I, for one, am disappointed that the Web is as much a visual medium as it's become. I like to think that the Web is best when it's a information, primarily written language, medium. I'm not looking forward to the day when high speed access turns it the Web into just Interactive TV.
Unfortinuately I believe that the webs "graphical medium" is what is the big draw for it.
<aside> When I first got onto the net (ala '89 or so), long before URLs, it was fun. The first time I saw Mosaic, I thought: "Great, another Gopher." -- That is until I saw that it had pictures. Then I knew it was going to be big. </aside>
Now, I have a cousin who has lost all of his sight. He does quite well with his computer. He reads his snail-mail by putting it onto the scanner, and his OCR software reads it to him. He uses the Net and does email and everything else a "sighted" person can do (minus seeing graphics, obvously) -- hell he's an MP3 monster as well:)
If you are playing Quake, let me give you a quick hint on how to make your own Gamer Modem -- at home. Turn off all of the compression and error checking codes on your modem. Not only will you:
Have better transmission times (no buffers to fill then to compress)
Have better error recovery times -- NONE!
Doesn't require special hardware.
Sure, you'll probably end up with garbled packets, but thats what your TCP/IP stack is for. This way your modem won't be trying to retransmit a time-sensitive UDP packet -- you're machine will just toss it out and wait for the next one. -Jeff
I once ran into a guy at a waterpark (which is why I saw him without a shirt on) who was from an eastern European country. Under his left arm on his chest he had tattooed: His Name His Religion His Blood Type A number And one or two other things I forgot. Apparently they use it for a more simple reason... war. Its easier to 1. know what kind of blood to put in the soldiers quicker, and 2. what kind of headstone to give them if they are killed.
Glad to see that you're trying to help. While your at it, why don't you ask them to meet you out by the bikeracks so that you can beat them up?
Acting like a 12 year old doesn't work, of course, you probably didn't listen to any of the other things your parents should have told you...
Re:Why it's sooo hard for this stuff to happen....
on
911 Calls Linux
·
· Score: 2
Like he said, Motorolla and MS take the blame for the NT box that...
No no, see, the PHB wants to be able to tell his boss that its "Someone elses problem". Actual liability is somewhat of a farse in a life or death situation unless you have software specifically written for that purpose. Which, in this case, they didn't before, and they don't now.
I totally agree. The other thing that paying your bill every month gets you is a very good credit rating. I know someone who did everything with this checking/debit cards and he's having a Hell of a time getting a Mortage (for his first house).
Having a blank credit raiting is almost as bad as a bad credit rating if you want to buy a house!
Sorry, his main premise stands that you can use the netmask as part of the IP address and... blamo... you get 64 bits to use. Of course, datagrams don't USE the netmask when they are moving around.
And the bits about binary being different than decimal... those are actually pretty funny. "Mathematically" if you have two different sets that map 1:1 for all values (0-255, 00000000-11111111, 00-ff) then they are the same. The only way it would be "different" is if they encoded the "2" then encoded the "5" then encoded the "5", which of course, is wrong. (and what got us into the y2k problem in the first place)
This is because, in my mind, since it was encrypted in the first place, that means that there exists a code somewhere (or some "code" somewhere) that is the key to its undoing.
Yes and no. Breaking any one particular encrypted stream is one thing. Breaking the algorithm so that *ANY* stream can be decoded (quickly) is a totally different one. Cracking an RSA stream is possible (given enough effort), but cracking the entire algorithm is nearly impossible to do. (You could be rich if you could do that:) ).
The pot of gold is not the any one particular stream, but rather the generic code-removal algorithm.
Most of the Slashdot comments I have seen have portrayed Evolution as fact. When, in fact, Evolution is a theory, just as any other scientifically established theory. There are facts supporting Evolution, but there are also facts discrediting Evoultion.
Its BOTH. Its a theory just like the law of Gravity is a theory -- we can't prove that just because for as long as anyone can find that the Earth had a gravitational of 9.80665 m/(s**2) that tomorrow it will not be something else.
Since it cannot be proven to be false for all instances, it cannot be more than a theory. It is a fact based upon the definition of the word fact.
Ignorance is the enemy, truth is the answer. Well, before you spout off what a fact is, please look up the definition (#4).
While not exactly of the same style as the article (um, threads?), we've created a distrubited batch system called Condor that uses an oppertunistic model of batch queueing. In it we can run job-parallel (i.e. you run 100 jobs that all do a part of the problem), or PVM (Parallel virtual machine - message passing).
That said, if it was really nice and easy to make a program parallel, it would be done by now. The closest thing I've ever seen is some nifty magic done by the SGI power-compilers to take advantage of SMP machines... but thats only for one machine. That doesn't even touch inter-machine problems.
Now, if you are re-writing your code from scratch (as the paper seems to indicate that you'll have to), then by all means you can use an existing package (most likely MPI or PVM) to get the parallelism that you are looking for...
I work on a project called Condor that many people use to queue up things like renders.
Its an oppertunistic batch queueing system. Meaning that it will run jobs on your machine when you are not there. In the lab I work in, I'll easially be able to get a month of CPU time out of the machines on people's desks that would otherwise go unused...
While it has been in Madison for a long time, they now have a Milwaukee edition, and I believe (as someone else said) there's a Denever edition. If anything it's coming to a college near you.
I'm still sad that they got rid of a couple things over the years. The campus poliece blotter was extremely funny, not to mention Drunk of the Week. Too bad I never made D.O.W., but at least I knew someone who was...ah those were the days...
Back in the day, all the stories were based near/around Madison. Now they are generically based.
Telco's in DSL-land are already positioning to have an open-access protocol. Its called PPPOE (PPP Over Ethernet). Here's a text drawing:
END USER -----DSL-->TELCO-->AC------> ISP1 -----> RADIUS
(connect this to the AC..../ doesn't allow for tags) -----> ISP2
(imagine more connections to the AC...)
So the X is the end user. He/She connects to the telco using the DSL line. The DSL connects (some details have been left off) to the Telco (i.e. Ameritech where I'm at). The Telco has something called an AC. The AC is the Access Concentrator. Think of it like a router. Then connected to the AC are a bunch of big ISP's that have provisioned lines, Value added features, etc.
So, how does PPPOE work? You run a PPPOE client on your machine. And lets say for this example that you have selected ISP1 as your ISP. When you login (using name and password), a PPPOE packet is sent down the DSL. The Telco routes the PPPOE packet based upon the ISP you select. The PPPOE packet eventually makes its way to a PPPOE server (not shown) and that server authenticates you through RADIOUS (which all the ISP's have been already using and know how to bill from, etc.)
So, once you are authenticated, you get an IP address from the ISP! Any traffic to that address goes through the PPPOE server, down the DSL line encapsulated within PPPOE packets and eventually your machine. The advantages to this system are that:
The Telco doesn't have to do any real work. They just wait for the ISP's to provision lines, and then the ISP's pay them to have access to the AC. The end user ends up shelling out money for the DSL line.
The ISP's know Raidus really well. They have been using it for years and do all of their accounting/billing from it.
Unfortunately, this also means that they will be able to charge for a "network login", i.e. for each machine you have. Granted you could (and probably will) run a Linux machine and NAT all the rest of your machines.
Right now DHCP has a drawback that there isn't a really good way to authenticate someone before handing them an IP (without making everyone register their MAC addresses...what a pain.)
Being a fan is one thing. Shelling out the $70 bucks for two years (and whatever an ISP adds) seems a bit much for a "fan site". I would be more inclined to not trust: 1. anything that starts at the root of a site 2. anything that doesn't match: s!^http://[^/]+\.edu/~!
Alot of money seems to be circulating these days for completely useless things.
That most certainly is your opinion. A free market society determines the usefulness of a good/service by the amount that is paid for it. Sure for you (and me and lots of other people) it isn't worth 600 big ones, but to that persion/corporeation it was.
Just because you think something's worthless doesn't mean that someone else does.
Besides they have to shell out the annual fee just like everyone else.
If I recall correctly that the ultra-secret group that got together to determine the algorithm used a 32 bit key... but the first 16 digits were zeros...
That was shown by some grad students a few months back, anyway.
I would have thought that Africa would be the last place to introduce this technology(self repairing cable,...
Self healing cable has been around for a long time (SONET and dual-connected FDDI are two examples that have been around for a long time.)
Basically if one side of the ring around africa breaks, then all the traffic will go around the other side.
In any case, most of the bandwith will be most likely used by telcos for plain old voice traffic anyway (which they refer to as the "billions" that they are going to save...)
It's BetaSP, not BetaMAX
on
DIVX is dead
·
· Score: 1
Uh... Please, don't confuse the issue. Broadcast TV does *NOT* use BetaMAX, they use BetaSP. Yes, still made by Sony, but there is a world of difference (in picture, and in price).
Its analogous to the difference between Cassette audio tapes and DAT...
Now I really wouldn't mind owning a BetaSP deck (without the multi-thousand dollar cost).
If your looking for the exact link its: http://www.nwc.com/1010/1010f144.html#5 . Also, there is the video of the whole thing so you can see it announced. (thats the whole awards ceremony...)
While you probably are correct, another interpretation that I thought of when I read that was that they may be worried that the source code was changed (which is not a bad thing for them to worry about).
Slashdot Mirror
Offtopic slightly, I know, but here's the shamless plug, and pointer, to the BOFH Excuse server.
Cheers.
Unfortinuately I believe that the webs "graphical medium" is what is the big draw for it.
<aside>
When I first got onto the net (ala '89 or so), long before URLs, it was fun. The first time I saw Mosaic, I thought: "Great, another Gopher." -- That is until I saw that it had pictures. Then I knew it was going to be big.
</aside>
Now, I have a cousin who has lost all of his sight. He does quite well with his computer. He reads his snail-mail by putting it onto the scanner, and his OCR software reads it to him. He uses the Net and does email and everything else a "sighted" person can do (minus seeing graphics, obvously) -- hell he's an MP3 monster as well :)
- Have better transmission times (no buffers to fill then to compress)
- Have better error recovery times -- NONE!
- Doesn't require special hardware.
Sure, you'll probably end up with garbled packets, but thats what your TCP/IP stack is for. This way your modem won't be trying to retransmit a time-sensitive UDP packet -- you're machine will just toss it out and wait for the next one. -JeffI once ran into a guy at a waterpark (which is why I saw him without a shirt on) who was from an eastern European country. Under his left arm on his chest he had tattooed:
His Name
His Religion
His Blood Type
A number
And one or two other things I forgot. Apparently they use it for a more simple reason... war. Its easier to 1. know what kind of blood to put in the soldiers quicker, and 2. what kind of headstone to give them if they are killed.
Scairy...
Glad to see that you're trying to help. While your at it, why don't you ask them to meet you out by the bikeracks so that you can beat them up?
Acting like a 12 year old doesn't work, of course, you probably didn't listen to any of the other things your parents should have told you...
No no, see, the PHB wants to be able to tell his boss that its "Someone elses problem". Actual liability is somewhat of a farse in a life or death situation unless you have software specifically written for that purpose. Which, in this case, they didn't before, and they don't now.
I totally agree. The other thing that paying your bill every month gets you is a very good credit rating. I know someone who did everything with this checking/debit cards and he's having a Hell of a time getting a Mortage (for his first house).
Having a blank credit raiting is almost as bad as a bad credit rating if you want to buy a house!
And the bits about binary being different than decimal... those are actually pretty funny. "Mathematically" if you have two different sets that map 1:1 for all values (0-255, 00000000-11111111, 00-ff) then they are the same. The only way it would be "different" is if they encoded the "2" then encoded the "5" then encoded the "5", which of course, is wrong. (and what got us into the y2k problem in the first place)
Here's an example: http://18446744072649904394/
Which comes from this program: #include <iostream.h>
;
main() {
unsigned long long a;
a = (192<<24) + (215<<16)+ (17<<8) + (10)
cout << a << endl;
}
Yes and no. Breaking any one particular encrypted stream is one thing. Breaking the algorithm so that *ANY* stream can be decoded (quickly) is a totally different one. Cracking an RSA stream is possible (given enough effort), but cracking the entire algorithm is nearly impossible to do. (You could be rich if you could do that :) ).
The pot of gold is not the any one particular stream, but rather the generic code-removal algorithm.
Its BOTH. Its a theory just like the law of Gravity is a theory -- we can't prove that just because for as long as anyone can find that the Earth had a gravitational of 9.80665 m/(s**2) that tomorrow it will not be something else.
Since it cannot be proven to be false for all instances, it cannot be more than a theory. It is a fact based upon the definition of the word fact.
Ignorance is the enemy, truth is the answer.
Well, before you spout off what a fact is, please look up the definition (#4).
While not exactly of the same style as the article (um, threads?), we've created a distrubited batch system called Condor that uses an oppertunistic model of batch queueing. In it we can run job-parallel (i.e. you run 100 jobs that all do a part of the problem), or PVM (Parallel virtual machine - message passing).
That said, if it was really nice and easy to make a program parallel, it would be done by now. The closest thing I've ever seen is some nifty magic done by the SGI power-compilers to take advantage of SMP machines... but thats only for one machine. That doesn't even touch inter-machine problems.
Now, if you are re-writing your code from scratch (as the paper seems to indicate that you'll have to), then by all means you can use an existing package (most likely MPI or PVM) to get the parallelism that you are looking for...
I work on a project called Condor that many people use to queue up things like renders.
Its an oppertunistic batch queueing system. Meaning that it will run jobs on your machine when you are not there. In the lab I work in, I'll easially be able to get a month of CPU time out of the machines on people's desks that would otherwise go unused...
While it has been in Madison for a long time, they now have a Milwaukee edition, and I believe (as someone else said) there's a Denever edition. If anything it's coming to a college near you.
I'm still sad that they got rid of a couple things over the years. The campus poliece blotter was extremely funny, not to mention Drunk of the Week. Too bad I never made D.O.W., but at least I knew someone who was...ah those were the days...
Back in the day, all the stories were based near/around Madison. Now they are generically based.
END USER -----DSL-->TELCO-->AC------> ISP1 -----> RADIUS
(connect this to the AC... ./ doesn't allow for tags) -----> ISP2
(imagine more connections to the AC...)
So the X is the end user. He/She connects to the telco using the DSL line. The DSL connects (some details have been left off) to the Telco (i.e. Ameritech where I'm at). The Telco has something called an AC. The AC is the Access Concentrator. Think of it like a router. Then connected to the AC are a bunch of big ISP's that have provisioned lines, Value added features, etc.
So, how does PPPOE work? You run a PPPOE client on your machine. And lets say for this example that you have selected ISP1 as your ISP. When you login (using name and password), a PPPOE packet is sent down the DSL. The Telco routes the PPPOE packet based upon the ISP you select. The PPPOE packet eventually makes its way to a PPPOE server (not shown) and that server authenticates you through RADIOUS (which all the ISP's have been already using and know how to bill from, etc.)
So, once you are authenticated, you get an IP address from the ISP! Any traffic to that address goes through the PPPOE server, down the DSL line encapsulated within PPPOE packets and eventually your machine. The advantages to this system are that:
- The Telco doesn't have to do any real work. They just wait for the ISP's to provision lines, and then the ISP's pay them to have access to the AC. The end user ends up shelling out money for the DSL line.
- The ISP's know Raidus really well. They have been using it for years and do all of their accounting/billing from it.
- Unfortunately, this also means that they will be able to charge for a "network login", i.e. for each machine you have. Granted you could (and probably will) run a Linux machine and NAT all the rest of your machines.
- Right now DHCP has a drawback that there isn't a really good way to authenticate someone before handing them an IP (without making everyone register their MAC addresses...what a pain.)
-Jeff Ballard (jballard@nwc.com.nospam)Being a fan is one thing. Shelling out the $70 bucks for two years (and whatever an ISP adds) seems a bit much for a "fan site". I would be more inclined to not trust:
1. anything that starts at the root of a site
2. anything that doesn't match: s!^http://[^/]+\.edu/~!
-Jeff
Alot of money seems to be circulating these days for completely useless things.
That most certainly is your opinion. A free market society determines the usefulness of a good/service by the amount that is paid for it. Sure for you (and me and lots of other people) it isn't worth 600 big ones, but to that persion/corporeation it was.
Just because you think something's worthless doesn't mean that someone else does.
Besides they have to shell out the annual fee just like everyone else.
If I recall correctly that the ultra-secret group that got together to determine the algorithm used a 32 bit key... but the first 16 digits were zeros...
That was shown by some grad students a few months back, anyway.
I would have thought that Africa would be the last place to introduce this technology(self repairing cable, ...
Self healing cable has been around for a long time (SONET and dual-connected FDDI are two examples that have been around for a long time.)
Basically if one side of the ring around africa breaks, then all the traffic will go around the other side.
In any case, most of the bandwith will be most likely used by telcos for plain old voice traffic anyway (which they refer to as the "billions" that they are going to save...)
Uh... Please, don't confuse the issue. Broadcast TV does *NOT* use BetaMAX, they use BetaSP. Yes, still made by Sony, but there is a world of difference (in picture, and in price).
Its analogous to the difference between Cassette audio tapes and DAT...
Now I really wouldn't mind owning a BetaSP deck (without the multi-thousand dollar cost).
-Jeff
If your looking for the exact link its: http://www.nwc.com/1010/1010f144.html#5 .
Also, there is the video of the whole thing so you can see it announced. (thats the whole awards ceremony...)
While you probably are correct, another interpretation that I thought of when I read that was that they may be worried that the source code was changed (which is not a bad thing for them to worry about).
Gullible? Gullible is not in the dictionary...