Gates can blame himself. People look at the roadkill Microsoft has left behind, companies like Netscape, Go, NeXT, DEC, and even companies like Sun, IBM, and Apple, and they wonder: what's the point? If I do anything interesting, Microsoft is either going to kill me or buy me, and then they are just going to continue shipping their own stuff. Should someone get a CS degree just to become an Microsoft certified virus remover?
IBM is focused on Linux; it's a business thing for them. There is no reason why thy should distract people from that by open sourcing OS/2.
Heck, I have some working software I could open source that I don't want to open source, not because I ever want to make money from it, but because I think the software is just not good for you; I'd much rather you use one of the alternatives.
You just pointed out one of the key advantages of not having the whole desktop in a single address space. That's 1:0 for the current approach...
No, it's not an advantage, it's a workaround for a fundamental problem with C/C++. Almost no other modern programming language needs such a workaround, and the price you pay for it is horrendous, both in terms of efficiency and in terms of functionality.
What's ironic that some of the same people that keep complaining about X11's pretty well-designed client-server architecture see nothing wrong with the dozens of processes and inter process communication in Windows, Macintosh, Gnome, and KDE.
What would be the advantage from a user's perspective? And again, what is wrong with KParts?
Can I embed the KMail message list component in a side panel by dragging it there? Can I embed the media player component into a KWord document by dragging its window there? Can I add a button to the KMail user interface by dragging it there, then edit its function in place? KDE, like most current user interfaces, is extremely restricted in what it lets users and developers do. Look at the work Alan Kay has been doing over the last decade; unlike what KDE is doing, his stuff is actually innovative.
But the question is whether their use of BSD code is actually legal.
Quite apart from whether it's legal, there is the ethical question of whether a company should be able to make billions out of research and development paid for by taxpayers; and just about every technology Microsoft ships was paid for by taxpayers through taxes.
Downloading a bittorrent of Windows XP = "Not Stealing".
Over the years, I have paid for so many copies of Windows XP that I would not consider it "stealing" if I started downloading some copies that I did actually want (it would still be illegal, but legality and ethics don't always coincide). Of course, I still don't want it, so the question is academic.
Violating copyright is viewed as about as serious as jaywalking on slashdot, unless the specific copyright you violate is the GPL, then it's worse than murder.
Well, and have you stopped for a moment to think why that is? I'll tell you: many companies view GPL, BSD, and similar licenses as equivalent to public domain. In fact, many companies treat such software as public domain, which is why they still keep incorporating it into their commercial products without complying with the license. People in the open source community are getting pissed off by arrogant and ignorant commercial software vendors.
traditional Western medicine based on scientific evidence is less and less trusted by the general public.
Most doctors, in my experience (more than I like) seem completely incapable of understanding science, or statistics for that matter. Most doctors also don't seem to give a damn.
In the end, there probably are some serious hypochondriacs out there, but we are all better off being able to get up-to-date medical information on the web and being able to reach people with the same medical conditions as we quickly and easily.
Of course, doctors don't like it: their monopoly on knowledge is eroding.
Why, o why do they always make changing the C++ ABI such an effort?
Because that's the way C++ is designed: it has lots and lots of dependencies on data structure and vtable layouts. Why? Because it may save a few cycles in a few places that don't usually matter. The price you pay is that upgrades are a PITA.
Objective-C is a little better in this regard (but unfortunately has lots of other problems).
Other designers have already been there, and while there's a lot to say about the heavy legacy of various existing designs, they work and have billions of man/hours put into them.
Well, I think people ought to explore new directions. Trouble is: BeOS/Zeta isn't doing that--it is basically the same as mainstream systems, only with more bouncing heads and less modularity.
It's pointless to post lengthy responses to anonymous posts, but in brief...
What's wrong with KParts? What would you like to replace them with?
Integration of the entire desktop environment in a single address space and language support for such integration.
And what would you like to see instead of applications and files?
Applications shouldn't exist at all. Files should exist, but the user shouldn't ordinarily see them.
And what makes you think that you can't do that in C/C++ at all?
Think of what would happen if you took everything the user is running in KDE and put it in a single address space. A single pointer error anywhere would crash the whole desktop. And that's just one of the problems. It's not that it is theoretically impossible, it's that it is impractical.
as for applications not talking to each other.. well... wait a few months until there's something to play with. i think you may be impressed =)
You've been at it for nearly a decade, but you are going to make it all right in the next six months: millions of lines of application code, the language, the toolkit and libraries, etc. OK, well, if you say so...
well, you're wrong on just about every count. while plasma itself is being written in C++, language bindings for javascript+html, python, ruby and java are all committed to (e.g. we have people to do them =). a visual designer is also on the roadmap.
Statements like that show that you don't even understand what the issues are, let alone what to do about them.
As I was saying, compared to Microsoft Windows or OS X, KDE isn't all that bad. But it is not "innovative" in any way.
Alright, since you know what your talking about, I nominate you to go write a window manager in assembly code. God knows you can write better assembler than a compiler.*
Wrong direction. What makes KDE bloated, cumbersome, and inefficient is that it is using a low-level language like C++; going to assembler would make that worse, not better.
Honestly, have you ever written a Qt/KDE app? It's a breeze!
It's "a breeze" only if your standards are already low (MFC, Gtk+, etc.).
OO makes everything make sense, and even though in most cases its overused, for UI its perfect.
Well, wouldn't it be nice then if KDE actually were written in a real object-oriented language instead of C++.
Remember that Alan coined the phrase "Object-Oriented". He has this to say "I invented the term Object-Oriented, and I can tell you I did not have C++ in mind." (here)
Maybe Qt/KDE is a little nicer than Gtk+/Gnome, but from the point of view of the state of the art of OOP, GUIs, and programming environments, that's like worrying about whether flintstones or wooden sticks are a better way to start a fire. You guys are reinventing the stone wheel for the third time. Come on, get with the program and move into the 21st century.
What would constitue "a deep way", in your opinion?
Well, "talking to each other in a deep way" in particular would mean getting rid of "applications" and "files" and crutches like KParts. You can't do it in a C/C++-based environment at all. More generally, innovating in a deep way would mean radically changing the way people interact with the UI.
Note that I am not claiming that such "new" ways of doing things are necessarily better, I'm just saying that as long as you just tinker around with Windows/Mac-like GUIs, you aren't innovating. And as long as nobody seriously tries, we'll never know whether they are better.
(Actually, 20 years ago, people were trying some of those things, but those approaches were killed by the junk that Microsoft and Apple started shipping around then, not because Microsoft or Apple were any better, but because they were cheaper.)
C++ does have it's faults, but OO is a good fit for user interface coding
OO is an excellent fit for user interface coding; too bad that C++ implements it so poorly.
and C++ is certainly more efficient than any other OO language.
It most certainly is not. The problem is not that C++ gives you low-level access, the problem is that it forces you to worry about low-level concerns even when you should be concentrating on algorithms and design.
I see mostly more tinkering around the edges: a few gimmicks and UI features, but nothing substantial. It's still the same C++-based bloat, the same applications that don't really talk to each other in a deep way, still the same brittle system, still the same barrier between users and programmers.
When I'm done adding stuff to my cart, it takes about 30 seconds to be done with checkout, and I don't have to think about my order again until it arrives at my door.
That's how almost all on-line bookstores work.
In any case, if you value your time and want good service, there is an even better choice: give (E-mail, FAX, whatever) your list of books to an independent bookseller near you--they'll do all the searching and ordering for you.
I love it when people that have never used either system, talk about it as if they were experts.
I guess that about sums it up: you think that by using a system, you can determine innovation. You think that a bit of theming and some incremental engineering improvements equal innovation.
Name a single significant innovation that Apple has contributed to the world of desktop computing over the last decade.
Why has there been such a sudden lack in innovation as of late?
There hasn't been much innovation in the last 10-20 years. Windows NT is, at best, a rewrite of 1980's technology. Macintosh is actually shipping nearly the same system that NeXT came up with in the mid-1980's. Linux is a reimplementation of 1980's UNIX, and BSD is largely based on 1980's UNIX (with some tweaks). The Internet has been around since the 1960's, and people were using it for pretty much the same things they are using it for today 20 years ago (discussion groups, pr0n, news, even some on-line shopping).
What has happened is that existing technologies have become popularized. That's not "innovation", it's business.
5+ standard deviations against the control is interesting. Should be easy to reproduce. (or not).
That may be interesting for IQ, but it's not very interesting in this case: if this really works, you should be able to get many orders of magnitude above background. Having to talk about "5+ standard deviations" is itself an indication that the effect is miniscule and may simply be due to experimental error.
You can get tabletop fusion with a TV high voltage supply, a glass bulb, some wire, and deuteriums gas. That's been known for decades and is used as a neutron source commercially. People build those things for science fairs. It's called the "Farnsworth Fusor" (I know, in light of Futurama it sounds like a joke, but the fictional character was named after the real one).
Why don't we all have flying cars, then? Because you can't get a net energy gain with the Farnsworth Fusor--it seems to be impossible in general to do so, the numbers just don't work out.
Of course, even if you do make it efficient, it's not exactly "clean energy": even with so-called aneutronic fusion, a few percent of the fusion reactions will generate neutrons, which, for realistic power generation, results in a neutron flux that causes the power generation to be quite dirty. Not as dirty as fission--disposal should be easier--but don't expect something harmless you can just run in your basement.
So, tabletop fusion isn't really anything impressive: there are probably lots of ways of getting fusion on your tabletop. The question is how you make it efficient enough to useful amounts of energy out of it. And cavitation seems no more promising there than inertial confinement in the Farnsworth Fusor. But maybe if enough people keep playing around with this, someone will get lucky and find something that works.
So many things point to Amazon that it seems like a lot of hassle to buy anywhere else, but it's not.
Get yourself a copy of Book Burro; it will automatically annotate any Amazon page you go to with a list of other bookstores you can buy the book at, as well as the prices (often lower than Amazon).
I think there are too many competing technologies: QT vs. GTK, dpkg vs. rpm vs. ebuild vs. tgz, etc. If we could work out some good standards -- that everybody followed
And how are you going to make everybody follow that one standard?
Fact is that the variety of libraries and environments out there reflects people's preferences.
But if you want to standardize, why don't we start with you. Throw out your desktop environment and switch to twm and Lesstif. Hey, what are you complaining about? It's been the standard for decades.
The Xbox 360 is, for practical purposes, a completely new product.
As for what Microsoft wants to do, it doesn't really matter. The company has had far more failures (some of them devastating) than successes.
Even their "successful" products (meaning, non-negligible market share) generally only survive because their Windows monopoly pulls them through. Or do you seriously believe, any significant number of people would pay money for IIS or Access or SQL Server or Windows Media servers/clients if it weren't for their integration into Windows?
Slashdot Mirror
Gates can blame himself. People look at the roadkill Microsoft has left behind, companies like Netscape, Go, NeXT, DEC, and even companies like Sun, IBM, and Apple, and they wonder: what's the point? If I do anything interesting, Microsoft is either going to kill me or buy me, and then they are just going to continue shipping their own stuff. Should someone get a CS degree just to become an Microsoft certified virus remover?
I assume that when you state that the tax payers paid for this software that you are talking about the first BSD IP-stack being written in Berkley
No, I'm talking about most of Microsoft's technologies: window systems, graphics, networking, operating systems, etc.
To address your last statement, I highly doubt most companies consider GPL to be public domain, since most of them avoid it like the plauge
Except for the ones that ship GPL-derived binaries without telling anybody about it. There are hundreds if not thousands of those around.
Well, the people (loosly) control where the government spends it's money, and if they didn't want that
What planet are you from? The American people have no idea where their money comes from or goes.
The only real difference between BSD and public domain is that you must include a Copyright notice.
You go on believing that.
IBM is focused on Linux; it's a business thing for them. There is no reason why thy should distract people from that by open sourcing OS/2.
Heck, I have some working software I could open source that I don't want to open source, not because I ever want to make money from it, but because I think the software is just not good for you; I'd much rather you use one of the alternatives.
You just pointed out one of the key advantages of not having the whole desktop in a single address space. That's 1:0 for the current approach...
No, it's not an advantage, it's a workaround for a fundamental problem with C/C++. Almost no other modern programming language needs such a workaround, and the price you pay for it is horrendous, both in terms of efficiency and in terms of functionality.
What's ironic that some of the same people that keep complaining about X11's pretty well-designed client-server architecture see nothing wrong with the dozens of processes and inter process communication in Windows, Macintosh, Gnome, and KDE.
What would be the advantage from a user's perspective? And again, what is wrong with KParts?
Can I embed the KMail message list component in a side panel by dragging it there? Can I embed the media player component into a KWord document by dragging its window there? Can I add a button to the KMail user interface by dragging it there, then edit its function in place? KDE, like most current user interfaces, is extremely restricted in what it lets users and developers do. Look at the work Alan Kay has been doing over the last decade; unlike what KDE is doing, his stuff is actually innovative.
MS legally using BSD licensed code = "Stealing".
But the question is whether their use of BSD code is actually legal.
Quite apart from whether it's legal, there is the ethical question of whether a company should be able to make billions out of research and development paid for by taxpayers; and just about every technology Microsoft ships was paid for by taxpayers through taxes.
Downloading a bittorrent of Windows XP = "Not Stealing".
Over the years, I have paid for so many copies of Windows XP that I would not consider it "stealing" if I started downloading some copies that I did actually want (it would still be illegal, but legality and ethics don't always coincide). Of course, I still don't want it, so the question is academic.
Violating copyright is viewed as about as serious as jaywalking on slashdot, unless the specific copyright you violate is the GPL, then it's worse than murder.
Well, and have you stopped for a moment to think why that is? I'll tell you: many companies view GPL, BSD, and similar licenses as equivalent to public domain. In fact, many companies treat such software as public domain, which is why they still keep incorporating it into their commercial products without complying with the license. People in the open source community are getting pissed off by arrogant and ignorant commercial software vendors.
traditional Western medicine based on scientific evidence is less and less trusted by the general public.
Most doctors, in my experience (more than I like) seem completely incapable of understanding science, or statistics for that matter. Most doctors also don't seem to give a damn.
In the end, there probably are some serious hypochondriacs out there, but we are all better off being able to get up-to-date medical information on the web and being able to reach people with the same medical conditions as we quickly and easily.
Of course, doctors don't like it: their monopoly on knowledge is eroding.
Why, o why do they always make changing the C++ ABI such an effort?
Because that's the way C++ is designed: it has lots and lots of dependencies on data structure and vtable layouts. Why? Because it may save a few cycles in a few places that don't usually matter. The price you pay is that upgrades are a PITA.
Objective-C is a little better in this regard (but unfortunately has lots of other problems).
it's designed to be low latency rather than to have a network-aware window system and multi-user capabilities
First of all, the Linux GUI uses shared memory and IPC, just like almost every other desktop OS in common use.
Secondly, none of those things are mutually exclusive. For example, ultiuser capabilities have no impact on the Linux kernel or its performance.
Be stayed single user and concentrated on its multimedia specialisation.
Yes, and that's a needless specialization. All major desktop OSes are more than fast enough to handle multimedia.
Other designers have already been there, and while there's a lot to say about the heavy legacy of various existing designs, they work and have billions of man/hours put into them.
Well, I think people ought to explore new directions. Trouble is: BeOS/Zeta isn't doing that--it is basically the same as mainstream systems, only with more bouncing heads and less modularity.
The forward thinking population of /. will now mock it because:
It will mock it because it has the same problems as Linux, BSD, OS X, and Windows, and on top of that isn't even backwards compatible.
Diversity is good, and a billion people writing GNU-style apps for Linux is not diversity.
Diversity is good. Too bad that BeOS and its derivatives don't provide it.
It's pointless to post lengthy responses to anonymous posts, but in brief...
What's wrong with KParts? What would you like to replace them with?
Integration of the entire desktop environment in a single address space and language support for such integration.
And what would you like to see instead of applications and files?
Applications shouldn't exist at all. Files should exist, but the user shouldn't ordinarily see them.
And what makes you think that you can't do that in C/C++ at all?
Think of what would happen if you took everything the user is running in KDE and put it in a single address space. A single pointer error anywhere would crash the whole desktop. And that's just one of the problems. It's not that it is theoretically impossible, it's that it is impractical.
as for applications not talking to each other .. well ... wait a few months until there's something to play with. i think you may be impressed =)
You've been at it for nearly a decade, but you are going to make it all right in the next six months: millions of lines of application code, the language, the toolkit and libraries, etc. OK, well, if you say so...
well, you're wrong on just about every count. while plasma itself is being written in C++, language bindings for javascript+html, python, ruby and java are all committed to (e.g. we have people to do them =). a visual designer is also on the roadmap.
Statements like that show that you don't even understand what the issues are, let alone what to do about them.
As I was saying, compared to Microsoft Windows or OS X, KDE isn't all that bad. But it is not "innovative" in any way.
Wrong direction. What makes KDE bloated, cumbersome, and inefficient is that it is using a low-level language like C++; going to assembler would make that worse, not better.
Honestly, have you ever written a Qt/KDE app? It's a breeze!
It's "a breeze" only if your standards are already low (MFC, Gtk+, etc.).
OO makes everything make sense, and even though in most cases its overused, for UI its perfect.
Well, wouldn't it be nice then if KDE actually were written in a real object-oriented language instead of C++.
Maybe Qt/KDE is a little nicer than Gtk+/Gnome, but from the point of view of the state of the art of OOP, GUIs, and programming environments, that's like worrying about whether flintstones or wooden sticks are a better way to start a fire. You guys are reinventing the stone wheel for the third time. Come on, get with the program and move into the 21st century.
What would constitue "a deep way", in your opinion?
Well, "talking to each other in a deep way" in particular would mean getting rid of "applications" and "files" and crutches like KParts. You can't do it in a C/C++-based environment at all. More generally, innovating in a deep way would mean radically changing the way people interact with the UI.
Note that I am not claiming that such "new" ways of doing things are necessarily better, I'm just saying that as long as you just tinker around with Windows/Mac-like GUIs, you aren't innovating. And as long as nobody seriously tries, we'll never know whether they are better.
(Actually, 20 years ago, people were trying some of those things, but those approaches were killed by the junk that Microsoft and Apple started shipping around then, not because Microsoft or Apple were any better, but because they were cheaper.)
C++ does have it's faults, but OO is a good fit for user interface coding
OO is an excellent fit for user interface coding; too bad that C++ implements it so poorly.
and C++ is certainly more efficient than any other OO language.
It most certainly is not. The problem is not that C++ gives you low-level access, the problem is that it forces you to worry about low-level concerns even when you should be concentrating on algorithms and design.
I see mostly more tinkering around the edges: a few gimmicks and UI features, but nothing substantial. It's still the same C++-based bloat, the same applications that don't really talk to each other in a deep way, still the same brittle system, still the same barrier between users and programmers.
When I'm done adding stuff to my cart, it takes about 30 seconds to be done with checkout, and I don't have to think about my order again until it arrives at my door.
That's how almost all on-line bookstores work.
In any case, if you value your time and want good service, there is an even better choice: give (E-mail, FAX, whatever) your list of books to an independent bookseller near you--they'll do all the searching and ordering for you.
I love it when people that have never used either system, talk about it as if they were experts.
I guess that about sums it up: you think that by using a system, you can determine innovation. You think that a bit of theming and some incremental engineering improvements equal innovation.
Name a single significant innovation that Apple has contributed to the world of desktop computing over the last decade.
Why has there been such a sudden lack in innovation as of late?
There hasn't been much innovation in the last 10-20 years. Windows NT is, at best, a rewrite of 1980's technology. Macintosh is actually shipping nearly the same system that NeXT came up with in the mid-1980's. Linux is a reimplementation of 1980's UNIX, and BSD is largely based on 1980's UNIX (with some tweaks). The Internet has been around since the 1960's, and people were using it for pretty much the same things they are using it for today 20 years ago (discussion groups, pr0n, news, even some on-line shopping).
What has happened is that existing technologies have become popularized. That's not "innovation", it's business.
If this is really D+D fusion, what purpose could neutron seeding possibly serve?
5+ standard deviations against the control is interesting. Should be easy to reproduce. (or not).
That may be interesting for IQ, but it's not very interesting in this case: if this really works, you should be able to get many orders of magnitude above background. Having to talk about "5+ standard deviations" is itself an indication that the effect is miniscule and may simply be due to experimental error.
You can get tabletop fusion with a TV high voltage supply, a glass bulb, some wire, and deuteriums gas. That's been known for decades and is used as a neutron source commercially. People build those things for science fairs. It's called the "Farnsworth Fusor" (I know, in light of Futurama it sounds like a joke, but the fictional character was named after the real one).
Why don't we all have flying cars, then? Because you can't get a net energy gain with the Farnsworth Fusor--it seems to be impossible in general to do so, the numbers just don't work out.
Of course, even if you do make it efficient, it's not exactly "clean energy": even with so-called aneutronic fusion, a few percent of the fusion reactions will generate neutrons, which, for realistic power generation, results in a neutron flux that causes the power generation to be quite dirty. Not as dirty as fission--disposal should be easier--but don't expect something harmless you can just run in your basement.
So, tabletop fusion isn't really anything impressive: there are probably lots of ways of getting fusion on your tabletop. The question is how you make it efficient enough to useful amounts of energy out of it. And cavitation seems no more promising there than inertial confinement in the Farnsworth Fusor. But maybe if enough people keep playing around with this, someone will get lucky and find something that works.
So many things point to Amazon that it seems like a lot of hassle to buy anywhere else, but it's not.
Get yourself a copy of Book Burro; it will automatically annotate any Amazon page you go to with a list of other bookstores you can buy the book at, as well as the prices (often lower than Amazon).
I think there are too many competing technologies: QT vs. GTK, dpkg vs. rpm vs. ebuild vs. tgz, etc. If we could work out some good standards -- that everybody followed
And how are you going to make everybody follow that one standard?
Fact is that the variety of libraries and environments out there reflects people's preferences.
But if you want to standardize, why don't we start with you. Throw out your desktop environment and switch to twm and Lesstif. Hey, what are you complaining about? It's been the standard for decades.
The Xbox 360 is, for practical purposes, a completely new product.
As for what Microsoft wants to do, it doesn't really matter. The company has had far more failures (some of them devastating) than successes.
Even their "successful" products (meaning, non-negligible market share) generally only survive because their Windows monopoly pulls them through. Or do you seriously believe, any significant number of people would pay money for IIS or Access or SQL Server or Windows Media servers/clients if it weren't for their integration into Windows?