDA is equiped with a bluetooth chip (high bandwidth) and I turn the thing on to make notes in my agenda. Tell me security is not important here:).
All I'm trying to say here is that some of the qualities that make BSD so suitable for servers are also required in a PDA.
Of course there are some limitations as well (memory, speed), but then pda's are generally equiped with faster processors and more memory than some PC's still running BSD in many offices.
As for world domination of the OS. That is an old centralistic view of world. In the modern view a hetrogenous network of all sorts of devices and services replaces it. The OS kernel is pretty much irrelevant as long as the programs running on top of it behave well and follow industry standards. I don't care whether slashdot runs on IIS or Apache, as long as I get the content in a reasonable time I'm happy.
I don't want to play disc jockey, that's the whole point of large storage devices. I currently store 650 Mb on a 1$ device (a cdrom). I don't see how my situation is improved significantly by this new invention.
I have no use for a device with content control. In a market with competition a variant of such a device without the content control will soon emerge or the invention will simply vaporize (like most storage related inventions seem to do). I read slashdot regularly and anouncements of the next generation storage devices (holographic storage, new and improved optical storage, better harddrive) are about as frequent as discussions on Gnu license issues. So, my guess is that this will fail (provided it ever evolves into a product which I doubt). BTW. 500 Mb isn't even close to the actual size of my mp3 collection, I need something larger.
Also take into account that you need to at least double time predictions given by a software engineer, that would bring us a fresh frozen woody in about 2003 (based on the very rough estimate of about a year in the article), so maybe it will be running kernel 2.6:)
Slashdot Mirror
Universities are there to train people to do research and invent cool new things. Unfortunately, many institutes claiming to be a university are really technical schools whose only purpose is to learn students some very basic job skills. Indeed this is something that can be done individually by a disciplined, intelligent person. However, many people who claim to have done so are in fact idiots with a stack of VB for newbies books on their desks. They are the people being kicked out by bankrupt .coms (probably there is some correlation with whether these people actually had a degree).
Now, MIT is known for having produced some very bright people. Going there means you are educated by some very bright teachers. Getting a degree there means you are quite something. This is an experience you can't replicate by buying any book. Claiming that you can says more about you (dropped out of college?) than about universities.
I work at a university. I have some colleagues that came back to university after having worked in industry for a couple of years. All state as the reason that working in industry as a programmer did not satisfy their intellectual curiosity. To me working here is worth the lower salary. I pitty wage slaves coding 80 hours a week on stupid ecommerce sites, endlessly reinventing the wheel.
No idea, but since they are in the business of selling prepackaged gnome installations on CD, I can't imagine it will take them long to update their site. Rumor has it on debianplanet that they already have some .debs for nautilus. My guess is that they are finalizing red-carpet and will make the new Gnome available as a channel on that. Red-carpet is already very stable on my machine and I have no idea why they are keeping it back, other than perhaps waiting for gnome 1.4 to be released (which is happening as we speak probably).
Both problems can fixed easily. For KDE, point your browser to kde.org, find the page with the installation instructions, find the deb sources. Do a apt-get install task-kde. For Gnome, go to ximian.com, find the installation instructions (involves calling lynx and piping its contents to sh). It doesn't get easier. And mind you these instructions even work for potato. All this assumes you have a network connection of course. If you are on a modem, I can imagine you could just put the debs somewhere local.
So after ximian updates their site you (and I) will be able to upgrade. I love debian for this ease of use. A few days ago apt-get informed me that kde 2.1.1 was available, before that news reached me through the regualar news sites. 20 minutes later it was up and running (required no manual intervention at all, no questions asked).
I agree that it would be nice if debian was a little faster in adopting these things but as long as I can get them from third parties this way it is fine with me.
It is in their interest, it would be a reason for users to upgrade to XP. And considering there are very few other reasons to do so, it would be a good idea to include this feature. On the other hand, bluetooth seems to be a moving target right now and I agree with MS that it would be a bad idea to include a beta product with XP (assuming they have no intention of further delaying XP). IN addition, vendors of bluetooth enabled devices might include their own drivers for bluetooth, there are even open source versions of bluetooth drivers (ok they are for linux, not for win32).
You linux people begin to sound like government officials: not my job! go fuck yourself!
:-), but seriously. Binaries are acceptable for just about anything but the kernel. It is good enough for me that the source code is available. I have no wish to have it on my hard drive though. I (and I'm pretty sure the majority of linux users) have no intention of editing a single line of it (and again, if we do we are capable of finding the ftp site). The compilation process is time consuming and not something for newbies. In addition it seems to consist mainly of choosing what to compile and what not to compile. Yet it is the only way to get important security updates and other important stuff.
Just kidding
What is it that puts the kernel aside from other software packages?
To answer a few concerns you point out:
"Should it support athlon? P-III? 386? All?"
Who owns a 386 these days anyway? The majority of users would be happy to use a 586 build (Mandrake gets their whole revenue from 586 builds!).
"Compiling a binary for each separate hardware is not the task of the kernel.org-people"
Assuming they run tests before they release, they should have some builds readily available (at least for commonly used platforms like i586). If not, I'm sure some linux distributers would be quite capable of compiling (I hope).
"Which is why you probably use mandrake, redhat or suse or similar and are happy with it."
I run debian. I must admit I like apt-get. But the install procedure is naive at best (debian developers are completely unaware of the concept of usability if you ask me).
I understand that for some people it is handy to have source code around, but why are kernels always distributed as source:
- they're huge downloads
- all you do with them is compiling (and hope you don't screw up somewhere)
- most configuration takes place after compilation (inserting modules and that sort of stuff).
I'd love to play with this new kernel release, however I don't have the time nor the patience to build the thing and figure out which modules I need.
You're not listening. If I drop icq I have to convince about 20 or so people to drop it too AND adopt the same im client as me . Since the whole point of having an icq client is to communicate with my friends I have no other option than to keep using icq (or compatible clients).
It's the same reason windows is still popular and it's the same reason linux on the desktop is going nowhere.
Unfortunately all my friends use icq, so I'm locked into icq. Unless of course I no longer want to communicate with my friends. The problem is not that there are no alternatives because there are and some are quite good actually. However, if you want to use an IM tool to actually communicate with somebody, the person on the other side will need a compatible client. And that's where AOL has us locked in since a great deal of people are using ICQ.
yes, I just had the deja vu feeling of getting a soundcard working in linux. In Beos and MS Windows it just works. It's a soundblaster 64, what's the fucking problem?
/dev audio devices in such a way that ordinary users (i.e. non root) can also enjoy sound. The only thing is, it took me two full days to get to the last part due to a gross lack of feedback as to what was actually going wrong. That is user unfriendly. That is linux. Some distributions handle this sort of thing much better these days but the bottom line is that most Linux distributions require you to do this type of fiddling to get the most trivial tasks done. Want the wheel on your mouse to work? Oh just change this and that obscure line in an even more obscure text file. Oh you want readable fonts in konqueror (ironically even the KDE website is unreadable without any tweaking)? Here's the enormously tedious procedure to get truetype fonts working.
:). But to ordinary users who just expect their mp3 player to play their mp3 when they click play this is annoying.
It turns out I have to fiddle a bit with the module settings, run red-hats sndconfig utility and since I'm on Debian, manually chmod the
Now don't get me wrong, I actually enjoy this kind of fiddling and was able to resolve all of the above issues (the reason I'm running Debian
Now you can say what you want about apple but the fact remains that while the UNIX community was talking about making UNIX user friendly, Apple got it done. I'm almost sorry I don't own a mac so I can't play with MacOS X.
This will mean less home users will use red-hat. This is a problem, because this is also the type of user that convinces their employers to adopt linux. And which distribution will they recommend? Exactly, the one they use at home. Red-hat won't get rich from this small fee. However, they will loose a lot of customers who will simply use something else since they don't like to cough up the money. It seems like a stupid thing to do.
The main source of revenue for Red Hat should be support contracts. A secondary source is retail sales of Red Hat CDs. Asking money for updates (a potential third revenue source)will harm those two revenue sources so it is a bad idea.
And the people who do realize don't realize they are not addressing any performance bottlenecks by using it. Get over it, native compilation does not solve many Java performance problems. There's a reason for native compilers such as towerJ and gcj not being used that often: the performance gain is not that big and sometimes not even there!
But I do Know. I've been working with various UNIX versions over the past few years. I tried BeOs, heck I even had a commodore 64 once. Trust me I've seen quite a few GUIs and UIs. While things have improved over the years, Unix never really caught up. Both Gnome, KDE and even enlightenment can be visually very nice. However, it's just a facade over a commandline interface. In fact the latest Gnome and KDE incarnations can hardly be considered as moderately more than a facade. Windows dropped the commandline interface long time ago (well it is still there but you rarely need it), apple probably never had one (I wouldn't know really since I don't use apples). Unix however has one and even more, requires one. The whole UNIX paradigm is based on pipes & filters, small tools, powerfull scripting and so on. In the hands of a trained professional like a programmer or system administrator it is a powerfull tool, specifically designed to make their job easier. However, making such a system usable for non pros requires more than a facade. I have to admit I like both Gnome and KDE. I see a lot of potential, but it's not finished yet.
Now I have seen a lot of Unix people like you. So stuck in the Unix paradigm that they are no longer even capable of using a windows UI. A colleague of mine has programmed C on solaris for years. You should see the poor bastard struggling with windows to accomplish even the most simplest tasks. However it is not windows that is the problem, it is is strong Unix bias. The problem Linux currently has is that the people designing the GUI are this type of persons. They try to understand but they just don't get it.
Whatever the qualities are of windows 9x/NT/2000, one thing is for sure: The windows GUI has been pretty much the same since 95 and most applications written for it integrate nicely with it. Some projects under linux have gone to great length to duplicate and even improve on the features offered by the windows GUI. However, none have come to the point yet where most linux apps integrate nicely. Up until KDE 2.0, the linux desktop lacked a decent filebrowser (which is why everybody chooses to use the commandline). Now with Nautilus, there are two, reasonably fully featured file browsers (the other is konqueror).
So what makes Windows better? Consistency and standardized behavior (industry standard of course, but then, are there any other?). Theming is a nice gimmick, and yes there are ways to do it on win32. However, more advanced users usually refrain from using such features. Better of course is entirely subjective. However to me it is clear that under windows people actually use the GUI whereas under linux you almost have to use the commandline to get anything non trivial done. As long as this is the case windows/apple will be the superior desktop for many users (including me).
Don't get me wrong, I have the latest Gnome and KDE running on the Debian PC next to me and I love them both. But surfing, reading mail and doing serious stuff like writing a presentation or a paper is much more comfortable under windows 2000 at this point in time.
PDF is about spreading information. Some information is under copyright. This copyright should be respected. If printing can be disabled, it may convince authors to distribute a non printable version for free (very usefull if you want to preview a book for instance). If this feature is bypassed by default by popular viewers, authors may choose not to distribute their material at all. That would be our loss, not theirs.
Now of course adobe's feature is implemented rather clumsily if it is that easy to bypass. No doubt, cracked versions of the viewer will be distributed anyway, so in a way it is already too late. But since debian hackers are self proclaimed moralists (all that bullshit about free speech rather than beer and so on), they should behave in a consistent way and respect other peoples copyright regardless of the technical possibilities. After all, they expect other people to do the same with their stuff.
Not to mention "network congestion" or "packet loss" :)
But then, why would they? I should point out that there are java versions of the java compiler. I think jdk 1.1 used to ship with a java compiler written in java. So, if it fits in the memory of a cellphone, it can run there.
Sure it's cool but:
- most of the time involved with booting goes into starting operating services. My kernel takes about 1/2 seconds to get to the part where it starts doing stuff with modules and inetd, those live on my HD.
- it seems to be linux exclusively
- should you be booting that often?
I think all this is a start of what might be a good solution to an existing problem: the current BIOS is slow, does stuff that could be considered redundant and does so in a suboptimal way.
I would like to see sing a bit speedier, however my PC spends only 5 seconds or so in the BIOS. If this project can save me about 2 seconds (which is what they claim), that sure is great but I still have to sit through the rest of the booting process. Can't we speed that up? I would love to see a hibernate function as in W2k appear in linux for instance.
You're forgetting about all those little GUI glitches. No show stoppers, but they do become annoying if you use Mozilla long enough. I think it is good they inserted an additional development cycle. This will allow them to address some of the more annoying bugs. That, and they need to speed up news and mail. As it is now it is not very useful.
And then you could fry the fish before catching it!
Qt is indeed platform independent, however this wrapper also wraps KDE stuff and as far as I know KDE only runs on UNIX (though it might be possible to port it).
I don't understand your argument about binding code into the JVM. Of course you can create 2 GUIs for your app and choose dynamically which one to use. However, in practice this limits the way you can design your code and will cause you to write a lot of non trivial gluecode. And why write a second GUI if you have a crossplatform swing based GUI anyway?
>You're confusing terms. Commonly: heavy==slow,large. light==fast,small. However, in terms of Java GUI components, heavyweight components are native (a la AWT), whilst lightweight components are rendered entriely by Java.
Lets elaborate a little here. Java leaves the rendering to Java2D. On windows this means that your widgets are actually rendered using directdraw (i.e. hardware accelerated) since the Java2D rendering is just a layer on top of platform specific rendering. What AWT did was much more ugly, it tried to wrap native components in such a way that it would be platform independent, it was ugly, buggy and never really became a good solution for crossplatform GUIs. That, by the way is also the trouble with this KDE wrapper, it will look nice on linux, but what about other platforms? It sounds like AWT2 to me, I don't see how the problems AWT had are addressed in any way by the KDE wrapping. But then it probably never was intended as a crossplatform solution. So in short, the KDE2 wrapper has the same inherent problems the original AWT had, yet it is posed as an alternative to the much better (than AWT) swing solution.
"things like server-class stability are largely irrelevant"
:).
You couldn't be more wrong. Nobody wants to reset their palm PC and lose all their data. Typically the software on this type of machines is running continually. The machine is suspended rather than turned off. Because of this, stability is a very strong requirement.
Another important requirement that is also associated with BSD is security. Consider the following scenario: I walk into the office of my companies largest competitor. I'm carrying a PDA containing very sensitive data. The PDA is equiped with a bluetooth chip (high bandwidth) and I turn the thing on to make notes in my agenda. Tell me security is not important here
All I'm trying to say here is that some of the qualities that make BSD so suitable for servers are also required in a PDA.
Of course there are some limitations as well (memory, speed), but then pda's are generally equiped with faster processors and more memory than some PC's still running BSD in many offices.
As for world domination of the OS. That is an old centralistic view of world. In the modern view a hetrogenous network of all sorts of devices and services replaces it. The OS kernel is pretty much irrelevant as long as the programs running on top of it behave well and follow industry standards. I don't care whether slashdot runs on IIS or Apache, as long as I get the content in a reasonable time I'm happy.
I don't want to play disc jockey, that's the whole point of large storage devices. I currently store 650 Mb on a 1$ device (a cdrom). I don't see how my situation is improved significantly by this new invention.
I have no use for a device with content control. In a market with competition a variant of such a device without the content control will soon emerge or the invention will simply vaporize (like most storage related inventions seem to do). I read slashdot regularly and anouncements of the next generation storage devices (holographic storage, new and improved optical storage, better harddrive) are about as frequent as discussions on Gnu license issues. So, my guess is that this will fail (provided it ever evolves into a product which I doubt). BTW. 500 Mb isn't even close to the actual size of my mp3 collection, I need something larger.
Also take into account that you need to at least double time predictions given by a software engineer, that would bring us a fresh frozen woody in about 2003 (based on the very rough estimate of about a year in the article), so maybe it will be running kernel 2.6 :)