But see, that's the thing...StarOffice may be huge and bloated, but it has a huge amount of functionality. If, by opensourcing it, either a) a non-bloated version comes out, or b) the other gpl office suites use some of the code to further their developments, then this is a Very Good Thing(TM).
Basically, what this means is that we could get a *really good* free office suite in the forseeable future.
The problem then is that you can't know someone is online without them knowing that you know;-). That could work, but in order for you to be connected to the network would mean someone could send out a "ping" to X (X being the user you wanna know is online). The ping could be a fake IM with a null message (and clients would most likely not display null messages, cause that's just a smart thing to do:P). If the message is accepted, you know they're on.
Of course, it could be implemented so that that wouldn't be possible, but with a distributed system I'm sure there would be *some* way of getting around it if you really wanted to. Especially because it's distributed...meaning messages would be going through your computer to get to someone elses...meaning you could "sniff" traffic to see if the person had sent out the message.
The main problem with this (that I see) would be that ANYONE would know EVERYONE that's on at one time. This would mean huge amounts of spam, and would mean there's no way for you to say "I don't want X knowing I'm on." Also, it would have the same problem Gnutella has, where depending on who you connected through, you'll get different responses to searches, meaning someone could be on, but you wouldn't be guaranteed of knowing it.
Eh, actually, from what I've heard (from numerous sources), the xig drivers are of MUCH lower quality than X4.0, and are much less stable. This may not be true for all graphics cards, but I've heard (again, from numerous places) that at least on 3dfx cards it doesn't hold anything on the X 4.0 servers.
Eh, but just like not many programs are made with Tk in Windows (which has those hooks), I doubt too many will be made for OS X. Oh, and notice I didn't say anything about Java...but we're getting off topic...the fact is, most programs won't be ported to gtk or qt (which probably won't be ported themselves for at least a few months after OS X comes out)...they'll be ported to Carbon or Cocoa. I doubt AQUA will become opensource, because that would take away Apple's big advantage over Linux/FreeBSD/etc...a really nice gui and graphical config system.
OpenAT (http://www.openat.org) is the sound toolkit developed by Loki. Nice, but this is not something they're working on; they're just using it in Blender
Actually, once you get the hang of it it gets pretty easy to use. One hand on the mouse and one on the keyboard means it optimizes how quickly you can do things...read a few tutorials and you'll start to get the hang of it.
The GLUT library has both C, C++ (same as C), FORTRAN, and Ada programming bindings. The GLUT source code distribution is portable to nearly all OpenGL implementations for the X Window System and Windows 9x/NT. GLUT also works well with Brian Paul's Mesa, a freely available implementation of the OpenGL API.
The current version of the GLUT API is 3. The current source code distribution is GLUT 3.7.
So, um...it's not opensource? Why can I download the source from here then?
But see, 4 is not yet as optimized as 3 for OpenGL. If you benchmark xf86 3.3.6 vs 4 with a 3dfx card, you'd find either virtually no difference, or you'd find that 3.3.6 is actually faster right now (even if it still has bugs, it's still as fast or faster, because the 4 driver right now hasn't been optimized any more than the 3 one has.)
Another thing...they used to do this. Remember, in the old days when Win3.1 was still the norm, where if your machine wouldn't run a game fast, you would use a boot disk, so none of the drivers or anything would load, and you'd have extra memory? How annoying is it to reboot your machine just to play a game? And then have to reboot again if you want to do anything else? (I know, this is like booting from Linux/BeOS to Windows, but you can still DO things from windows...like play a different game, or surf the web, or...etc etc etc)
This is NOT benchmarking. Quake uses a very small amount of the OpenGL functions, meaning this is only testing which driver has been best optimized for quake (know "MiniGL" that used to come with 3d cards instead of OpenGL? That's a stripped down version of OpenGL, containing only the parts of OpenGL that quake/quake2 needed, and heavily optimized for both of them). In order to do a true benchmark, it would need to be a program that tests ALL of OpenGL's functionality, instead of a minority of it.
is MNG supported by IE? By a lot of graphics devel programs? The fact that it's in Mozilla is great, but if it's going to catch on, it needs something more used to support it. I don't know anything about MNG, but from what the original poster says, it sounds like it's a very new format...well, new formats need visible applications to support them before they take off. Hopefully we'll see that with MNG.
1) If they're writing them in Python, Perl, C, C++, (oh, and POSIX is not a language), they still need to use the Carbon or Cocoa APIs (or hooks to the APIs).
2) Cocoa and Carbon do NOT run on top of the BSD kernel, they run on top of Aqua, the GUI
3) As Cocoa and Carbon are not OpenSource, and Apple has no reason to port them, they will not get ported
4) Eh, not sure what hardware drivers you're talking about, but Aqua, the GUI, is NOT OpenSource, meaning the drivers for the graphics hardware won't be...meaning only Apple can port them.
One Microsoft Way
Yes, but very useful for servers
on
Linux BIOS
·
· Score: 1
I agree completely with what you're saying, but there is one place where this would fit: servers. Servers need as close to 100% uptime as possible, and (if it's high end) will probably have some SCSI in it, meaning long boot times. Taking the 2 minutes out of the boot time for that would be well worth it in a server situation, where you're not worried about having to put (or have) another OS on it.
but I find it hard to believe that personal electronics actually have an effect on jets (mainly because I've been on plenty of flights surrounded by people ignoring those rules -- and I've yet to be involved in a crash).
That's like saying "well, because I've drunk and driven before, and I've been on the road with people who are been drinking and driving, and I haven't crashed yet, I should keep drinking and driving, because it is therefore not possible that I will crash."
Hey, anyone know if there's anything like this in the US? I know there's an online company that will tape shows for you and then you can watch them later, but I haven't heard of anything like this.
Wow, you actually thought this post was meant to be realistic? All I have to say is...this post is funnier than the one it's responding to if that's true.
I asked Daryll about this recently on the dri-devel mailing list...here is was he said:
3dfx does have a high-performance OpenGL implementation. They put a reasonable amount of manpower into it. I think this is showing you the performance improvement you can get by putting real resources behind the project.
The nVidia situation is interesting. In that case, they are using essentially the same code base between Linux and Windows. The question to ask is why the Linux version is then slower? It could be an OS issue, a compiler issue, a driver issue, or something else entirely. If we close that gap, whatever it is, all the implementations get better. By the way, 2.3 kernels have been signficantly faster, so the 2.4 release may help with the difference.
The argument is that Open Source efforts can do it better, but you have to qualify that a bit. What defines better? In some cases we're not concentrating on the same focus. For example, Mesa tries very hard to be a complete and conforming version of OpenGL. In some cases that may mean losing some performance compared to tweaking of Q3A at the expense of everything else. Some of the security and stability fixes in the MGA DRI code mean we lose a bit of performance. The 3dfx in-a-window implementation under Windows is quite a bit slower than their full screen mode.
You also have to compare manpower efforts. 3dfx and ATI put a lot of people on working on their drivers. They are each paying PI for one engineer. That limits not only how fast the drivers can be produced but how good they will be. We're also really not seeing much help from the community. We'd love to have more people contribute.
The 3dfx driver has remained mostly unchanged (except for bug fixes) for the last year. That's because I've been payed to get it running on different boards (V3/V4/V5), to fix some bugs, and improve the infrastructure (DRI). We really haven't had the resources to spend doing optimizations and to rewrite some of the really ugly parts in the code.
The bottom line is that there is room to improve. I have no doubt that with the right attention all the drivers would be very close to their windows counterparts. We just need some good people to do the work.
Slashdot Mirror
But see, that's the thing...StarOffice may be huge and bloated, but it has a huge amount of functionality. If, by opensourcing it, either a) a non-bloated version comes out, or b) the other gpl office suites use some of the code to further their developments, then this is a Very Good Thing(TM).
Basically, what this means is that we could get a *really good* free office suite in the forseeable future.
It's just like the woman he fell in love with...a replicant designed not to know he's a replicant.
The problem then is that you can't know someone is online without them knowing that you know ;-). That could work, but in order for you to be connected to the network would mean someone could send out a "ping" to X (X being the user you wanna know is online). The ping could be a fake IM with a null message (and clients would most likely not display null messages, cause that's just a smart thing to do :P). If the message is accepted, you know they're on.
Of course, it could be implemented so that that wouldn't be possible, but with a distributed system I'm sure there would be *some* way of getting around it if you really wanted to. Especially because it's distributed...meaning messages would be going through your computer to get to someone elses...meaning you could "sniff" traffic to see if the person had sent out the message.
One Microsoft Way
The main problem with this (that I see) would be that ANYONE would know EVERYONE that's on at one time. This would mean huge amounts of spam, and would mean there's no way for you to say "I don't want X knowing I'm on." Also, it would have the same problem Gnutella has, where depending on who you connected through, you'll get different responses to searches, meaning someone could be on, but you wouldn't be guaranteed of knowing it.
One Microsoft Way
dri.sourceforge.net it takes a long time to download (cvs), and about an hour and a half to build on my machine (dual PII 300 MHz).
One Microsoft Way
Eh, actually, from what I've heard (from numerous sources), the xig drivers are of MUCH lower quality than X4.0, and are much less stable. This may not be true for all graphics cards, but I've heard (again, from numerous places) that at least on 3dfx cards it doesn't hold anything on the X 4.0 servers.
One Microsoft Way
Eh, but just like not many programs are made with Tk in Windows (which has those hooks), I doubt too many will be made for OS X. Oh, and notice I didn't say anything about Java...but we're getting off topic...the fact is, most programs won't be ported to gtk or qt (which probably won't be ported themselves for at least a few months after OS X comes out)...they'll be ported to Carbon or Cocoa. I doubt AQUA will become opensource, because that would take away Apple's big advantage over Linux/FreeBSD/etc...a really nice gui and graphical config system.
One Microsoft Way
OpenAT (http://www.openat.org) is the sound toolkit developed by Loki. Nice, but this is not something they're working on; they're just using it in Blender
Actually, it's OpenAL (http://www.openal.org).
One Microsoft Way
Actually, once you get the hang of it it gets pretty easy to use. One hand on the mouse and one on the keyboard means it optimizes how quickly you can do things...read a few tutorials and you'll start to get the hang of it.
One Microsoft Way
Uh, what are you talking about?
From http://www.opengl.org/Documentation/G LUT.html
The GLUT library has both C, C++ (same as C), FORTRAN, and Ada programming bindings. The GLUT source code distribution is portable to nearly all OpenGL implementations for the X Window System and Windows 9x/NT. GLUT also works well with Brian Paul's Mesa, a freely available implementation of the OpenGL API.
The current version of the GLUT API is 3. The current source code distribution is GLUT 3.7.
So, um...it's not opensource? Why can I download the source from here then?
One Microsoft Way
But see, 4 is not yet as optimized as 3 for OpenGL. If you benchmark xf86 3.3.6 vs 4 with a 3dfx card, you'd find either virtually no difference, or you'd find that 3.3.6 is actually faster right now (even if it still has bugs, it's still as fast or faster, because the 4 driver right now hasn't been optimized any more than the 3 one has.)
One Microsoft Way
Another thing...they used to do this. Remember, in the old days when Win3.1 was still the norm, where if your machine wouldn't run a game fast, you would use a boot disk, so none of the drivers or anything would load, and you'd have extra memory? How annoying is it to reboot your machine just to play a game? And then have to reboot again if you want to do anything else? (I know, this is like booting from Linux/BeOS to Windows, but you can still DO things from windows...like play a different game, or surf the web, or...etc etc etc)
One Microsoft Way
This is NOT benchmarking. Quake uses a very small amount of the OpenGL functions, meaning this is only testing which driver has been best optimized for quake (know "MiniGL" that used to come with 3d cards instead of OpenGL? That's a stripped down version of OpenGL, containing only the parts of OpenGL that quake/quake2 needed, and heavily optimized for both of them). In order to do a true benchmark, it would need to be a program that tests ALL of OpenGL's functionality, instead of a minority of it.
One Microsoft Way
This isn't the DVD-rental shit. This is a video codec based on MPEG-4.
One Microsoft Way
In your email address, you say "THISWORDOPTIONAL", but which word are you talking about?
One Microsoft Way
is MNG supported by IE? By a lot of graphics devel programs? The fact that it's in Mozilla is great, but if it's going to catch on, it needs something more used to support it. I don't know anything about MNG, but from what the original poster says, it sounds like it's a very new format...well, new formats need visible applications to support them before they take off. Hopefully we'll see that with MNG.
One Microsoft Way
1) If they're writing them in Python, Perl, C, C++, (oh, and POSIX is not a language), they still need to use the Carbon or Cocoa APIs (or hooks to the APIs).
2) Cocoa and Carbon do NOT run on top of the BSD kernel, they run on top of Aqua, the GUI
3) As Cocoa and Carbon are not OpenSource, and Apple has no reason to port them, they will not get ported
4) Eh, not sure what hardware drivers you're talking about, but Aqua, the GUI, is NOT OpenSource, meaning the drivers for the graphics hardware won't be...meaning only Apple can port them.
One Microsoft Way
I agree completely with what you're saying, but there is one place where this would fit: servers. Servers need as close to 100% uptime as possible, and (if it's high end) will probably have some SCSI in it, meaning long boot times. Taking the 2 minutes out of the boot time for that would be well worth it in a server situation, where you're not worried about having to put (or have) another OS on it.
One Microsoft Way
The Quicktime 4 movie codec is made by Sorenson...
One Microsoft Way
but I find it hard to believe that personal electronics actually have an effect on jets (mainly because I've been on plenty of flights surrounded by people ignoring those rules -- and I've yet to be involved in a crash).
That's like saying "well, because I've drunk and driven before, and I've been on the road with people who are been drinking and driving, and I haven't crashed yet, I should keep drinking and driving, because it is therefore not possible that I will crash."
One Microsoft Way
Hey, anyone know if there's anything like this in the US? I know there's an online company that will tape shows for you and then you can watch them later, but I haven't heard of anything like this.
One Microsoft Way
unfortunately, dialpad is windows-only. (Or was when i last checked a couple months ago).
One Microsoft Way
Wow, you actually thought this post was meant to be realistic? All I have to say is...this post is funnier than the one it's responding to if that's true.
One Microsoft Way
the DRI drivers aren't as fast as the 3.3.6 ones yet tho.
One Microsoft Way
I asked Daryll about this recently on the dri-devel mailing list...here is was he said:
3dfx does have a high-performance OpenGL implementation. They put a reasonable amount of manpower into it. I think this is showing you the performance improvement you can get by putting real resources behind the project.
The nVidia situation is interesting. In that case, they are using essentially the same code base between Linux and Windows. The question to ask is why the Linux version is then slower? It could be an OS issue, a compiler issue, a driver issue, or something else entirely. If we close that gap, whatever it is, all the implementations get better. By the way, 2.3 kernels have been signficantly faster, so the 2.4 release may help with the difference.
The argument is that Open Source efforts can do it better, but you have to qualify that a bit. What defines better? In some cases we're not concentrating on the same focus. For example, Mesa tries very hard to be a complete and conforming version of OpenGL. In some cases that may mean losing some performance compared to tweaking of Q3A at the expense of everything else. Some of the security and stability fixes in the MGA DRI code mean we lose a bit of performance. The 3dfx in-a-window implementation under Windows is quite a bit slower than their full screen mode.
You also have to compare manpower efforts. 3dfx and ATI put a lot of people on working on their drivers. They are each paying PI for one engineer. That limits not only how fast the drivers can be produced but how good they will be. We're also really not seeing much help from the community. We'd love to have more people contribute.
The 3dfx driver has remained mostly unchanged (except for bug fixes) for the last year. That's because I've been payed to get it running on different boards (V3/V4/V5), to fix some bugs, and improve the infrastructure (DRI). We really haven't had the resources to spend doing optimizations and to rewrite some of the really ugly parts in the code.
The bottom line is that there is room to improve. I have no doubt that with the right attention all the drivers would be very close to their windows counterparts. We just need some good people to do the work.
One Microsoft Way