Open Source Community gets despised by RMS (and fanatics like him) for not urging the "political issues", and in the same breath they want to put their name on and take credit for something made by the Open Source Community. Further, calling it GNU/Linux will give you the impression that a Linux distribution is made out of GNU software and the Linux kernel. This is far from true. A normal Linux distribution contains the Linux kernel, plus software from a lot of different projects like, BSD, Apache, Perl, etc... and GNU. And finaly, Linux is NOT a "GNU variant". Linux was not put into the GNU operating system. Instead, some GNU software (and just as important software ogirinating from other projects) was collected and put around the Linux kernel to form a complete operation system.
Binaries are important, but not even close to the importance of screenshots. Without screenshots I don't bother downloading (unless I've seen it before, etc). Putting up screenshots must be one of the simplest ways to attract users. Even if the app is text-based I want to see an xterm screenshot (unless it's a non-issue, like a daemon app, etc).
I tried the first alpha release of opera (released months ago) and was not very impressed. However, this release was really nice and fast. Of course it crashed a couple of times, and some thing still didn't work, but it looks very promesing. I'm really looking forward to the final 4.0.
If they would have done this two years ago they would have had ~90% of the Linux GUI market. However, they didn't, and now QT is just another player in the GUI-field. Bad business move.
As soona as a few hundred sun-developers are working on your favourite pet-project[...]
A few hundred developers? You obviously don't have a clue about the size of Sun. They don't have the resources to put "a few hundred" developers on the Gnome project. In total (world wide), they have about 4000 people working with software (SunOS, compilers, cluster software, libraries, CDE, etc, etc.), the rest (and a much bigger part) is working with hardware, sales, etc. So thinking that Sun will put more than 20-30 people (tops) on the Gnome project is probably not very familiar with the industry.
My guess on what the future will look like is something like this:
Mozilla, as a whole, will never be very populare because it's bloated, slow, and gigantic.
For Unix/Linux users the Galeon browser will become THE browser, becuase it's fast, small, and ONLY a browser.
Since Galeon is based on gecko (the mozilla rendering engine), Gecko will continue to evolve and become the only major part of the original Mozilla project that "survive".
...and then what? open the.doc file in emacs? I don't think he wanted a HTML or PDF because he didn't know how to open the zip/exe, but because he couldn't view the.doc file.
Most garbage collectors work by sweeping over the active areas of the heap, and copying anything which is reachable.
The algorithm you are refering to is called Copying (or Copying Garbage Collection). It is NOT the most common algorithm for collecting garbage. Why? Well, simply because of its disadvantages. First it requires a heap twice the size of what the progam needs (often refered to as From-space and To-space). As the amount of memory used by a program increases, less memory is deallocated each time the garbage collector is run and consequently the garbage collections become more frequent. Further, this algorithm can be very inefficient if the number of resident objects is large. If the amount of memory occupied by active objects is large, a lot of copying has to be performed in order to transfer all active objects from the From-space to the To-space.
The Mark Sweep and Reference Counting algorithms are probably the most common ones today (but not Copying).
Reference counting is NOT the same as garbage collection.
WRONG! Reference counting is indeed one way of collecting garbage. The fact that it can not handle circular data structures doesn't mean that it's no longer a garbage collector, it's just one of its disadventages. Other garbage collection algorithms have disadventages as well, e.g. Mark Sweep sufferes the start-stop behavior, Copying requires a heap twice the size of what the program needs, etc. No algorithm is more of a garbage collector than the other.
This is the first time I ever heard about this site. Perhaps I've visited it some time, but since I didn't bookmark it (and didn't even remember the name) that site couldn't have been that valuable. There are already too many linux sites out there, loosing some of them will probably only do good in the long run.
Since more and more applications become available on Linux the use of Wine gets smaller and smaller. There are already office software, browsers, etc. Personally I don't miss any vital application that doesn't under Linux already.
Did this Poag ever do anything that even slightly resembles a useful contribution to the open source community. To me it was just another lame I-can't-code-but-look-at-my-screenshots site.
If they were to do this, it would probably increase performance. But marginally. At least in linux programs, when I've done profiles, most of the cycles were taken up in console i/o. Printf, etc.
malloc/free is a major bottleneck in many systems, especially in SMPs. So, yes a faster (without locks) malloc/free would speed up a program several times (not just marginally). I have several examples of programs that spends up to 80% of their time in malloc/free. But of course this depends on the kind of program. Multithreaded object-oriented programs running on SMP hardware is defenetly one category of programs that would benefit from this.
It would also require a modification to the kernel - so that the brk() system call would simply call the hardware variants.
Most likely, but is that really a problem?
I think that it probably *would* remove flexibility, but at the same time I think that's a good thing. Part of the problem with memory leaks and stuff is the flexibility...
Yes, it would probably remove some flexibility (which is not a good thing). But I don't see what you mean when you say that flexibility is part of the memory leak problems.
After reading the one page description I have some comments:
They are going to insert malloc/free into hardware, that doesn't mean they will build garbage collection into the hardware (which is a completely different thing). However, as the project description says, it could improve garbage collection performance, but that is another issue.
DMM (Dynamic Memory Management) is a well known term, not something new as the slashdot article tries to make it sound like.
Building DMM into hardware it not a new idea. However, I'm not really sure if it has been successfully implemented in any "real" CPUs.
Please, stop pretending you don't know.... it is well known that the Mozilla interface SUCKS in terms of responsiveness/redraw time.
The slow GUI is the main reason I don't use it. I really hope removeal of the debug code will make it substantially faster, or else I'm afraid this browser is going nowhere.
BeroLinux, for those who don't know, was the first 2.2kernel distribution with everything recompiled to be pentium optimized.
Sorry Dan, but this is not true. Even if it doesn't really matter who did what first I can not resist to point out that Stampede Linux was the first 2.2 kernel distribution that was compiled with pentium optimization.
Slashdot Mirror
Open Source Community gets despised by RMS (and fanatics like him) for not urging the "political issues", and in the same breath they want to put their name on and take credit for something made by the Open Source Community. Further, calling it GNU/Linux will give you the impression that a Linux distribution is made out of GNU software and the Linux kernel. This is far from true. A normal Linux distribution contains the Linux kernel, plus software from a lot of different projects like, BSD, Apache, Perl, etc... and GNU. And finaly, Linux is NOT a "GNU variant". Linux was not put into the GNU operating system. Instead, some GNU software (and just as important software ogirinating from other projects) was collected and put around the Linux kernel to form a complete operation system.
No, because a Beowulf cluster (by definition) uses an Open Source OS, and Solaris is not Open Source.
Binaries are important, but not even close to the importance of screenshots. Without screenshots I don't bother downloading (unless I've seen it before, etc). Putting up screenshots must be one of the simplest ways to attract users. Even if the app is text-based I want to see an xterm screenshot (unless it's a non-issue, like a daemon app, etc).
I tried the first alpha release of opera (released months ago) and was not very impressed. However, this release was really nice and fast. Of course it crashed a couple of times, and some thing still didn't work, but it looks very promesing. I'm really looking forward to the final 4.0.
If they would have done this two years ago they would have had ~90% of the Linux GUI market. However, they didn't, and now QT is just another player in the GUI-field. Bad business move.
As soona as a few hundred sun-developers are working on your favourite pet-project[...]
A few hundred developers? You obviously don't have a clue about the size of Sun. They don't have the resources to put "a few hundred" developers on the Gnome project. In total (world wide), they have about 4000 people working with software (SunOS, compilers, cluster software, libraries, CDE, etc, etc.), the rest (and a much bigger part) is working with hardware, sales, etc. So thinking that Sun will put more than 20-30 people (tops) on the Gnome project is probably not very familiar with the industry.
Mozilla, as a whole, will never be very populare because it's bloated, slow, and gigantic.
For Unix/Linux users the Galeon browser will become THE browser, becuase it's fast, small, and ONLY a browser.
Since Galeon is based on gecko (the mozilla rendering engine), Gecko will continue to evolve and become the only major part of the original Mozilla project that "survive".
...and then what? open the .doc file in emacs? I don't think he wanted a HTML or PDF because he didn't know how to open the zip/exe, but because he couldn't view the .doc file.
Most garbage collectors work by sweeping over the active areas of the heap, and copying anything which is reachable.
The algorithm you are refering to is called Copying (or Copying Garbage Collection). It is NOT the most common algorithm for collecting garbage. Why? Well, simply because of its disadvantages. First it requires a heap twice the size of what the progam needs (often refered to as From-space and To-space). As the amount of memory used by a program increases, less memory is deallocated each time the garbage collector is run and consequently the garbage collections become more frequent. Further, this algorithm can be very inefficient if the number of resident objects is
large. If the amount of memory occupied by active objects is large, a lot of copying
has to be performed in order to transfer all active objects from the From-space to the To-space.
The Mark Sweep and Reference Counting algorithms are probably the most common ones today (but not Copying).
Reference counting is NOT the same as garbage collection.
WRONG! Reference counting is indeed one way of collecting garbage. The fact that it can not handle circular data structures doesn't mean that it's no longer a garbage collector, it's just one of its disadventages. Other garbage collection algorithms have disadventages as well, e.g. Mark Sweep sufferes the start-stop behavior, Copying requires a heap twice the size of what the program needs, etc. No algorithm is more of a garbage collector than the other.
Read a book on the subject and you'll get the idea, I can recommend Garbage Collection - Algorithms for Automatic Dynamic Memory Management.
How about actually reading the article before posting a question about its contents? The answer can be found on the first page... doh!
You don't have to have WINE installed to run this program! WINE is NOT neede by xmps!
This is the first time I ever heard about this site. Perhaps I've visited it some time, but since I didn't bookmark it (and didn't even remember the name) that site couldn't have been that valuable. There are already too many linux sites out there, loosing some of them will probably only do good in the long run.
Is Wine really needed?
Since more and more applications become available on Linux the use of Wine gets smaller and smaller. There are already office software, browsers, etc. Personally I don't miss any vital application that doesn't under Linux already.
[Pro64] is fully compatible with the Red Hat Cygnus gcc and g++ offerings
what does this mean? is it a gcc back-end or is it a complete compiler of its own?
Did this Poag ever do anything that even slightly resembles a useful contribution to the open source community. To me it was just another lame I-can't-code-but-look-at-my-screenshots site.
oh, well... there goes my karma
If they were to do this, it would probably increase performance. But marginally. At least in linux programs, when I've done profiles, most of the cycles were taken up in console i/o. Printf, etc.
malloc/free is a major bottleneck in many systems, especially in SMPs. So, yes a faster (without locks) malloc/free would speed up a program several times (not just marginally). I have several examples of programs that spends up to 80% of their time in malloc/free. But of course this depends on the kind of program. Multithreaded object-oriented programs running on SMP hardware is defenetly one category of programs that would benefit from this.
It would also require a modification to the kernel - so that the brk() system call would simply call the hardware variants.
Most likely, but is that really a problem?
I think that it probably *would* remove flexibility, but at the same time I think that's a good thing. Part of the problem with memory leaks and stuff is the flexibility...
Yes, it would probably remove some flexibility (which is not a good thing). But I don't see what you mean when you say that flexibility is part of the memory leak problems.
They are going to insert malloc/free into hardware, that doesn't mean they will build garbage collection into the hardware (which is a completely different thing). However, as the project description says, it could improve garbage collection performance, but that is another issue.
DMM (Dynamic Memory Management) is a well known term, not something new as the slashdot article tries to make it sound like.
Building DMM into hardware it not a new idea. However, I'm not really sure if it has been successfully implemented in any "real" CPUs.
The security button is removed with:
Netscape*toolBar.viewSecurity.isEnabled: false
"[...]OpenStep, which has a very similar object hierarchy to NeXTSTEP, but uses a better object allocation model[...]"
better object allocation model? would you care to explain that a bit further? I would be interested in hearing more about that.
I got this too. I removed my old ~/.mozilla directory (used by M11) and then it worked fine.
Please, stop pretending you don't know.... it is well known that the Mozilla interface SUCKS in terms of responsiveness/redraw time.
The slow GUI is the main reason I don't use it. I really hope removeal of the debug code will make it substantially faster, or else I'm afraid this browser is going nowhere.
I second that.
BeroLinux, for those who don't know, was the first 2.2kernel distribution with everything recompiled to be pentium optimized.
Sorry Dan, but this is not true. Even if it doesn't really matter who did what first I can not resist to point out that Stampede Linux was the first 2.2 kernel distribution that was compiled with pentium optimization.
eh? What makes you think he said that? He said "Win2000 will only run on Intel".. and Merced done by Intel!