That's funny-- I always find it odd when the British and Indian folks call math "maths." It's an interesting cultural difference. And I disagree with your abbreviation argument-- "math" is a prefix of "mathematics" while "maths" is not. In fact, pluralizing "math" makes it seem like you concede that there does exist a "mathematic" singular, which you abbreviate to "math," and then pluralize again to mimic the original word.
I wonder how much counterfeit money is directed toward vending machines. It would seem that a machine would be easier to fool than a person, since presumably one only has to fool the built-in scanner, without need to reproduce the properties of the paper (smell, quality, feel, etc.).
Re:I wish I knew where I could find the MS fonts
on
Libranet 2.8 Review
·
· Score: 1
I've never encountered a font, on any platform (or at least linux, windows, dos, mac, atari st, amiga, sun), that looks better for text editing and writing code than the default fixed 6x9 font used by xterms. The characters are small enough not to be overly wasteful of space, but somehow appear very clear and unambiguous-- they are drawn really well.
Sure, *this one* is fixed, but it sure doesn't inspire confidence in the security of their service. Who knows if there are other holes left for crackers to exploit...
How about a grant model? This already works pretty well for research, where the results are made available to the public. An agency (government, consortium of corporations who realize software is useful, but don't want to develop it alone, etc.) puts out a call for proposals for desired project, and funds one which seems best (by reputation, plan, etc.). The grant covers the developers' time, administrivia, etc.
If this results in a smaller total number of employed software developers, does that mean the previous model was inefficient? Maybe eliminating that inefficiency is good in the long term-- the extra developers can now work on new challenges that will further increase our standard of living, instead of reinventing the wheel in all the proprietary software shops.
With NAT, client machines can't accept incoming connections because they are not visible to the network. Yes, you can set up port forwarding at the NAT point, but you can only do it for one machine per port. If all 100 machines behind your NAT want to serve Web pages from port 80, you're out of luck.
This is fundamentally bad. The Internet's purpose is free exchange of information, where any node can serve bits to the rest of the network. The reason letting kludges like NAT take root is dangerous is that it makes some nodes less free than others, and potentially prevents innovations that would be possible in the original open network.
I got my Red Hat ISOs from Red Hat with no problem.... I don't see any analysis of why I might want to "donate" my bandwidth to some person I don't know, or why I should expect bandwidth from anyone except the person I'm downloading from.
There is no need for any analysis to convince you. If the method is effective, it will be used simply due to economics. Suppose a thousand people behind cable modems want a video from a server on a modem link. Here is a simple CS101 problem: figure out the schedule that lets that file propagate to all thousand users in the shortest amount of time. Hint: it does not happen by everyone downloading from the server.
Now tell me why I shouldn't open up some of my bandwidth for the privilege of being among these thousand users and getting the file I want as soon as possible?
It doesn't work well with NAT, as mentioned above, and the solution is not to proclaim NAT "broken"
The solution is to banish NAT back to the depths of hell where it came from and roll out IPv6. Oh well, one can hope...
I keep all my ports closed anyway, I'm not going to open them for someone else.
NAT breaks the Internet's noble property of openness, whereby any node can accept and send IP datagrams to any other node. It is unfortunate that NAT took root as it did, but fortunately it is becoming an increasing nuisance as more services spring up which require ability to accept incoming connections. Hopefully IPv6 will some day let us get rid of NAT's festering carcass.
In my next residence I will register my phone under an alias. If anyone calls asking for the alias, then they are immediately identified as a telemarketer and I will tell them there is no one here by that name. This crap has gone far enough.
Hey, that's an awesome idea! Register yourself under the name Mr. Mudder Fokker, so that they avoid calling you in fear of getting in trouble for obscene phone calls:)
OT: I think making a constitutional amendment mandating a balanced budget may be going too far, but make it so that you can't be re-elected as president if you have a budget in the red (or something like that -- though not sure what to do about second-term presidents)..
Here are a couple of dumb ideas for dealing with a president that ends up with a budget deficit:
Give a small vote penalty to his party in the next elections. Say, if Bush ends up with a deficit, give the Repubs a -10% penalty in the next elections.
This is a pretty natural thing to do with any kind of graph (PGP key servers, blogs, p2p network topologies, you name it). And the larger graph one can get, the more interesting it gets. I drool when I think about the kind of analyses the people at AOL must do with their buddy list database...
The companies have so far invested more than $100,000 in the deal, excluding investments in the pilot roaming service.
Wow, they invested a whopping $100,000! Are they sure they can afford that much? With that amount of support, I'm sure the WLAN services will be everywhere in no time!
The scalability of the mesh network depends on the connectivity patterns. Note that there is no backbone in one of these-- all links have the same capacity. Thus, if all nodes just talk to their neighbors, there is no limit to the size of the network-- obviously, distant parts have no effect on each other.
However, if the connectivity patterns are global, the mesh won't scale as well. For example, when arbitrary pairs of nodes are just as likely to want to talk to each other, one can show that the capacity alotted to such pairs diminishes as 1/sqrt(N) on a 2-D mesh (slice the mesh in half, and note that on the order of N/2 nodes must pass through the sqrt(N) nodes along the dividing line).
The obvious conclusion is that we won't be able to build wide area mesh networks out of a single type of link. But that's what backbones are for, and the ad hoc networks are still damn useful in local environments, such as meetings, towns, etc. The missing part is some sort of standard resource discovery protocol whereby a node can find services available to it, such as routing, exchange business cards or ebooks or mp3s, control the stereo, download photos from the digital camera, get readings from the sensor network embedded in the building, download the local building's map, etc.
A question about your universe
on
Ask Larry Niven
·
· Score: 5, Interesting
Mr. Niven,
First, thanks for the awesome books that you have written-- I am a big fan.
My biggest question about the universe in which many of the stories take place is about the "Outsiders." They seem to be extremely technologically advanced; they jump in whenever something impossible needs to be done, such as the Puppeteers moving their worlds around. How come the Outsiders didn't end up competing with Protectors, or do something on the scale of building a ringworld?
One one hand, I agree that with a clean room implementation it's easier to argue that there is no infringement. But on the other hand, not having done it doesn't automatically mean that they are infringing on the original work... though it may be harder to prove. I wonder how this would be decided in court-- how similar does a program have to be to qualify as derived work?
Slashdot Mirror
That's funny-- I always find it odd when the British and Indian folks call math "maths." It's an interesting cultural difference. And I disagree with your abbreviation argument-- "math" is a prefix of "mathematics" while "maths" is not. In fact, pluralizing "math" makes it seem like you concede that there does exist a "mathematic" singular, which you abbreviate to "math," and then pluralize again to mimic the original word.
I wonder how much counterfeit money is directed toward vending machines. It would seem that a machine would be easier to fool than a person, since presumably one only has to fool the built-in scanner, without need to reproduce the properties of the paper (smell, quality, feel, etc.).
I've never encountered a font, on any platform (or at least linux, windows, dos, mac, atari st, amiga, sun), that looks better for text editing and writing code than the default fixed 6x9 font used by xterms. The characters are small enough not to be overly wasteful of space, but somehow appear very clear and unambiguous-- they are drawn really well.
Sure, *this one* is fixed, but it sure doesn't inspire confidence in the security of their service. Who knows if there are other holes left for crackers to exploit...
How about a grant model? This already works pretty well for research, where the results are made available to the public. An agency (government, consortium of corporations who realize software is useful, but don't want to develop it alone, etc.) puts out a call for proposals for desired project, and funds one which seems best (by reputation, plan, etc.). The grant covers the developers' time, administrivia, etc.
If this results in a smaller total number of employed software developers, does that mean the previous model was inefficient? Maybe eliminating that inefficiency is good in the long term-- the extra developers can now work on new challenges that will further increase our standard of living, instead of reinventing the wheel in all the proprietary software shops.
Speaking of which, there seems to be a GPL-ed port with Unicode support at tvision.sourceforge.net. Here is a screenshot.
I think I remember it. It used half-full characters (ASCII 218 or so) to draw beveled buttons in text mode :)
Count the flashbacks to Turbovision!
This reminds me of when I was young.
:^)
Let me guess, all sections of your railroad track went uphill!
With NAT, client machines can't accept incoming connections because they are not visible to the network. Yes, you can set up port forwarding at the NAT point, but you can only do it for one machine per port. If all 100 machines behind your NAT want to serve Web pages from port 80, you're out of luck.
This is fundamentally bad. The Internet's purpose is free exchange of information, where any node can serve bits to the rest of the network. The reason letting kludges like NAT take root is dangerous is that it makes some nodes less free than others, and potentially prevents innovations that would be possible in the original open network.
I got my Red Hat ISOs from Red Hat with no problem. ...
:)
I don't see any analysis of why I might want to "donate" my bandwidth to some person I don't know, or why I should expect bandwidth from anyone except the person I'm downloading from.
There is no need for any analysis to convince you. If the method is effective, it will be used simply due to economics. Suppose a thousand people behind cable modems want a video from a server on a modem link. Here is a simple CS101 problem: figure out the schedule that lets that file propagate to all thousand users in the shortest amount of time. Hint: it does not happen by everyone downloading from the server.
Now tell me why I shouldn't open up some of my bandwidth for the privilege of being among these thousand users and getting the file I want as soon as possible?
It doesn't work well with NAT, as mentioned above, and the solution is not to proclaim NAT "broken"
The solution is to banish NAT back to the depths of hell where it came from and roll out IPv6. Oh well, one can hope...
I keep all my ports closed anyway, I'm not going to open them for someone else.
Sounds like an old maid
NAT breaks the Internet's noble property of openness, whereby any node can accept and send IP datagrams to any other node. It is unfortunate that NAT took root as it did, but fortunately it is becoming an increasing nuisance as more services spring up which require ability to accept incoming connections. Hopefully IPv6 will some day let us get rid of NAT's festering carcass.
I tried to calculate the answer to your question, but all I got are these division by zero errors. Go figure.
In my next residence I will register my phone under an alias. If anyone calls asking for the alias, then they are immediately identified as a telemarketer and I will tell them there is no one here by that name. This crap has gone far enough.
:)
Hey, that's an awesome idea! Register yourself under the name Mr. Mudder Fokker, so that they avoid calling you in fear of getting in trouble for obscene phone calls
Once a janitor, always a janitor!
Every cool OS technology has been invented in Multics. Ask any OS researcher :)
Here are a couple of dumb ideas for dealing with a president that ends up with a budget deficit:
Follow this link.
This is a pretty natural thing to do with any kind of graph (PGP key servers, blogs, p2p network topologies, you name it). And the larger graph one can get, the more interesting it gets. I drool when I think about the kind of analyses the people at AOL must do with their buddy list database...
The companies have so far invested more than $100,000 in the deal, excluding investments in the pilot roaming service.
Wow, they invested a whopping $100,000! Are they sure they can afford that much? With that amount of support, I'm sure the WLAN services will be everywhere in no time!
Maybe the author meant to write $100M.
The scalability of the mesh network depends on the connectivity patterns. Note that there is no backbone in one of these-- all links have the same capacity. Thus, if all nodes just talk to their neighbors, there is no limit to the size of the network-- obviously, distant parts have no effect on each other.
However, if the connectivity patterns are global, the mesh won't scale as well. For example, when arbitrary pairs of nodes are just as likely to want to talk to each other, one can show that the capacity alotted to such pairs diminishes as 1/sqrt(N) on a 2-D mesh (slice the mesh in half, and note that on the order of N/2 nodes must pass through the sqrt(N) nodes along the dividing line).
The obvious conclusion is that we won't be able to build wide area mesh networks out of a single type of link. But that's what backbones are for, and the ad hoc networks are still damn useful in local environments, such as meetings, towns, etc. The missing part is some sort of standard resource discovery protocol whereby a node can find services available to it, such as routing, exchange business cards or ebooks or mp3s, control the stereo, download photos from the digital camera, get readings from the sensor network embedded in the building, download the local building's map, etc.
Mr. Niven,
First, thanks for the awesome books that you have written-- I am a big fan.
My biggest question about the universe in which many of the stories take place is about the "Outsiders." They seem to be extremely technologically advanced; they jump in whenever something impossible needs to be done, such as the Puppeteers moving their worlds around. How come the Outsiders didn't end up competing with Protectors, or do something on the scale of building a ringworld?
Purchasing manager:
One one hand, I agree that with a clean room implementation it's easier to argue that there is no infringement. But on the other hand, not having done it doesn't automatically mean that they are infringing on the original work... though it may be harder to prove. I wonder how this would be decided in court-- how similar does a program have to be to qualify as derived work?