He's shown us how to get a circuit established cheaply. Actually doing it may be made difficult by your phone company, but it shows how they are trying to rape data services for so much more money than than things like security systems.
The real problem is that you want connectivity to the internet. Even if you find someone who's willing to piggy back you on their circuit, chances are they're violating their terms of service by doing it. That may get them cut off if they're caught. If they're doing NAT it would be hard for their ISP to find out.
If you really want to offer legitament ISP services, then you'd get your circuit to another ISP, and you'd ask to buy transit rights. Unfortunetly, these don't come cheap. You have to pay them for allowing your data to cross their network, and they probably have to pay transit fees to another ISP which they'll pass along to you.
It's a great idea. I'd like to see lots of free bandwidth. There's just many many hands between you and the global internet and they all want their cut. By the time you're done it isn't cheap.
Another major problem with this scheme is that the video quality is going to be awful. 500Mb means it's going to be equivalent to VCD/MPEG1 which is about the same as a cheap VCR.
A DVD stream uses MPEG2 and gets up to 10 Mb/s. If you assume an average of 5Mb/s (which is low) and a 90 minute movie, that's 27GB.
They could be planning to use MPEG4 which is better quality, but even then they'll be compressing the hell out of it to get a film down to that size.
I'd rather drive to the video store or use NetFlix and watch a real DVD. Not to mention I can do that a lot sooner after the films release.
Re:Problems with the PS2 linux...
on
X-server for PS2
·
· Score: 3, Informative
1) I don't know if they're making you sign an NDA. I've been allowed to talk about the unit at this point.
2) Yes, you are correct. You can not make a stand alone app. They only allow their CD/DVDs to work. You have no access to firmware and boot. This is Sony's bread and butter and they do want to protect it. If you can make a stand alone app you don't have to pay them royaltees. Royaltees are how they make money. Good or bad, that's how it works.
3) That's not correct. You get full system/programming documentation and get direct access to the sound and graphics. You get very low level libraries, so it isn't very friendly but you can do it. If I recall correctly nothing is statically linked.
4) Yep. Again, this goes back to #2.
5) Yes, it uses sync-on-green. That used to be a standard. It's not really a protection scheme. In fact, I suspect they just don't have enough lines to put the sync anywhere else.
6) Sony is looking at releasing it in the US. I have no information on how that's going. My guess is that it went over well enough in Japan that they'll do it here. If they were worried about hacking they wouldn't have released it in japan.
Sony's Linux Comes with XFree 3.3.6
on
X-server for PS2
·
· Score: 4, Insightful
There seems to be some real confusion here. The Sony PS2 Linux comes with XFree 3.3.6 running on their hardware.
Metrolink is selling their Micro-X product. That's based on XFree 4.0 and runs in a much smaller footprint. You don't need it, but with the memory limits of the PS2 it's not a bad idea.
I've got a Japanese PS2 running Linux, so if people have questions, maybe I should do a more complete piece on Slashdot. I did a presentation for my local users group not long ago that could be adapted.
Re:They are targeting set tops, etc.
on
X-server for PS2
·
· Score: 3, Informative
Sorry, you don't have to buy the X server. It comes with XFree 3.3.6. I should know. I've got a japanese playstation running Linux in my bedroom.
Any other questions?
Is OpenNIC's method fair? How did they get to be the defined operator? Didn't they just show up and say they were doing it? Isn't this the same complaint you're making about ICANN?
How do you know the ballot boxes weren't stuffed? When do we change Atlantic's rullership of.net? What do we do when it changes?
You're inventing the same mechanisms and just putting different names on it.
Congratulations you have just invented, get this, a single root. I mentioned this in my post.
TLD's are NOT roots. They are top level domains. One stop down from the root.
How are you going to get everyone to agree on who should be the TLD servers? You can setup any organization you want. Somehow they'll have to agree, perhaps by voting. Guess what, you've just invented equivalent of ICANN.
Don't you people understand why a single root IS required?
Let's make this clear. Right now there is one root. It means there's one name space. If I look up slashdot.org, I'm sure I get this site. Now, lets say there are other root servers that some "new" organization runs. What happens when this new organization puts in an entry for slashdot.org that points somewhere else?
The answer is that you no longer have any faith that any address you use is going to work for yourself. Even if it works for you, if you give an address to someone else you don't know what root server they're using and whether or not it works. Your email address becomes useless, because you can't be sure it actually works for any other user. You can also get into all the criminal activity if users connect to fake sites.
Before someone makes the argument, that the various root servers will just "agree" on how they share the name space, that ends up being the same as having a single root! The "agreement" just happens in the single version of the root server data file.
So, the fact is, that a single consistant root really is needed unless you want to go back to IP addresses for everything.
Now, I do agree that ICANN should be moving faster in granting new domains. They've had a hard time creating policies, and frankly creating policies that work for a VERY diverse group of people is extremely tough. But people should try to understand the TECHNICAL issues instead of just bashing on ICANN as a new form of government.
I have a lot of equipment in my study and I needed a better way to store it. Most shelving really isn't up to the task.
I ended up buying industrial shelving. It's cheap (~$100 for 36" wide by 72" high with 5 shelves) Each shelf is rated to 400 lbs. I bought the grid style shelving, which means it is made up bars every 3 inches or so. That makes cabling nice and easy.
No I don't get to screw my equipment into it, but it sits on it nicely, and even supports a Compaq DS20 without any hassle. (In fact, there are 3 other "normal" machines on the same shelf.) It isn't very pretty or very sexy, but cheap and efficient. (Think warehouse shelving)
You can get this sort of thing at any commercial office supply place. I'm not sure if some place like OfficeDepot would have it or not.
Loki had the same sort of problems when they ported Tribes2. They switched over to a freely available GSM encoding (from a university in Germany). It worked so well they're adding the code to the Windows version so you can chat between versions.
Mutt makes it about as easy as it gets. It has builtin pgp support. Red Hat 7.1 ships with that enabled in the RPMS, so I assume others do as well.
Add the line:
set pgp_autosign=yes
In your config file and it will automatically sign all your outgoing messages.
To encrypt a message, you just compose it and then before sending hit 'p' to go to the PGP menu then 'e' to encrypt, or b to sign and encrypt. It prompts you for your PGP passphase and off it goes. It also remembers the pass phase for the duration of the mutt session to save retyping it. If you want to tell it to forget the pass phrase during the session just hit ctrl-f.
This all great and wonderful, but you have to have installed mutt, installed pgp or gpg, have setup your PGP keyring, and it doesn't hurt to have registered with pgp.net so that everyone can find it. That's not hard to do if you have instructions, but it isn't clear and easy for most users.
Who wants to setup a service where you type a full name of an a band and song, and it assigns you a unique number?
It would be trivial to implement and include in a napster application an automatic translation from names to numbers. Filtering becomes completely useless because only the numbers are stored in Napster.
That was my point. VCRs let you do it, but it was difficult. You had to setup the VCR and you had to wait until the show was over to watch it. None of that is true with a Tivo.
With the Tivo it is trivial to record all your shows. Also you can watch a show and skip the commercials at the same time. (They can play and record simultaneously)
So, even if I want to watch a prime time lineup, I just start watching a little later in the evening. A 20-30 minutes delay for each hour of TV lets me skip all the commercials.
It is hard to describe the difference between a PVR and VCR. It isn't just a VCR with a disk. It is a whole new interface to TV. Yes, it sounds like an add, but even my parents saw the difference in a few days after I bought them a Tivo.
The biggest problem with Tivo selling the data is that the television networks are going to figure out that Tivo users never watch commercials! I watch almost nothing in real-time and always skip the commercials. It's great, but not what television executives want to hear.
There was a piece on 60 minutes recently about PVRs. This was one of their main points, if people don't watch ads then they lose the revenue to sustain the shows. They suggested in-program product placements (the hero holds up a pepsi) and pay-per-view for normal TV as alternatives.
They are expecting PVRs to be the biggest commercial electonics product launch in history. (faster then CDs, DVDs, etc) We're already seeing the problems web sites are having with being advertiser driven. The only thing stopping people from doing that with TV was that it was difficult. Now PVRs make it really easy. It'l be interesting to see how it works out.
4.0.3 is a bugfix release. XFree is now being
more careful about bugfix releases versus feature releases. XFree 4.1 will be the next big feature release and you'll see a lot of improvements in 3D support as well as things like the RENDER extension.
If you don't want even your anonymous viewing information (information that does not identify you or your household) used in any way, simply tell us by calling our toll-free telephone number and it will not be.
Sounds like you can decide whether or not to participate.
John is on several technical advisory boards, including 3dfx and nVidia. It isn't surpising that he'd be invited and accept one for the X-box as well. (I have no first hand knowledge either way.)
Why? Microsoft gets feedback from an industry expert. John gets early access to information about hardware. It is beneficial to both parties.
Is that selling out? No way. From what I can tell, John is perfectly happy to speak his mind regardless of what the hardware manufacturer wants to hear. He still does care about Linux. He joined our (PI) technical meeting a few months ago. You're just seeing less of him recently because he's busy with his next product.
TUX includes a variety of kernel and apache changes. Can you give a rough measure of how each of the changes improved the http performance? I'm interested in the amount of improvement as well as why it improved performance. Do those particular changes have negative impact on the performance of other applications?
What problems are you having with the 3dfx boards? In general it would help if people supplied bug reports intead of just complaining.
Let me correct a few facts that have shown up in this thread. First, all the 3dfx boards that do 2D are supported under XFree 4.0. That means that the Voodoo Graphics and Voodoo2 are not, but since they don't use the X server at all they don't need any special support to run under X.
2D performance of the 3dfx boards is much better under 4.0 than it was under 3.3. The new architecture of 4.0 and the release of 3dfx specs allows the server to run much better.
3D performance is a more complicated matter. I suspect this is where people are complaining. The DRI is a new architecture and it has somewhat different goals than the hack I put together for doing 3D under 3.3. It is designed to run multiple applications, in a window, reliably, and securely. Those extra features do take a toll on performance. Our goal for 4.0 was always to make it work correctly before worrying about the performance. We've just gotten to that point and are now putting some attention on performance. I've made some substantial improvements in the last week, that will be showing up in 4.0 soon.
Finally, Precision Insight and now VA Linux Systems is doing he 2D and 3D work for 3dfx. That work is being done primarily by me. I also did most of the 2D and 3D work under XFree 3.3. (I did have some good contributions by few other developers) So mostly the same people are doing the 3.3 work and 4.0 work. It both cases it was done mostly by a very small group (less than 3) contributors.
The XFree 4.0 support for 3dfx boards is far from laughable. It's one of the best boards supported under 4.0. XFree 4.0 had ambitious goals and we've done a remarkable job pulling them all together. Is the job complete? No. It is running nicely for a lot of people? Absoltely. If you'd like to see it improve then quit complaining and contribute code or bug reports.
No, that's not true about the Alpha. I do have one, and started playing with a port. I would have needed to get a different license from 3dfx, but since I never asked for it, they never gave me one. I don't think there would have been any problem, just a little time/hassle involved.
Although nVidia did release something that could be called open source, the state of the code made it very difficult to use. It was run through a preprocessor. To me that doesn't count.
Actually, Precision Insight has been trying to open up our development process as much as possible. We've recently setup a site on SourceForge called dri.sourceforge.net that has the latest public XFree release and ALL our development. In fact, it is our primary CVS repository at this point.
So, I hope you see that PI is trying to not only do open source development, but also do it in as open a development model as possible.
All the sources and all the specs are available for the 3dfx cards as you mentioned. The GLX list does have a lot of people on it, but in reality only about a half dozen are probably active developers.
I'm the primary author for the 3dfx drivers. I'd love to have more people help with them. That's why we setup dri.sourceforge.net. That's got the current 3dfx drivers, and will include all the drivers from Precision Insight including the Matrox and ATI drivers.
Slashdot Mirror
He's shown us how to get a circuit established cheaply. Actually doing it may be made difficult by your phone company, but it shows how they are trying to rape data services for so much more money than than things like security systems.
The real problem is that you want connectivity to the internet. Even if you find someone who's willing to piggy back you on their circuit, chances are they're violating their terms of service by doing it. That may get them cut off if they're caught. If they're doing NAT it would be hard for their ISP to find out.
If you really want to offer legitament ISP services, then you'd get your circuit to another ISP, and you'd ask to buy transit rights. Unfortunetly, these don't come cheap. You have to pay them for allowing your data to cross their network, and they probably have to pay transit fees to another ISP which they'll pass along to you.
It's a great idea. I'd like to see lots of free bandwidth. There's just many many hands between you and the global internet and they all want their cut. By the time you're done it isn't cheap.
Another major problem with this scheme is that the video quality is going to be awful. 500Mb means it's going to be equivalent to VCD/MPEG1 which is about the same as a cheap VCR.
A DVD stream uses MPEG2 and gets up to 10 Mb/s. If you assume an average of 5Mb/s (which is low) and a 90 minute movie, that's 27GB.
They could be planning to use MPEG4 which is better quality, but even then they'll be compressing the hell out of it to get a film down to that size.
I'd rather drive to the video store or use NetFlix and watch a real DVD. Not to mention I can do that a lot sooner after the films release.
2) Yes, you are correct. You can not make a stand alone app. They only allow their CD/DVDs to work. You have no access to firmware and boot. This is Sony's bread and butter and they do want to protect it. If you can make a stand alone app you don't have to pay them royaltees. Royaltees are how they make money. Good or bad, that's how it works.
3) That's not correct. You get full system/programming documentation and get direct access to the sound and graphics. You get very low level libraries, so it isn't very friendly but you can do it. If I recall correctly nothing is statically linked.
4) Yep. Again, this goes back to #2.
5) Yes, it uses sync-on-green. That used to be a standard. It's not really a protection scheme. In fact, I suspect they just don't have enough lines to put the sync anywhere else.
6) Sony is looking at releasing it in the US. I have no information on how that's going. My guess is that it went over well enough in Japan that they'll do it here. If they were worried about hacking they wouldn't have released it in japan.
Metrolink is selling their Micro-X product. That's based on XFree 4.0 and runs in a much smaller footprint. You don't need it, but with the memory limits of the PS2 it's not a bad idea.
I've got a Japanese PS2 running Linux, so if people have questions, maybe I should do a more complete piece on Slashdot. I did a presentation for my local users group not long ago that could be adapted.
Sorry, you don't have to buy the X server. It comes with XFree 3.3.6. I should know. I've got a japanese playstation running Linux in my bedroom. Any other questions?
How do you know the ballot boxes weren't stuffed? When do we change Atlantic's rullership of .net? What do we do when it changes?
You're inventing the same mechanisms and just putting different names on it.
TLD's are NOT roots. They are top level domains. One stop down from the root.
How are you going to get everyone to agree on who should be the TLD servers? You can setup any organization you want. Somehow they'll have to agree, perhaps by voting. Guess what, you've just invented equivalent of ICANN.
Let's make this clear. Right now there is one root. It means there's one name space. If I look up slashdot.org, I'm sure I get this site. Now, lets say there are other root servers that some "new" organization runs. What happens when this new organization puts in an entry for slashdot.org that points somewhere else?
The answer is that you no longer have any faith that any address you use is going to work for yourself. Even if it works for you, if you give an address to someone else you don't know what root server they're using and whether or not it works. Your email address becomes useless, because you can't be sure it actually works for any other user. You can also get into all the criminal activity if users connect to fake sites.
Before someone makes the argument, that the various root servers will just "agree" on how they share the name space, that ends up being the same as having a single root! The "agreement" just happens in the single version of the root server data file.
So, the fact is, that a single consistant root really is needed unless you want to go back to IP addresses for everything.
Now, I do agree that ICANN should be moving faster in granting new domains. They've had a hard time creating policies, and frankly creating policies that work for a VERY diverse group of people is extremely tough. But people should try to understand the TECHNICAL issues instead of just bashing on ICANN as a new form of government.
I ended up buying industrial shelving. It's cheap (~$100 for 36" wide by 72" high with 5 shelves) Each shelf is rated to 400 lbs. I bought the grid style shelving, which means it is made up bars every 3 inches or so. That makes cabling nice and easy.
No I don't get to screw my equipment into it, but it sits on it nicely, and even supports a Compaq DS20 without any hassle. (In fact, there are 3 other "normal" machines on the same shelf.) It isn't very pretty or very sexy, but cheap and efficient. (Think warehouse shelving)
You can get this sort of thing at any commercial office supply place. I'm not sure if some place like OfficeDepot would have it or not.
Loki had the same sort of problems when they ported Tribes2. They switched over to a freely available GSM encoding (from a university in Germany). It worked so well they're adding the code to the Windows version so you can chat between versions.
Mutt makes it about as easy as it gets. It has builtin pgp support. Red Hat 7.1 ships with that enabled in the RPMS, so I assume others do as well.
Add the line:
set pgp_autosign=yes
In your config file and it will automatically sign all your outgoing messages.
To encrypt a message, you just compose it and then before sending hit 'p' to go to the PGP menu then 'e' to encrypt, or b to sign and encrypt. It prompts you for your PGP passphase and off it goes. It also remembers the pass phase for the duration of the mutt session to save retyping it. If you want to tell it to forget the pass phrase during the session just hit ctrl-f.
This all great and wonderful, but you have to have installed mutt, installed pgp or gpg, have setup your PGP keyring, and it doesn't hurt to have registered with pgp.net so that everyone can find it. That's not hard to do if you have instructions, but it isn't clear and easy for most users.
It would be trivial to implement and include in a napster application an automatic translation from names to numbers. Filtering becomes completely useless because only the numbers are stored in Napster.
Actually I prefer Coca-Cola. Don't endorsements make more money? :-)
With the Tivo it is trivial to record all your shows. Also you can watch a show and skip the commercials at the same time. (They can play and record simultaneously)
So, even if I want to watch a prime time lineup, I just start watching a little later in the evening. A 20-30 minutes delay for each hour of TV lets me skip all the commercials.
It is hard to describe the difference between a PVR and VCR. It isn't just a VCR with a disk. It is a whole new interface to TV. Yes, it sounds like an add, but even my parents saw the difference in a few days after I bought them a Tivo.
The biggest problem with Tivo selling the data is that the television networks are going to figure out that Tivo users never watch commercials! I watch almost nothing in real-time and always skip the commercials. It's great, but not what television executives want to hear. There was a piece on 60 minutes recently about PVRs. This was one of their main points, if people don't watch ads then they lose the revenue to sustain the shows. They suggested in-program product placements (the hero holds up a pepsi) and pay-per-view for normal TV as alternatives. They are expecting PVRs to be the biggest commercial electonics product launch in history. (faster then CDs, DVDs, etc) We're already seeing the problems web sites are having with being advertiser driven. The only thing stopping people from doing that with TV was that it was difficult. Now PVRs make it really easy. It'l be interesting to see how it works out.
4.0.3 is a bugfix release. XFree is now being more careful about bugfix releases versus feature releases. XFree 4.1 will be the next big feature release and you'll see a lot of improvements in 3D support as well as things like the RENDER extension.
Sounds like you can decide whether or not to participate.
Linux Helps Bring Titanic to Life
- |Daryll
Why? Microsoft gets feedback from an industry expert. John gets early access to information about hardware. It is beneficial to both parties.
Is that selling out? No way. From what I can tell, John is perfectly happy to speak his mind regardless of what the hardware manufacturer wants to hear. He still does care about Linux. He joined our (PI) technical meeting a few months ago. You're just seeing less of him recently because he's busy with his next product.
- |Daryll
TUX includes a variety of kernel and apache changes. Can you give a rough measure of how each of the changes improved the http performance? I'm interested in the amount of improvement as well as why it improved performance. Do those particular changes have negative impact on the performance of other applications?
Let me correct a few facts that have shown up in this thread. First, all the 3dfx boards that do 2D are supported under XFree 4.0. That means that the Voodoo Graphics and Voodoo2 are not, but since they don't use the X server at all they don't need any special support to run under X.
2D performance of the 3dfx boards is much better under 4.0 than it was under 3.3. The new architecture of 4.0 and the release of 3dfx specs allows the server to run much better.
3D performance is a more complicated matter. I suspect this is where people are complaining. The DRI is a new architecture and it has somewhat different goals than the hack I put together for doing 3D under 3.3. It is designed to run multiple applications, in a window, reliably, and securely. Those extra features do take a toll on performance. Our goal for 4.0 was always to make it work correctly before worrying about the performance. We've just gotten to that point and are now putting some attention on performance. I've made some substantial improvements in the last week, that will be showing up in 4.0 soon.
Finally, Precision Insight and now VA Linux Systems is doing he 2D and 3D work for 3dfx. That work is being done primarily by me. I also did most of the 2D and 3D work under XFree 3.3. (I did have some good contributions by few other developers) So mostly the same people are doing the 3.3 work and 4.0 work. It both cases it was done mostly by a very small group (less than 3) contributors.
The XFree 4.0 support for 3dfx boards is far from laughable. It's one of the best boards supported under 4.0. XFree 4.0 had ambitious goals and we've done a remarkable job pulling them all together. Is the job complete? No. It is running nicely for a lot of people? Absoltely. If you'd like to see it improve then quit complaining and contribute code or bug reports.
- |Daryll
Although nVidia did release something that could be called open source, the state of the code made it very difficult to use. It was run through a preprocessor. To me that doesn't count.
- |Daryll
David will be doing pretty much the same work as he did before. His first priority at Precision Insight is to get XFree 4.0 out the door.
- |Daryll
Actually, Precision Insight has been trying to open up our development process as much as possible. We've recently setup a site on SourceForge called dri.sourceforge.net that has the latest public XFree release and ALL our development. In fact, it is our primary CVS repository at this point.
So, I hope you see that PI is trying to not only do open source development, but also do it in as open a development model as possible.
- |Daryll
All the sources and all the specs are available for the 3dfx cards as you mentioned. The GLX list does have a lot of people on it, but in reality only about a half dozen are probably active developers.
I'm the primary author for the 3dfx drivers. I'd love to have more people help with them. That's why we setup dri.sourceforge.net. That's got the current 3dfx drivers, and will include all the drivers from Precision Insight including the Matrox and ATI drivers.
- |Daryll