Two things: 1) Microsoft is claiming WINCE, in addition to running batteries down on palm computers, is a realtime OS (its not). They envision blood monitors and such running WINCE (ACK!!!). 2) Real time at the interrupt level on Linux? Perhaps you have a different definition of realtime than I do. Having the SCSI layer turn off interrupts for 10ms at a time kills any deterministic response more than that. And god forbit you're running an IDE hard drive with UDMA! Unless you were speaking of RT/Linux, but that's an entirely different (and spartan) beast.
I've spent quite a bit of time lately trying to get a patent (Don't kill me, its for hardware I designed), so I have a few thoughts on the whole process.
The patent office as it stands right now is so heavily biased towards large corporations that the only solution I can see is to rebuild it from scratch. Origionally, the inventor would write up a patent, (with free help from a patent examiner!) and submit it. It would be evaluated and either rejected or accepted. It cost about a hundred bucks.
Now, don't even THINK about trying to get a patent without a lawyer. Most likely, you'll get rejected, and if you manage to get accepted, you'll end up with a mostly worthless patent. If you're serious, and have a good, patentable idea, expect to pay about 3000 dollars for a decent patent.
However, if you're a large company, and have a staff of good lawyers, you can get just about anything patented, no matter how ridiculous. And companies usually try to price their licenses so that its cheaper to just submit to extortion than to fight their patent (which is so expensive that an individual shouldn't even consider it).
Large companies have their own patent portfolio, and if they need someone else's patent, they arrange a cross-licensing agreement, instead of paying fees, since they can most likely extort the other patent holders as well.
The current system has been manipulated to stratify the status quo, and protect slow moving large companys from small innovators. Write your congresscritter, not that'll it'll do any good.
Re:So what I want to know is when my TNT works...
on
No Next Q3Test
·
· Score: 1
Just a quick note: The 3DFX drivers are a quick 3D only hack (not that this is a bad thing, they're the best way to play q3 on linux right now).
The Nvidia drivers are a traditional GLX driver, in which all textures and drawing commands must get funneled through a unix socket. This means that all drawing happens in the context of the X server, which is great if you have a SMP box, but sucks otherwise.
Theoretically, nvidia could have made a full-screen only hack like 3DFX, but evidently they have choosen to wait for XF4 and PI's DRI to do 3D fullscreen and in a window at full speed. And I can't say I really blame them. Of course, I'm not a game junkie, and the current nvidia driver is pretty damn fast on my SMP machine. Primarily because my code does no texturing.
Piracy serves an important purpose. Say John Q. Businessman needs to create MS Office documents, since all his clients use Office (and MS claims there is no such thing as network effects!). He has to go buy a win32 environment to run Office (ignoring macs for now). Since MS is the only place to get an OS to run Win32 apps, he must buy their software.
Since he doesn't have a choice, MS can charge whatever they want. But wait, they don't! There is some competitive force keeping prices down (not very well, but its there). At some point, Windows would just cost too much, and the fines for piracy, times the risk that they'll get caught is less than the price of Windows. The smart businessman then pirates windows.
Microsoft does have competition. Its the software pirates. Obviously this doesn't make piracy good; I'd prefer that said businessman used free software, but it is certainly something to think about.
I found those articles "telecosm" in forbes: www.forbes.com/asap/gilder however, the page appears to be gone. They allude to the publishing of the series, so it may have been taken down to boost sales of the book, published by Simon & Schuster. Its a very interesting series on the future of telecommunications, Dark Fiber, and bandwidth. It might even be worth buying the book if you're interested in this subject.
At least in theory, dual CPUs should help quite a bit in X. The client sends requests to the server, which then does the drawing. With a dual box, they both get their own CPU to work on, meaning the client can prepare the next request while the server is drawing the last one.
In the X FAQ, one of the suggestions for speeding X up is to "swap" machines with a coworker, and use his machine set to your display, and vice versa. The point of this exercise is to reduce the constant context switches (1 per X request). Of course, with extensions like MIT-SHM this is (obviously) no longer a win, however on a SMP box, it very well could be.
A better answer is to edit out everything in/etc/inetd.conf that you don't explicitly need. Also, check your init scripts (/etc/rc.d/*) for things you don't need (There are pretty little graphical tools to help you here). Now keep an eye on whatever distro/os you use's security page. If they don't have a security page, switch to one that does.
Of course, these suggestions apply equally to Linux or OpenBSD; The reasons for OpenBSD's security have little to do with its default install, and a lot more to do with careful code auditing.
Everyone with 1/10 of a clue knows that all vendor (especially apple) provided benchmarks are just a twist of words away from outright lies. There really isn't much point complaining about it. Would you trust the claims of a used car salesman, or would you test drive the car? It works out OK, since the only people who believe this tripe don't need much computing power anyway.
"internet explosion" is not a thing, its a magic phrase. It works like this; you're giving a presentation to stupid people with lots of money who keep hearing about this internet thing. Prefix our magic phrase with the word "tapping" and you'll invoke the grand money gods. You will suddenly find yourself with ridiculous amounts of cash. I suggest the next available flight to somewhere with no diplomatic ties to the US.
Quake2 has a software renderer, and Carmack has stated that it uses GL's transformation pipeline. It does not, however, make user of the lighting engine, as Quake (and most ofther FPS games) use lightmaps instead of vertex lighting.
transmeta.COM name server ns.transmeta.COM transmeta.COM name server ns1.best.COM transmeta.COM name server ns2.best.COM transmeta.COM name server ns3.best.COM neosilicon.transmeta.COM has address 206.184.214.14 best-gw-aux.transmeta.COM has address 206.184.214.33 ssl.transmeta.COM has address 206.184.214.15 localhost.transmeta.COM has address 127.0.0.1 neon-best.transmeta.COM has address 206.184.214.10 www.transmeta.COM has address 206.184.214.11 loghost.transmeta.COM has address 127.0.0.1 best-gw.transmeta.COM has address 206.184.214.1 ns.transmeta.COM has address 206.184.214.14
neosilicon? new silicon? Interesting. Please commence out-of-control speculation... NOW.
Having used both QT and GTK--, my opinion is that I like GTK-- better, mainly because I also code in C for smaller apps, and I want a consistant widget set. (QT/C bindings were just to prove it can be done, they suck to actually use -- I have never seen a real app developed with them). The big advantage for QT is that it does a *LOT* more, stuff like socket objects, and various things you would not strictly consider to be part of a widget set. If I get some free time, I will be working on a companion library to GTK that does all the non-gui support tasks like threads and sockets in a portable, OO way.
This is definately not a beginners project. Writing this driver would require simulating the functions of a DSP chip on a general purpose CPU. It would require excellent understanding of how modems work at the signal level. It requires hard real time for acceptable performance. The reason no one has attempted reverse engineering one is that doing so would be very specific to a certain winmodem, and the work involved would be orders of magnitude greater than required for just about any other driver.
They might be just acronyms to you, but I've written code for all of the above APIs. Let me tell ya something; if you don't care that the APIs are not from MS (DirectWhatever), then Linux is just as good for game development, excluding 3D sound support, which I mentioned. The moderation worked perfectly, the origional poster though that writing games for linux was like writing games for DOS. This is obviously untrue, as anyone who has written code for both systems can tell. I pointed people to resources to actually learn about these APIs, so they would not make incorrect, unqualified statements like the origional posters, and yours.
go to www.opengl.org and compare that to the DOS of 10 years ago. libGGI is a lot less of a pain in the ass than the VBE(or god forbid, before VBE). The only place linux is really lacking in terms of APIs for games is something like Window's DirectSound3D, and the more advanced extensions to it.
Check out the n64 demo scene. Check out the net yaroze. Once you buy a machine there is no way to keep you from reverse engineering it and making your own tools. Or are you mad because they're selling you a specialized graphics workstation at a HUGE loss and want to make part of it back on their tools?
This could be done by linking with winelib and writing a custom binary loader (similar to the origional wine binary, but for DLLs). I looked into doing this for x11amp, to load winamp plugins, but it would take some unholy hackery. It'd be a lot of work that would be better spent making cool GPLed plugins.
I haven't seen the RT Message client, but the IRC client is a bunch of javascript. When running, it just looks like a web page. This does not add any signifigant code the mozilla browser itself. Even if the IM client adds hundreds of thousands of lines of code to mozilla, *YOU* have the source, and *YOU* can recompile it with whatever you want in there. I'm sure after mozilla gets released people will rpm or deb slimmed down browser only mozillas.
Here at school, they converted an NT lab to a linux lab, that NFS mounts your home dir. Users cannot screw up anything on the computer, except their own files. Beats the hell out of the old system, where people would come and ask to use the computer you're working on, because they "saved their paper in c:\temp" Any centrally managed unix system gives you most of the same advantages of this thin client, while not overloading your network with things that are better done locally.
I find it somewhat telling that the article didn't mention Andrea Arcangeli and Gerhard Wichert, workin at SuSe and Siemens respectively and wrote this patch pretty much dual-handedly. I suppose with the corpratization of linux, the companies are more important now than the actual people who make linux what it is. At the very least, a link to Andrea's archived message on l-k would give credit where it is due.
my comment was rather tongue in cheek, pointing out how silly the origional comment was by taking it to its logical conclusion. (I suppose I could have gone all the way to assembling your CPU with a soldering iron to make it more obvious). Sorry I was unclear in my intent, I'll tag my sarcasm next time.
Slashdot Mirror
Two things:
1) Microsoft is claiming WINCE, in addition to running batteries down on palm computers, is a realtime OS (its not). They envision blood monitors and such running WINCE (ACK!!!).
2) Real time at the interrupt level on Linux? Perhaps you have a different definition of realtime than I do. Having the SCSI layer turn off interrupts for 10ms at a time kills any deterministic response more than that. And god forbit you're running an IDE hard drive with UDMA! Unless you were speaking of RT/Linux, but that's an entirely different (and spartan) beast.
I've spent quite a bit of time lately trying to get a patent (Don't kill me, its for hardware I designed), so I have a few thoughts on the whole process.
The patent office as it stands right now is so heavily biased towards large corporations that the only solution I can see is to rebuild it from scratch. Origionally, the inventor would write up a patent, (with free help from a patent examiner!) and submit it. It would be evaluated and either rejected or accepted. It cost about a hundred bucks.
Now, don't even THINK about trying to get a patent without a lawyer. Most likely, you'll get rejected, and if you manage to get accepted, you'll end up with a mostly worthless patent. If you're serious, and have a good, patentable idea, expect to pay about 3000 dollars for a decent patent.
However, if you're a large company, and have a staff of good lawyers, you can get just about anything patented, no matter how ridiculous. And companies usually try to price their licenses so that its cheaper to just submit to extortion than to fight their patent (which is so expensive that an individual shouldn't even consider it).
Large companies have their own patent portfolio, and if they need someone else's patent, they arrange a cross-licensing agreement, instead of paying fees, since they can most likely extort the other patent holders as well.
The current system has been manipulated to stratify the status quo, and protect slow moving large companys from small innovators. Write your congresscritter, not that'll it'll do any good.
Just a quick note: The 3DFX drivers are a quick 3D only hack (not that this is a bad thing, they're the best way to play q3 on linux right now).
The Nvidia drivers are a traditional GLX driver, in which all textures and drawing commands must get funneled through a unix socket. This means that all drawing happens in the context of the X server, which is great if you have a SMP box, but sucks otherwise.
Theoretically, nvidia could have made a full-screen only hack like 3DFX, but evidently they have choosen to wait for XF4 and PI's DRI to do 3D fullscreen and in a window at full speed. And I can't say I really blame them. Of course, I'm not a game junkie, and the current nvidia driver is pretty damn fast on my SMP machine. Primarily because my code does no texturing.
Piracy serves an important purpose. Say John Q. Businessman needs to create MS Office documents, since all his clients use Office (and MS claims there is no such thing as network effects!). He has to go buy a win32 environment to run Office (ignoring macs for now). Since MS is the only place to get an OS to run Win32 apps, he must buy their software.
Since he doesn't have a choice, MS can charge whatever they want. But wait, they don't! There is some competitive force keeping prices down (not very well, but its there). At some point, Windows would just cost too much, and the fines for piracy, times the risk that they'll get caught is less than the price of Windows. The smart businessman then pirates windows.
Microsoft does have competition. Its the software pirates. Obviously this doesn't make piracy good; I'd prefer that said businessman used free software, but it is certainly something to think about.
www.alsa-project.org -- This has an excellent driver for the 4D wave. Plus, Alsa beats the hell out of OSS.
I found those articles "telecosm" in forbes: www.forbes.com/asap/gilder however, the page appears to be gone. They allude to the publishing of the series, so it may have been taken down to boost sales of the book, published by Simon & Schuster. Its a very interesting series on the future of telecommunications, Dark Fiber, and bandwidth. It might even be worth buying the book if you're interested in this subject.
Am I the only one thinking an X terminal for every room of the house?
At least in theory, dual CPUs should help quite a bit in X. The client sends requests to the server, which then does the drawing. With a dual box, they both get their own CPU to work on, meaning the client can prepare the next request while the server is drawing the last one.
In the X FAQ, one of the suggestions for speeding X up is to "swap" machines with a coworker, and use his machine set to your display, and vice versa. The point of this exercise is to reduce the constant context switches (1 per X request). Of course, with extensions like MIT-SHM this is (obviously) no longer a win, however on a SMP box, it very well could be.
A better answer is to edit out everything in /etc/inetd.conf that you don't explicitly need. Also, check your init scripts (/etc/rc.d/*) for things you don't need (There are pretty little graphical tools to help you here). Now keep an eye on whatever distro/os you use's security page. If they don't have a security page, switch to one that does.
Of course, these suggestions apply equally to Linux or OpenBSD; The reasons for OpenBSD's security have little to do with its default install, and a lot more to do with careful code auditing.
This ones a real beauty, previously owned by an little old lady who only drove it to church on sundays!
Ignore the sawdust overflowing from the transmission! What smoke?
Everyone with 1/10 of a clue knows that all vendor (especially apple) provided benchmarks are just a twist of words away from outright lies. There really isn't much point complaining about it. Would you trust the claims of a used car salesman, or would you test drive the car? It works out OK, since the only people who believe this tripe don't need much computing power anyway.
"internet explosion" is not a thing, its a magic phrase. It works like this; you're giving a presentation to stupid people with lots of money who keep hearing about this internet thing. Prefix our magic phrase with the word "tapping" and you'll invoke the grand money gods. You will suddenly find yourself with ridiculous amounts of cash. I suggest the next available flight to somewhere with no diplomatic ties to the US.
Quake2 has a software renderer, and Carmack has stated that it uses GL's transformation pipeline. It does not, however, make user of the lighting engine, as Quake (and most ofther FPS games) use lightmaps instead of vertex lighting.
transmeta.COM name server ns.transmeta.COM
transmeta.COM name server ns1.best.COM
transmeta.COM name server ns2.best.COM
transmeta.COM name server ns3.best.COM
neosilicon.transmeta.COM has address 206.184.214.14
best-gw-aux.transmeta.COM has address 206.184.214.33
ssl.transmeta.COM has address 206.184.214.15
localhost.transmeta.COM has address 127.0.0.1
neon-best.transmeta.COM has address 206.184.214.10
www.transmeta.COM has address 206.184.214.11
loghost.transmeta.COM has address 127.0.0.1
best-gw.transmeta.COM has address 206.184.214.1
ns.transmeta.COM has address 206.184.214.14
neosilicon? new silicon? Interesting. Please commence out-of-control speculation... NOW.
Having used both QT and GTK--, my opinion is that I like GTK-- better, mainly because I also code in C for smaller apps, and I want a consistant widget set. (QT/C bindings were just to prove it can be done, they suck to actually use -- I have never seen a real app developed with them). The big advantage for QT is that it does a *LOT* more, stuff like socket objects, and various things you would not strictly consider to be part of a widget set. If I get some free time, I will be working on a companion library to GTK that does all the non-gui support tasks like threads and sockets in a portable, OO way.
This is definately not a beginners project. Writing this driver would require simulating the functions of a DSP chip on a general purpose CPU. It would require excellent understanding of how modems work at the signal level. It requires hard real time for acceptable performance. The reason no one has attempted reverse engineering one is that doing so would be very specific to a certain winmodem, and the work involved would be orders of magnitude greater than required for just about any other driver.
They might be just acronyms to you, but I've written code for all of the above APIs. Let me tell ya something; if you don't care that the APIs are not from MS (DirectWhatever), then Linux is just as good for game development, excluding 3D sound support, which I mentioned. The moderation worked perfectly, the origional poster though that writing games for linux was like writing games for DOS. This is obviously untrue, as anyone who has written code for both systems can tell. I pointed people to resources to actually learn about these APIs, so they would not make incorrect, unqualified statements like the origional posters, and yours.
go to www.opengl.org and compare that to the DOS of 10 years ago. libGGI is a lot less of a pain in the ass than the VBE(or god forbid, before VBE). The only place linux is really lacking in terms of APIs for games is something like Window's DirectSound3D, and the more advanced extensions to it.
Check out the n64 demo scene. Check out the net yaroze. Once you buy a machine there is no way to keep you from reverse engineering it and making your own tools. Or are you mad because they're selling you a specialized graphics workstation at a HUGE loss and want to make part of it back on their tools?
This could be done by linking with winelib and writing a custom binary loader (similar to the origional wine binary, but for DLLs). I looked into doing this for x11amp, to load winamp plugins, but it would take some unholy hackery. It'd be a lot of work that would be better spent making cool GPLed plugins.
I haven't seen the RT Message client, but the IRC client is a bunch of javascript. When running, it just looks like a web page. This does not add any signifigant code the mozilla browser itself. Even if the IM client adds hundreds of thousands of lines of code to mozilla, *YOU* have the source, and *YOU* can recompile it with whatever you want in there. I'm sure after mozilla gets released people will rpm or deb slimmed down browser only mozillas.
Here at school, they converted an NT lab to a linux lab, that NFS mounts your home dir. Users cannot screw up anything on the computer, except their own files. Beats the hell out of the old system, where people would come and ask to use the computer you're working on, because they "saved their paper in c:\temp" Any centrally managed unix system gives you most of the same advantages of this thin client, while not overloading your network with things that are better done locally.
I find it somewhat telling that the article didn't mention Andrea Arcangeli and Gerhard Wichert, workin at SuSe and Siemens respectively and wrote this patch pretty much dual-handedly. I suppose with the corpratization of linux, the companies are more important now than the actual people who make linux what it is. At the very least, a link to Andrea's archived message on l-k would give credit where it is due.
my comment was rather tongue in cheek, pointing out how silly the origional comment was by taking it to its logical conclusion. (I suppose I could have gone all the way to assembling your CPU with a soldering iron to make it more obvious). Sorry I was unclear in my intent, I'll tag my sarcasm next time.
True geeks build their own machine from parts that they carefully evaluated and picked out themselves.
Please don't call yourself a geek if you simply buy whatever massmarket crap comes out of our major corporations.