That's I think where Apollo, Silverlight, or JavaFX want to take us; it'd be nice if there was an OSS stack (like the LAMP stack) that matched these capabilities...
How about Linux / GNOME / Firefox / Java applets? It's been done for ages and found wanting. Any of those three options you listed had better go through and see why Java isn't used for this, and the challenges Java had to get through to go as far as it did.
I don't know how sure the GP is, but I'm pretty damn sure Wine will be packaged shortly, if it isn't already. The guy who makes the packages currently was just accepted as a Ubuntu Developer. But don't take my word on whether it's packaged or not, check the repos.
The biggest problem with distro packaging isn't that you cant get them to accept random super package managers (which is every bit a social problem as it is technical, and MUST be approached that way if you expect ANY progress), it's that not all projects run well on a six month stable release cycle. Theoretically backport repos could address this, but there's a balance between new features and not crashing that users expect you to find and upstream developers are untrained in. This balance is pretty hard to find within a week of testing.
But WINE is a good example: generally speaking WINE is a work in progress. For some uses it's fantastic, for others it's not. I've rarely had an upgrade of WINE break previously working apps (but I'm sure it's happened). So I'd much prefer new packages sooner rather than later. To the extent that WINE should probably get updates in backports, and if you dont want them turn that repo off or pin the package. dpkg is pretty flexible and I'm still learning the ins and outs of it's power.
It's not very hard to imagine. I can even give you a simple scenario. Joe-bob wants a program to do something. If we're lucky, he looks at what programs are visible in the Applications Menu, and finds it. We're not; the function he wants isn't installed by default. But fortunately Joe-bob remembers a program he used often in windows. So he does what windows users do whenever they need more software: he downloads the program off the internet, and runs the installer.
At this point GNOME detects the.exe and mime type and decides WINE should be launched. Which works reasonably well for most software. Except installers. Installshield is a real bitch and won't work without copying files from the Windows partition he doesn't have on his Dell machine. Suddenly, in a very real sense, all these projects with the "well you need a Windows license to use this, but who doesn't have one?" clause have a solid counter-example.
Anyways, the installer crashes, and Dell gets another call asking why his software doesn't work, when X Y and Z do. The obvious answer (why should Windows programs work in Linux any more than Mac programs?) is sad, but any better solution requires the sort of Apple like Reality Distortion Field coupled with mac user like Devotion to work. WINE is not parallels (at least in quality). It's still great software that I'm amazed by more and more every month, but I wouldn't put it on laptops sold if I was Dell. Not without the sort of unrecoverable investment Dell is unlikely to make.
In summary, if you provide WINE, you'd better put your money where your mouth is and make it damn near flawless. If you don't provide WINE, you'd probably still need to educate switchers. But I don't expect any "CIS 101: Introduction to the Linux Desktop" classes any time soon.
More importantly, JavaOne is going on right now, which Sun uses to promote all their products under the guise of a conference on Java. So it's no surprise that so many "big announcements" are coming up on Slashdot. Just remember that slashdot is driven by users, many of whom are probably attending JavaOne right now.
The USA has certain properties that allow the world's best to flourish. We have a strong university system that creates a massive "Brain Drain" whereby many of the best students researchers wind up here. The rules to create a new company are simple enough that overseas entrepreneurs come here to start and do business. In a world with over 4 billion people, excellence almost requires extraordinary dedication, and this immigration reinforces the performance culture. But most important is that people are rewarded in proportion to their productivity, in other words, you're not sacrificing quality of life for work. Without that aspect, the French work week makes perfect sense.
At least for me, the hardest part of being productive is the first hour. After the first hour or so I'm engrossed enough to where I can be productive for several hours. It's an unfortunate aspect of a somewhat creative process like programming. It's much harder to force myself to think about things than force myself to do manual labor, probably because most such work can be done while letting your mind wander.
You keep repeating this story to Slashdot, but I'm missing any credible reference to it. Perhaps you could provide a few more specifics besides the name RMS?
Which is to say, after having worked with and seen plenty more international teams, that other cultures find dedication to excellence a shocking American practice and expectation. I have no idea why this is so, when every example of its absence results in mediocrity.
Taking credit for the whole project was not what the parent was talking about, rather taking responsibility to see a good idea come through in the face of opposition to it. Ideally you don't want just one such person, but two or three. This might make the rest of the team pointless, as it seems like people best able to promote a technology within are also best at using it. At which point I can understand why international teams find this stuff shocking, in the same way that 2.0 "passed with a C" engineers wind up as a "Sales Engineer".
So why doesn't one of the candidates call people out on it? Surely there's room in a ten man race for an outspoken, tell-all-apologize-for-it-later approach that you want.
Also, I think the majority watches CNN on accident. As in "oops, I meant channel 63, not 36" accident. I don't think they spend enough time watching to bother forming opinions on things, as long as the impact is distant and abstract.
The answer is simpler than you'd think. The graduating class will not be well rounded. For a while my department had a natural science with lab requirement. A lot of people took geology (apparently the easiest way to satisfy this), I took chemistry, and a lot of the double major engineering types took Physics. Almost nobody took biology. I know I wouldn't have if they hadn't added it on top of the science requirement. Partly this IS the biology field's problem: everyone perceives it as classification/soft science, and teaching focuses on memorization (quick, what are the major kinds of biological molecules? A: Lipids, proteins, and nucleic acids, if I recall). For engineers this is the most boring aspect. Relationships, especially mathematical ones, are most interesting to us. My perception is that if you like science but hate math, biology is the field to go into. If you disagree, go check your college's Biology undergrad math requirements.
The good news is that I was more prepared to understanding what folding@home etc do when they came along. And I'm taking a class now that deals with genomics. The course deals with algorithmic tools for biologists, and while the course is half biology, half computer science, the biologists struggle to keep up with the maths involved in things like BLAST searches (it finds genetic matches for sequences). And the software tools are often horrible.
For my final project we tested a heuristic based SNP prediction tool that we felt was similar to a statistical method another paper presented. The software was horrible. A perl script, written in traditional fuck-you readable perl style. At first I thought something was horribly wrong with our pipeline feeding it data, since it crashed on a divide by zero error. I went about making it continue when a div by zero would have happened, only to discover that the program no longer halted. A div by zero error was the terminating condition! I only shudder to think what else was hiding in that code, but the results were simple to present: it runs fine on its own test data, but the data we had from an insect failed to present any SNPs. There still could be problems with our input to it, but at least I'm sure now the data's not responsible for the div by zero error.
Biologists are generally not best of breed programmers, though I know of a few exceptions. And CS students appear to be avoiding the field--a year after the change, at the next cirriculum review, the Biology requirement was removed because of loud student complaint. And so our students remain "well rounded", by which I mean they all take the same courses that cover much of the same ground they've already been exposed to.
The thing about AI is that what it is keeps changing. Twenty years ago, Expert Systems was AI. Chess programs used to be considered AI, and some people still do, but generally the pattern seems to be that AI is about unsolved problems. Once we solve them, it's no longer intelligent. Some of this I can understand, simply because as we discover how things work they lose some mystery and intelligence is one of the most mysterious properties we assign to humanity. But very little of the early stuff tried to do learning. As if once we knew how to win at chess, we could take that and apply it to the rest of the universe, treating it like a bigger instance of a chess game.
There's a couple reasons for it not being in the kernel. First, it misleads users who expect some degree of data security. The good news is that sort of person likely follows kernel patches to the FS and would likely be aware of the problem, possibly even writing a script that replaces rm with a real-rm.
The second argument is that it's better handled in user space, so the OS doesn't have to make that sort of policy. There's no reason you can't just alias rm to some.Trash, or configure your Desktop Environment to do so (GNOME does, for example). There's all sorts of things you have to decide that might not suit everyone. For example, if I delete a file on a USB drive, does it go in a.Trash storage in the USB drive, or do we copy it over to a main.Trash folder? Many people don't realize they have to empty the trash to reclaim space on their thumbdrive in GNOME.
The final argument I can come up with is security problems. We can't have one global.Trash bin in a multiuser system. And quotas. And permissions.
Reading historic archives of the LKML suggests it's at least come up once. I guess Torvald's opinion is that anything that CAN go in the userspace SHOULD. Can't explain the webserver in kernel though. Perhaps that opinion has changed some time in the last 10 years?
You have to admit the line "this Computer Scientist had never taken a CS class! He was just good at algorithms" makes you sound retarded, since the only constant among the broad regions you've described is precisely algorithms.
All you've done is demonstrate that most all the choices you have on Linux are there on Microsoft too. In addition to the choices that don't exist on the Linux platform. I didn't really mean to harp on the paid aspect quite as much, though I'm sure that someone could highlight a few more commercial tools.
What I meant by java is that should you really want to, you can run your entire large scale applcation on java. iPlanet, tomcat, java db, servlets and jsps, netBeans, the Java language, the JVM, etc. Nobody uses all of this, because there's better choices out there. Postgresql, for example works well. I can't really speak to what developers with.NET use, though the author suggests.NET developers stay with MS tools. Point is, if you're looking for best of breed, you have to stop caring about vertical integration. You have the choice to use only parts of the vertical integration.
On an unrelated note, I wasn't aware the.NET SDK was free. Interesting, but I can't see the justification in using F# over Ocaml. Of course, if you're using.NET you might as well get an MSDN kit and visual studio, so I can see where they're going with it.
There's plenty of choice on Windows. The only difference is that these choices involve paying money for things whose worth you can't evaluate until you've used them for longer than a month. Branding helps tremendously in such a situation, as does bundling, both of which MS has in spades.
Some examples of choices developers have on windows platforms: * IDEs - visual studio, eclipse, netbeans, dev-c++, codewarrior, just to name a few I've used * The various.NET languages * Databases * Webservers, IIS, apache, or something else? * antivirus, Vista tried pretty hard to end all of these though.
If you're just moaning about how Microsoft has a large vertically integrated set of tools, well, there's Java. Nobody does this, because its stupid and they have the choice not to.
Given the mix of master's international students that remain in the US, I'd say only marginally better. The good news is they communicate well with their foreign counterparts, and India has some history with English.
Slashdot Mirror
How about Linux / GNOME / Firefox / Java applets? It's been done for ages and found wanting. Any of those three options you listed had better go through and see why Java isn't used for this, and the challenges Java had to get through to go as far as it did.
I don't know how sure the GP is, but I'm pretty damn sure Wine will be packaged shortly, if it isn't already. The guy who makes the packages currently was just accepted as a Ubuntu Developer. But don't take my word on whether it's packaged or not, check the repos.
The biggest problem with distro packaging isn't that you cant get them to accept random super package managers (which is every bit a social problem as it is technical, and MUST be approached that way if you expect ANY progress), it's that not all projects run well on a six month stable release cycle. Theoretically backport repos could address this, but there's a balance between new features and not crashing that users expect you to find and upstream developers are untrained in. This balance is pretty hard to find within a week of testing.
But WINE is a good example: generally speaking WINE is a work in progress. For some uses it's fantastic, for others it's not. I've rarely had an upgrade of WINE break previously working apps (but I'm sure it's happened). So I'd much prefer new packages sooner rather than later. To the extent that WINE should probably get updates in backports, and if you dont want them turn that repo off or pin the package. dpkg is pretty flexible and I'm still learning the ins and outs of it's power.
It's not very hard to imagine. I can even give you a simple scenario. Joe-bob wants a program to do something. If we're lucky, he looks at what programs are visible in the Applications Menu, and finds it. We're not; the function he wants isn't installed by default. But fortunately Joe-bob remembers a program he used often in windows. So he does what windows users do whenever they need more software: he downloads the program off the internet, and runs the installer.
.exe and mime type and decides WINE should be launched. Which works reasonably well for most software. Except installers. Installshield is a real bitch and won't work without copying files from the Windows partition he doesn't have on his Dell machine. Suddenly, in a very real sense, all these projects with the "well you need a Windows license to use this, but who doesn't have one?" clause have a solid counter-example.
At this point GNOME detects the
Anyways, the installer crashes, and Dell gets another call asking why his software doesn't work, when X Y and Z do. The obvious answer (why should Windows programs work in Linux any more than Mac programs?) is sad, but any better solution requires the sort of Apple like Reality Distortion Field coupled with mac user like Devotion to work. WINE is not parallels (at least in quality). It's still great software that I'm amazed by more and more every month, but I wouldn't put it on laptops sold if I was Dell. Not without the sort of unrecoverable investment Dell is unlikely to make.
In summary, if you provide WINE, you'd better put your money where your mouth is and make it damn near flawless. If you don't provide WINE, you'd probably still need to educate switchers. But I don't expect any "CIS 101: Introduction to the Linux Desktop" classes any time soon.
More importantly, JavaOne is going on right now, which Sun uses to promote all their products under the guise of a conference on Java. So it's no surprise that so many "big announcements" are coming up on Slashdot. Just remember that slashdot is driven by users, many of whom are probably attending JavaOne right now.
The obvious answer is because it's easy.
It could be worse. My university has chosen to write their own software, and is now inflicting their creation on other schools by selling it to them.
The USA has certain properties that allow the world's best to flourish. We have a strong university system that creates a massive "Brain Drain" whereby many of the best students researchers wind up here. The rules to create a new company are simple enough that overseas entrepreneurs come here to start and do business. In a world with over 4 billion people, excellence almost requires extraordinary dedication, and this immigration reinforces the performance culture. But most important is that people are rewarded in proportion to their productivity, in other words, you're not sacrificing quality of life for work. Without that aspect, the French work week makes perfect sense.
At least for me, the hardest part of being productive is the first hour. After the first hour or so I'm engrossed enough to where I can be productive for several hours. It's an unfortunate aspect of a somewhat creative process like programming. It's much harder to force myself to think about things than force myself to do manual labor, probably because most such work can be done while letting your mind wander.
You keep repeating this story to Slashdot, but I'm missing any credible reference to it. Perhaps you could provide a few more specifics besides the name RMS?
Which is to say, after having worked with and seen plenty more international teams, that other cultures find dedication to excellence a shocking American practice and expectation. I have no idea why this is so, when every example of its absence results in mediocrity.
Taking credit for the whole project was not what the parent was talking about, rather taking responsibility to see a good idea come through in the face of opposition to it. Ideally you don't want just one such person, but two or three. This might make the rest of the team pointless, as it seems like people best able to promote a technology within are also best at using it. At which point I can understand why international teams find this stuff shocking, in the same way that 2.0 "passed with a C" engineers wind up as a "Sales Engineer".
So why doesn't one of the candidates call people out on it? Surely there's room in a ten man race for an outspoken, tell-all-apologize-for-it-later approach that you want.
Also, I think the majority watches CNN on accident. As in "oops, I meant channel 63, not 36" accident. I don't think they spend enough time watching to bother forming opinions on things, as long as the impact is distant and abstract.
Actually, Devil Jesus is the best worst thing I've ever heard of!
True, but I've never seen a prefetch algorithm that can make up for a crappy demand pager.
The answer is simpler than you'd think. The graduating class will not be well rounded. For a while my department had a natural science with lab requirement. A lot of people took geology (apparently the easiest way to satisfy this), I took chemistry, and a lot of the double major engineering types took Physics. Almost nobody took biology. I know I wouldn't have if they hadn't added it on top of the science requirement. Partly this IS the biology field's problem: everyone perceives it as classification/soft science, and teaching focuses on memorization (quick, what are the major kinds of biological molecules? A: Lipids, proteins, and nucleic acids, if I recall). For engineers this is the most boring aspect. Relationships, especially mathematical ones, are most interesting to us. My perception is that if you like science but hate math, biology is the field to go into. If you disagree, go check your college's Biology undergrad math requirements.
The good news is that I was more prepared to understanding what folding@home etc do when they came along. And I'm taking a class now that deals with genomics. The course deals with algorithmic tools for biologists, and while the course is half biology, half computer science, the biologists struggle to keep up with the maths involved in things like BLAST searches (it finds genetic matches for sequences). And the software tools are often horrible.
For my final project we tested a heuristic based SNP prediction tool that we felt was similar to a statistical method another paper presented. The software was horrible. A perl script, written in traditional fuck-you readable perl style. At first I thought something was horribly wrong with our pipeline feeding it data, since it crashed on a divide by zero error. I went about making it continue when a div by zero would have happened, only to discover that the program no longer halted. A div by zero error was the terminating condition! I only shudder to think what else was hiding in that code, but the results were simple to present: it runs fine on its own test data, but the data we had from an insect failed to present any SNPs. There still could be problems with our input to it, but at least I'm sure now the data's not responsible for the div by zero error.
Biologists are generally not best of breed programmers, though I know of a few exceptions. And CS students appear to be avoiding the field--a year after the change, at the next cirriculum review, the Biology requirement was removed because of loud student complaint. And so our students remain "well rounded", by which I mean they all take the same courses that cover much of the same ground they've already been exposed to.
Ridiculus maybe. But I wouldn't mind making teamkilling a felony :P
The thing about AI is that what it is keeps changing. Twenty years ago, Expert Systems was AI. Chess programs used to be considered AI, and some people still do, but generally the pattern seems to be that AI is about unsolved problems. Once we solve them, it's no longer intelligent. Some of this I can understand, simply because as we discover how things work they lose some mystery and intelligence is one of the most mysterious properties we assign to humanity. But very little of the early stuff tried to do learning. As if once we knew how to win at chess, we could take that and apply it to the rest of the universe, treating it like a bigger instance of a chess game.
There's a couple reasons for it not being in the kernel. First, it misleads users who expect some degree of data security. The good news is that sort of person likely follows kernel patches to the FS and would likely be aware of the problem, possibly even writing a script that replaces rm with a real-rm.
.Trash, or configure your Desktop Environment to do so (GNOME does, for example). There's all sorts of things you have to decide that might not suit everyone. For example, if I delete a file on a USB drive, does it go in a .Trash storage in the USB drive, or do we copy it over to a main .Trash folder? Many people don't realize they have to empty the trash to reclaim space on their thumbdrive in GNOME.
.Trash bin in a multiuser system. And quotas. And permissions.
The second argument is that it's better handled in user space, so the OS doesn't have to make that sort of policy. There's no reason you can't just alias rm to some
The final argument I can come up with is security problems. We can't have one global
Reading historic archives of the LKML suggests it's at least come up once. I guess Torvald's opinion is that anything that CAN go in the userspace SHOULD. Can't explain the webserver in kernel though. Perhaps that opinion has changed some time in the last 10 years?
I don't think seniority holds much here. You could disagree with me I suppose, but it wouldn't be very consistent :P
"Why would you trust this process to an indivudal who had no idea what it was like to even go to college?"
Probably because the idea of individuals who do know what its like to go through college marketing MIT to girls wasn't working.
You have to admit the line "this Computer Scientist had never taken a CS class! He was just good at algorithms" makes you sound retarded, since the only constant among the broad regions you've described is precisely algorithms.
Why bother with all this categorization of what people said they did when ohlo goes straight to the point of open source software: the source code.
Because in an better world, you don't have to fight your enemies to the death.
Pray tell, how did you eliminate the possibility of existing components growing, in order to conclude modularity itself is the problem?
All you've done is demonstrate that most all the choices you have on Linux are there on Microsoft too. In addition to the choices that don't exist on the Linux platform. I didn't really mean to harp on the paid aspect quite as much, though I'm sure that someone could highlight a few more commercial tools.
.NET use, though the author suggests .NET developers stay with MS tools. Point is, if you're looking for best of breed, you have to stop caring about vertical integration. You have the choice to use only parts of the vertical integration.
.NET SDK was free. Interesting, but I can't see the justification in using F# over Ocaml. Of course, if you're using .NET you might as well get an MSDN kit and visual studio, so I can see where they're going with it.
What I meant by java is that should you really want to, you can run your entire large scale applcation on java. iPlanet, tomcat, java db, servlets and jsps, netBeans, the Java language, the JVM, etc. Nobody uses all of this, because there's better choices out there. Postgresql, for example works well. I can't really speak to what developers with
On an unrelated note, I wasn't aware the
There's plenty of choice on Windows. The only difference is that these choices involve paying money for things whose worth you can't evaluate until you've used them for longer than a month. Branding helps tremendously in such a situation, as does bundling, both of which MS has in spades.
.NET languages
Some examples of choices developers have on windows platforms:
* IDEs - visual studio, eclipse, netbeans, dev-c++, codewarrior, just to name a few I've used
* The various
* Databases
* Webservers, IIS, apache, or something else?
* antivirus, Vista tried pretty hard to end all of these though.
If you're just moaning about how Microsoft has a large vertically integrated set of tools, well, there's Java. Nobody does this, because its stupid and they have the choice not to.
Given the mix of master's international students that remain in the US, I'd say only marginally better. The good news is they communicate well with their foreign counterparts, and India has some history with English.