You are feeding the arguments for what I have been saying for years: Not only have the MS monopoly keeped the OS development back. It has also stifled the chip-development. All developers developed for MS-Intel. Nothing else. Most propetary software isn't portable. Therefore there was no applications for NT4 on Alpha. Therefore MS have given up on WindowsNT on PPC even though the OS might be easily portable in itself. Therefore all the superiour architectures to i386 has died.
If we are going to get new architectures - if it is not already too late - on the desktop we have to take either the route of open source, where each user (in principle) can recompile the application for his architecture, or the route of virtual machines (Java or.Net), which can do hotspot compilation locally. The distribution system of current properetary, closed source software, where precompiled binaries are distributed, kills every attempt to make an architecture which isn't compatible with the dominat i386 (maybe x64 in the future).
There is a good thing happening though: Intel and AMD seems to got stuck wrt. clock-speed. They can't make the CPUs run any faster now. They have to go for hypethreading or multi-core chips. For that to give any performance benifit most applications have to be rewritten. If while doing that people start to think about portability ther might be a chance that those rewritten applications will also run on other architectures. Even PC programmers aren't living in the near-assambler programming world as they did in the 80's and beginning of the 90s anymore!
.. was scientific research something you voted true/false?
What does it count if someone states something in an article and the only argument is references to other articles! I bet that is what is done in the 90% of the 75% of the articles. You can keep pepeating the same "fact" over and over again and everybody starts believing it!
I work as a system level programmer (device drivers, protocol stacks etc.) I have Ph.D. in physics. I was original hired because I am a geek with experience with Linux and especially gcc. I thought I was good and could compete with people with master degrees in computer science.
I have since started to take classes in CS in my spare time. I found out that I had so much to learn!
In (some of) those classes I really learn a lot which I would never have learned "by experience".
The conclusion: If I was the boss, I would hire people with a degree. And ofcourse one you can trust have substance. At a good university you learn stuff in a few years that would have taken lifetimes to learn "by experience".
You mix math with other sciences, like physics. Physics is indeed like what you descripe. Math isn't. Math is about starting from some simple axioms and prove all the rest with logic, not observations.
Do you have code in Sun's version of OO? Is it GPL and do you still have to copyright?
If so I think you can sue Sun for copyright violation if they put patented code into OO, as that would be the same as putting further license restrictions on OO which GPL explicitly forbid.
Wind is actually considered a good starter. If the wind is blowing a wind turbine can go from being idling to produce full power within 30 secs.
Here in Denmark the power companies have started to see wind turbines placed in big windfarms as stabilising the grid, whereas they the old individual turbines destabelished the grid.
It is not that simple! Even if you want to give girls legos they prefer the dolls. If you try to give boys the dolls they prefer the legos anyway.
To a large extent the effect is biological.
What happens is that the testetorone in the fetus is creating the brain slightly differently.
What is sad about all this is that people look at the statistics and argue we don't have equal oppotunities and want to "force" women into forinstance computer science. By doing that they remove the equal oppotunies!
What must matter is that whatever you want to do you should be allowed to as long as you are qualified. If 10 times as many boys than girls want to study computerscience then that is how it is. By trying to make meassures to make it 50-50% you will probably end up destroying computer science for everybody.
Isn't really very much... The company I work for allready sell windtubines at 3MW. Other companies sell even bigger ones (4.5MW I believe.)
These turbines takes a lot of manpower to keep running. Stuff needs to be repaired every month or so. I can't start to imagine the problems one would have when trying to put them down into the salty waters of East River!
But then again: One have to try and get the technology running. That was how the windturbine-buisness got started, too, and that is big buisness these days.
Now IE is in it's standeard configuration a lot more insecure than Mozilla. But when it comes to netbanking the security problems are not as such in the browser but in the OS you are running it on: What is the probability of gettting a trojan that sniffs your password and other security keys to letting the cracker empty you bankaccount? If they control your PC they can even do it from your very own PC!
Now it is really bad when they rely on you turning on ActiveX or something else insecure making your PC even more vulnerable to trojans!
I run Linux at home, but I still don't dare use netbanking (also because I have had insights into the system my bank uses from my professional life).
I considered getting an account in another bank where they don't rely so much on your PC to be secure: Once in a while they snail mail you a small physical card with a table of random numbers on it. When you want to do a transfer of money it asks you to look up into the table and type in the corresponding number. This way they can make sure you not only know the password but also have the physical card. Thus if a cracker takes over your PC they can't transfer money from your account anyway - only see what you have on your account. This solution is ofcourse not very elegant but it is much more secure than what any of the other banks can offer.
Where does it state in the GPL that you can't restrict downloads? Even charge for downloads?
(MontaVista is doing the same thing I believe).
What you can't do, however, is to put any other licence but the GPL on the downloaded material. I.e. anyone who downloads those rpms are free to destribute it. Where is that license? There is nothing about it on their website.
On the other hand, no matter what kind of passwords they put on their service and they don't put on additional licences they are still distributing the code under GPL - i.e. giving you the rights to use whatever code they might actually have copyright to under the GPL.
And the IDE WindRiver offer on top of gcc/make/gdb, called Tornado, sucks (slow full of bugs etc.) They have started to ship a new IDE called Sniff but when I tested it was a mess. Might have improved now - it got some potential!
For people who can live without a fancy IDE and can settle with the command line (or use emacs as a wrapper) Linux is a very good option. For people used to Visual Studie, Tornado is closer to what they are used to, but it is still far from satisfying the Windows developer.
But with Eclipse around it should be hard for MontaVista or others to create an IDE much better than Tornado.
At work we use VxWorks but I ported a Linux to some of our processor boards. Quite frankly it wasn't so hard: Linux had drivers for most of our hardware. VxWorks didn't, so we had to buy them or spend a lot of man power developing it.
But VxWorks got the realtime performance - don't forget that!I tried to do a small meassurement of the (worst case) latency for tasks/process scheduling. VxWorks: ~100us, Linux (with preemtive patch): ~2ms. Since then I have improved VxWorks latency a lot by removing a big interrupt lock out in one of our drivers.
Slashdot Mirror
You are feeding the arguments for what I have been saying for years: Not only have the MS monopoly keeped the OS development back. It has also stifled the chip-development. All developers developed for MS-Intel. Nothing else. Most propetary software isn't portable. Therefore there was no applications for NT4 on Alpha. Therefore MS have given up on WindowsNT on PPC even though the OS might be easily portable in itself. Therefore all the superiour architectures to i386 has died.
.Net), which can do hotspot compilation locally. The distribution system of current properetary, closed source software, where precompiled binaries are distributed, kills every attempt to make an architecture which isn't compatible with the dominat i386 (maybe x64 in the future).
If we are going to get new architectures - if it is not already too late - on the desktop we have to take either the route of open source, where each user (in principle) can recompile the application for his architecture, or the route of virtual machines (Java or
There is a good thing happening though: Intel and AMD seems to got stuck wrt. clock-speed. They can't make the CPUs run any faster now. They have to go for hypethreading or multi-core chips. For that to give any performance benifit most applications have to be rewritten. If while doing that people start to think about portability ther might be a chance that those rewritten applications will also run on other architectures. Even PC programmers aren't living in the near-assambler programming world as they did in the 80's and beginning of the 90s anymore!
What does it count if someone states something in an article and the only argument is references to other articles! I bet that is what is done in the 90% of the 75% of the articles. You can keep pepeating the same "fact" over and over again and everybody starts believing it!
I have since started to take classes in CS in my spare time. I found out that I had so much to learn! In (some of) those classes I really learn a lot which I would never have learned "by experience".
The conclusion: If I was the boss, I would hire people with a degree. And ofcourse one you can trust have substance. At a good university you learn stuff in a few years that would have taken lifetimes to learn "by experience".
You mix math with other sciences, like physics. Physics is indeed like what you descripe. Math isn't. Math is about starting from some simple axioms and prove all the rest with logic, not observations.
If so I think you can sue Sun for copyright violation if they put patented code into OO, as that would be the same as putting further license restrictions on OO which GPL explicitly forbid.
Here in Denmark the power companies have started to see wind turbines placed in big windfarms as stabilising the grid, whereas they the old individual turbines destabelished the grid.
To a large extent the effect is biological. What happens is that the testetorone in the fetus is creating the brain slightly differently.
What is sad about all this is that people look at the statistics and argue we don't have equal oppotunities and want to "force" women into forinstance computer science. By doing that they remove the equal oppotunies!
What must matter is that whatever you want to do you should be allowed to as long as you are qualified. If 10 times as many boys than girls want to study computerscience then that is how it is. By trying to make meassures to make it 50-50% you will probably end up destroying computer science for everybody.
Isn't really very much... The company I work for allready sell windtubines at 3MW. Other companies
sell even bigger ones (4.5MW I believe.)
These turbines takes a lot of manpower to keep running. Stuff needs to be repaired every month or so. I can't start to imagine the problems one would have when trying to put them down into the salty waters of East River!
But then again: One have to try and get the technology running. That was how the windturbine-buisness got started, too, and that is big buisness these days.
Try to use JFFS2 - it is designed for flash and is journalled! And it can't be worse than WindRivers current TFFS/DosFS combo!
Now it is really bad when they rely on you turning on ActiveX or something else insecure making your PC even more vulnerable to trojans!
I run Linux at home, but I still don't dare use netbanking (also because I have had insights into the system my bank uses from my professional life).
I considered getting an account in another bank where they don't rely so much on your PC to be secure: Once in a while they snail mail you a small physical card with a table of random numbers on it. When you want to do a transfer of money it asks you to look up into the table and type in the corresponding number. This way they can make sure you not only know the password but also have the physical card. Thus if a cracker takes over your PC they can't transfer money from your account anyway - only see what you have on your account. This solution is ofcourse not very elegant but it is much more secure than what any of the other banks can offer.
What you can't do, however, is to put any other licence but the GPL on the downloaded material. I.e. anyone who downloads those rpms are free to destribute it. Where is that license? There is nothing about it on their website.
On the other hand, no matter what kind of passwords they put on their service and they don't put on additional licences they are still distributing the code under GPL - i.e. giving you the rights to use whatever code they might actually have copyright to under the GPL.
And the IDE WindRiver offer on top of gcc/make/gdb, called Tornado, sucks (slow full of bugs etc.) They have started to ship a new IDE called Sniff but when I tested it was a mess. Might have improved now - it got some potential!
For people who can live without a fancy IDE and can settle with the command line (or use emacs as a wrapper) Linux is a very good option. For people used to Visual Studie, Tornado is closer to what they are used to, but it is still far from satisfying the Windows developer.
But with Eclipse around it should be hard for MontaVista or others to create an IDE much better than Tornado.
At work we use VxWorks but I ported a Linux to some of our processor boards. Quite frankly it wasn't so hard: Linux had drivers for most of our hardware. VxWorks didn't, so we had to buy them or spend a lot of man power developing it.
But VxWorks got the realtime performance - don't forget that!I tried to do a small meassurement of the (worst case) latency for tasks/process scheduling. VxWorks: ~100us, Linux (with preemtive patch): ~2ms. Since then I have improved VxWorks latency a lot by removing a big interrupt lock out in one of our drivers.
I wonder how TRON (The Real-time Operation Nucleus)
is related to the real time operating system you offered at
http://www.atinucleus.com/
Is the word "Nucleus" related anyway? There are a few faint references to TRON on the Nucleus-site