I write C for a living, and enjoy it, but then I do systems level stuff. I would hate to write a modern application in C. It'd be crazy. Java or C# would be pushing it too, they're just not high-level enough.
Take for instance a "modern application" like gtk-gnutella. It is developped in C. Other Gnutella clients exist that are developped in Java. I claim that it is because gtk-gnutella is in C that it is more efficient for the end-user (in terms of memory footprint, CPU usage).
True, a Gnutella client can be seen as "low-level" because it essentially runs close to the system: it manipulates sockets and files (I/Os), and implements network protocols (with the need to construct and decompile complex binary packets).
However, being written in C is not a burden, it's a breeze.
At the risk of being pedantic, please note that LimeWire is NOT a network, as was Napster for instance. LimeWire is just the name of a Gnutella client, as BearShare or gtk-gnutella also are.
So the right wording would have been: "Despite the posts trying to paint this into the next Napster/Gnutella/P2P, I think....".
I welcome your informed opinion. I seldom encounter someone who actually knows that much about Gnutella.
Yes, Gnutella's architecture makes it a better sharing system than pure BitTorrent, for the simple reason that BT creates torrents for a single file, whereas Gnutella servents have been maintaining download meshes (the name of "torrrents" in Gnutella-speak) for many files at once.
One point that perhaps is not known by the general public is that Gnutella is going to have a Distributed Hash Table soon. Something that BT has for some time now, but the addition of a DHT to Gnutella will make magnet URIs finally work right.
The question is then: will it be faster to locate a torrent file or get your hands on a magnet URI that can be fed to any serious Gnutella servent to start a download?
I think Gnutella and BT have much more in common that what a superficial comparison could tell. The architecture is different, the technical implementation radically different, but in the end, both systems will be ultimately converging to something similar. BT will become searchable and Gnutella will have full magnet support, further blurring the distinction.
The only difference is that in Gnutella, you cannot just download a file via a command-line client the way you can do that in BT. But it's just that servent authors never bothered to implement it, and I believe this would be much like emacs/emacsclient: if you have a running Gnutella servent, it is easy to create a command-line tool that will contact that servent to initiate the download, the output being fed to the command-line tool instead of being written to disk, the command-line tool being in charge of writing the data to the disk.
In a sense, BT is simply a re-invention of the Gnutella idea from another angle.
That's because the English spellling of those words is heavily inspired by the French speilling. I like the English spelling better than the American one, I wonder why?:-)
You see, "favour" in French is "faveur", and "colour" in French in "couleur". Oh, and "realise" in French is "realiser".
The basic property rights are that when I own something, I have the right to dispose of it as I wish, or resell it to someone who wants it.
How am I supposed to exerce my right to sell it if I cannot re-record easily new biometrics data for the buying party? And what if the buying party wants to offer the thing to some other third party? (this applies to first hand purchases by the way).
Looks like a dream that will never fly in practice.
Instead of limiting our rights and trying to make copying increasingly more difficult, why not make it less interesting for people to actually copy the thing?
Obviously, there is Demand for movies, but the price is set to high (so that the firm captures profit). Maybe reducing the price would allow demand and offer to meet and reduce copying to a marginal level. There would be less profits, that's for sure.
There are laws against monopolies because in a monopoly, a firm sets the quantities and the prices in such a way that it maximizes its profits. And since that is not good for society as a whole, anti-monopolistic laws exist.
Perhaps there should be laws against firms that try to extract more profit out of the goods they sell than the ones they should be entitled to in a normal competitive situation? Utopia, I know, but this is just to show that there is something that is not right in the way those businesses are run: they clearly have more power to make profit than the ones they would get in a purely competitive market.
Actually, patenting this sort of stuff is genius. Now only Macrovision will be allowed to try and spoof hashes, etc. So P2P freedom fighters need only bankrupt/hijack 1 corporation!
Indeed. They have to actively defend their patents though. If they do not, I believe the patents become moot (no longer enforceable). However, I'm sure it will be their pleasure to actually license the patents to some companies.
You're confusing property with copyright. When you refer to words such as stolen property, this applies to something that belongs to someone else and that you acquired physically.
We're talking about copyright infringement here, which has nothing to do with property deprivation: it has to do with illegal distribution of a work for which you have no rights to distribute, without actually taking ownership of something.
To go back to your book analogy, it would be as if I was redistributing copies of a book, without actually stealing the book. I may not have the right to redistribute the copyrighted work that makes up the book, but I can sell or give away the physical book I own, along with its copyrighted material.
You see, just because you're redistributing Coyrighted material does not make it loose its Copyrighted status. Therefore, the nth degree person taking the work and selling it is the one committing prejudice, and that prejudice cannot be transitively transferred back.
Also, for illegal redistribution of Copyrighted material, infringement must be used instead of piracy.
* Copyright infringement, for profit (illegal reselling of a copyrighted work for which you have no rights). * Copyright infringement without profit (illegal redistribution, for free).
Would you agree that the second is of a lesser degree. And that it is closer to "free advertisement" (as in: "look, I like this movie/music, I think it's cool, don't you share my view?").
Following up on myself: they are using ethanol and not water because water is much harder to splash than ethanol.
I found the answer in another article dealing with the subject.
How would superfluids behave?
on
Bang But No Splash
·
· Score: 3, Interesting
It would be interesting to investigate how superfluids behave.
Since the article hints that the more viscosity, the lower the pressure must be to avoid splashing of the droplet, would superfluids (which have no viscosity at all) behave as expected even under the atmospheric pressure, or even a higher pressure?
Offhand, why are they using ethanol and not water for their study though?
Yes, the basic blocks are 1K blocks. That means that for large files, the whole tree is not kept, only the topmost levels (the topmost 10, I'd say, but that depends on clients).
Moreover, Gnutella is using an alternate hash (Tiger) to construct a hash tree, and the root of that hash tree is appended to the SHA1 of the file to make a unique "bitprint", as understood by Bitzi (http://www.bitzi.com/).
Although it *is* true that a firewall is supposed to guard you against spurious connection attempts to potentially unsecured application listening on those TCP ports, I think most P2P users who know enough simply open the TCP port of their P2P application and possibly forward it to their machine if they are behind a masquerading firewall.
Don't forget that, as a P2P user, your experience with an unfirewalled servent is much more pleasant. The same will go with UDP ports. If you run a Gnutella servent supporting UDP (such as gtk-gnutella, the forthcoming 0.95 release), you can have more results for your searches, delivered directly to you by those who have the files you're searching for.
Therefore, the step to build a reliable TCP-like stream on top of UDP was a natural one, which LimeWire took.
However, we'll see how serious the LimeWire folks are about the claimed "openess" of the Gnutella network: until they publish the specification to implement this TCP-like stream, no other Gnutella servent can support it in an inter-operable way. In effect, currently, you have to run LimeWire to benefit from this "technology". Which gives them a competitive advantage for the time beeing...
You are right that gtk-gnutella works just fine through firewalls.
However, gtk-gnutella (or any other servent) cannot establish a TCP connection between two firewalled hosts.
With the so-called FW2FW transfers, an UDP connection is made by having the two firewalled hosts "punch a hole" through the firewalls by sending datagrams to each other, each knowing the port on which the whole was opened through a third party (a "push-proxy" if you want the gory details).
Once each side can receive the datagrams of the other party, a "reliable" stream of communication is made on top of this UDP link. That's something gtk-gnutella does not support yet.
Now I can promise you that gtk-gnutella will implement this "technology" as soon as LimeWire publishes the specifications for FW2FW.
All of the above are altruistic pursuits. While open source development only serves to inflate your ego.
You miss the point I think. When I develop something that I give away, I don't do it for you. I do it for myself. Because it serves my interests. Then I give it away because I'm a nice guy and because it might be useful to you.
Ego only appears in the last step, when I give it away. I like to get credit for what I did, like everyone else I guess. But ego is not the motivation for doing the stuff in the first place.
Remember something: it costs to release software if you care about maintenance. I only released about 20% of the software I ever wrote for myself, because I would not find the time to maintain it anyway, or it's too specific and would need to be generalized to become useful to someone else.
Whatever Bill Gates thinks about open source software is not going to make me change my way of doing things... Even if Gates's point was proven (which I highly doubt it is), because if the stuff I wrote for me could be useful for you, the chances that you would actually buy it from me are slim, and I would not want to depend on you in any way. Which is why I just give it away in the first place, for you to (ab)use.
I see your point, and I don't dispute it. I'm just saying that software patenting has become an arm race where the system auto-feeds itself, and the more money you have, the more patents you try to submit in order to construct an IP portfolio that could save you should you start to infringe a patent and get sued for that.
In this race, the notion of good/bad is no longer appropriate. Patents are viewed as an investment.
Right, and because this is very easy to do in software (the circumvention), it means that software patents are ridiculous because they only advantage the wealthier companies.
By comparison, try to improve a patented chemical product, like the special glue that is used on Post-Its. Good luck;-) It will not be so easy to do as it is to circumvent a software patent...
Larry Wall's "mailagent" initial implementation appeared in comp.sources.unix in 1989 IIRC, in his "dist" package (the package that contains metaconfig, the Configure script generator).
This early implementation was sending acks to people and was also processing "commands" embedded in the message. I think it pre-dates procmail.
The "mailagent" script finally grew outside of the "dist" package and became a standalone mail filter, with a nicer syntax than procmail. Search for it in CPAN if you want to compare.
Flag me as a troll if you wish, but I would not say that, because they were awarded this patent after submitting it, they are necessarily a bad company. Let me explain:
We have no idea what they are going to do with this patent. If they don't enforce it, maybe they submitted it to protect themselves against a competitor doing the same thing and then trying to enforce it?
We will be able to say that McAfee is "bad" when, and only when, they try to enforce their newly awarded patent against anyone, because only then will we know that their intent was not legitimate protection of their business but rather destruction of other ones, be they for-profit or not.
Let's not forget that this kind of broad patent can be easily circumvented. You only have to do one single improvement over the patented method, and you no longer fall under the patent. It's not the idea of doing Bayesian filterting that can be patented, it is rather HOW it is done. You patent a specific implementation. Ideas cannot be patented, in Europe (don't know about the US system, but I would say it is similar in intent).
That's why people look at prior ART to dispute a patent.
I don't know about SpamAssassin, but I use SpamBayes and I know that their algorithm involves more than just a Bayesian filter. I doubt they would fall under the patent that was just granted, or it would be really bad luck, doing the exact same things the exact same way as it is documented in the patent.
Slashdot Mirror
I write C for a living, and enjoy it, but then I do systems level stuff. I would hate to write a modern application in C. It'd be crazy. Java or C# would be pushing it too, they're just not high-level enough.
Take for instance a "modern application" like gtk-gnutella. It is developped in C. Other Gnutella clients exist that are developped in Java. I claim that it is because gtk-gnutella is in C that it is more efficient for the end-user (in terms of memory footprint, CPU usage).
True, a Gnutella client can be seen as "low-level" because it essentially runs close to the system: it manipulates sockets and files (I/Os), and implements network protocols (with the need to construct and decompile complex binary packets).
However, being written in C is not a burden, it's a breeze.
At the risk of being pedantic, please note that LimeWire is NOT a network, as was Napster for instance. LimeWire is just the name of a Gnutella client, as BearShare or gtk-gnutella also are.
So the right wording would have been: "Despite the posts trying to paint this into the next Napster/Gnutella/P2P, I think....".
Raphael
I welcome your informed opinion. I seldom encounter someone who actually knows that much about Gnutella.
Yes, Gnutella's architecture makes it a better sharing system than pure BitTorrent, for the simple reason that BT creates torrents for a single file, whereas Gnutella servents have been maintaining download meshes (the name of "torrrents" in Gnutella-speak) for many files at once.
One point that perhaps is not known by the general public is that Gnutella is going to have a Distributed Hash Table soon. Something that BT has for some time now, but the addition of a DHT to Gnutella will make magnet URIs finally work right.
The question is then: will it be faster to locate a torrent file or get your hands on a magnet URI that can be fed to any serious Gnutella servent to start a download?
I think Gnutella and BT have much more in common that what a superficial comparison could tell. The architecture is different, the technical implementation radically different, but in the end, both systems will be ultimately converging to something similar. BT will become searchable and Gnutella will have full magnet support, further blurring the distinction.
The only difference is that in Gnutella, you cannot just download a file via a command-line client the way you can do that in BT. But it's just that servent authors never bothered to implement it, and I believe this would be much like emacs/emacsclient: if you have a running Gnutella servent, it is easy to create a command-line tool that will contact that servent to initiate the download, the output being fed to the command-line tool instead of being written to disk, the command-line tool being in charge of writing the data to the disk.
In a sense, BT is simply a re-invention of the Gnutella idea from another angle.
Raphael
That's because the English spellling of those words is heavily inspired by the French speilling. I like the English spelling better than the American one, I wonder why? :-)
:-)
You see, "favour" in French is "faveur", and "colour" in French in "couleur". Oh, and "realise" in French is "realiser".
No need to americanize the spelling, really.
In Europe, of course. We even have a special court to deal with that at the European level.
The basic property rights are that when I own something, I have the right to dispose of it as I wish, or resell it to someone who wants it.
How am I supposed to exerce my right to sell it if I cannot re-record easily new biometrics data for the buying party? And what if the buying party wants to offer the thing to some other third party? (this applies to first hand purchases by the way).
Looks like a dream that will never fly in practice.
Instead of limiting our rights and trying to make copying increasingly more difficult, why not make it less interesting for people to actually copy the thing?
Obviously, there is Demand for movies, but the price is set to high (so that the firm captures profit). Maybe reducing the price would allow demand and offer to meet and reduce copying to a marginal level. There would be less profits, that's for sure.
There are laws against monopolies because in a monopoly, a firm sets the quantities and the prices in such a way that it maximizes its profits. And since that is not good for society as a whole, anti-monopolistic laws exist.
Perhaps there should be laws against firms that try to extract more profit out of the goods they sell than the ones they should be entitled to in a normal competitive situation? Utopia, I know, but this is just to show that there is something that is not right in the way those businesses are run: they clearly have more power to make profit than the ones they would get in a purely competitive market.
Indeed. They have to actively defend their patents though. If they do not, I believe the patents become moot (no longer enforceable). However, I'm sure it will be their pleasure to actually license the patents to some companies.
You're confusing property with copyright. When you refer to words such as stolen property, this applies to something that belongs to someone else and that you acquired physically.
We're talking about copyright infringement here, which has nothing to do with property deprivation: it has to do with illegal distribution of a work for which you have no rights to distribute, without actually taking ownership of something.
To go back to your book analogy, it would be as if I was redistributing copies of a book, without actually stealing the book. I may not have the right to redistribute the copyrighted work that makes up the book, but I can sell or give away the physical book I own, along with its copyrighted material.
You see, just because you're redistributing Coyrighted material does not make it loose its Copyrighted status. Therefore, the nth degree person taking the work and selling it is the one committing prejudice, and that prejudice cannot be transitively transferred back.
Also, for illegal redistribution of Copyrighted material, infringement must be used instead of piracy.
Is Copyright Infringement a crime or a delict?
No harm done.
I have an excellent e-mail filtering system called "mailagent".
Granted, but what about this further distinction:
* Copyright infringement, for profit (illegal reselling of a copyrighted work for which you have no rights).
* Copyright infringement without profit (illegal redistribution, for free).
Would you agree that the second is of a lesser degree. And that it is closer to "free advertisement" (as in: "look, I like this movie/music, I think it's cool, don't you share my view?").
A prison sentence for copyright infrigement? Are we loosing the sense of proportion here?
What about murderers? Oh right, they also get prison sentences.
Following up on myself: they are using ethanol and not water because water is much harder to splash than ethanol.
I found the answer in another article dealing with the subject.
It would be interesting to investigate how superfluids behave.
Since the article hints that the more viscosity, the lower the pressure must be to avoid splashing of the droplet, would superfluids (which have no viscosity at all) behave as expected even under the atmospheric pressure, or even a higher pressure?
Offhand, why are they using ethanol and not water for their study though?
Yes, the basic blocks are 1K blocks. That means that for large files, the whole tree is not kept, only the topmost levels (the topmost 10, I'd say, but that depends on clients).
Correction: Gnutella is using SHA1, not MD5.
Moreover, Gnutella is using an alternate hash (Tiger) to construct a hash tree, and the root of that hash tree is appended to the SHA1 of the file to make a unique "bitprint", as understood by Bitzi (http://www.bitzi.com/).
Although it *is* true that a firewall is supposed to guard you against spurious connection attempts to potentially unsecured application listening on those TCP ports, I think most P2P users who know enough simply open the TCP port of their P2P application and possibly forward it to their machine if they are behind a masquerading firewall.
Don't forget that, as a P2P user, your experience with an unfirewalled servent is much more pleasant. The same will go with UDP ports. If you run a Gnutella servent supporting UDP (such as gtk-gnutella, the forthcoming 0.95 release), you can have more results for your searches, delivered directly to you by those who have the files you're searching for.
Therefore, the step to build a reliable TCP-like stream on top of UDP was a natural one, which LimeWire took.
However, we'll see how serious the LimeWire folks are about the claimed "openess" of the Gnutella network: until they publish the specification to implement this TCP-like stream, no other Gnutella servent can support it in an inter-operable way. In effect, currently, you have to run LimeWire to benefit from this "technology". Which gives them a competitive advantage for the time beeing...
You are right that gtk-gnutella works just fine through firewalls.
However, gtk-gnutella (or any other servent) cannot establish a TCP connection between two firewalled hosts.
With the so-called FW2FW transfers, an UDP connection is made by having the two firewalled hosts "punch a hole" through the firewalls by sending datagrams to each other, each knowing the port on which the whole was opened through a third party (a "push-proxy" if you want the gory details).
Once each side can receive the datagrams of the other party, a "reliable" stream of communication is made on top of this UDP link. That's something gtk-gnutella does not support yet.
Now I can promise you that gtk-gnutella will implement this "technology" as soon as LimeWire publishes the specifications for FW2FW.
You miss the point I think. When I develop something that I give away, I don't do it for you. I do it for myself. Because it serves my interests. Then I give it away because I'm a nice guy and because it might be useful to you.
Ego only appears in the last step, when I give it away. I like to get credit for what I did, like everyone else I guess. But ego is not the motivation for doing the stuff in the first place.
Remember something: it costs to release software if you care about maintenance. I only released about 20% of the software I ever wrote for myself, because I would not find the time to maintain it anyway, or it's too specific and would need to be generalized to become useful to someone else.
Whatever Bill Gates thinks about open source software is not going to make me change my way of doing things... Even if Gates's point was proven (which I highly doubt it is), because if the stuff I wrote for me could be useful for you, the chances that you would actually buy it from me are slim, and I would not want to depend on you in any way. Which is why I just give it away in the first place, for you to (ab)use.
I see your point, and I don't dispute it. I'm just saying that software patenting has become an arm race where the system auto-feeds itself, and the more money you have, the more patents you try to submit in order to construct an IP portfolio that could save you should you start to infringe a patent and get sued for that.
In this race, the notion of good/bad is no longer appropriate. Patents are viewed as an investment.
Right, and because this is very easy to do in software (the circumvention), it means that software patents are ridiculous because they only advantage the wealthier companies.
;-) It will not be so easy to do as it is to circumvent a software patent...
By comparison, try to improve a patented chemical product, like the special glue that is used on Post-Its. Good luck
Larry Wall's "mailagent" initial implementation appeared in comp.sources.unix in 1989 IIRC, in his "dist" package (the package that contains metaconfig, the Configure script generator).
This early implementation was sending acks to people and was also processing "commands" embedded in the message. I think it pre-dates procmail.
The "mailagent" script finally grew outside of the "dist" package and became a standalone mail filter, with a nicer syntax than procmail. Search for it in CPAN if you want to compare.
Flag me as a troll if you wish, but I would not say that, because they were awarded this patent after submitting it, they are necessarily a bad company. Let me explain:
We have no idea what they are going to do with this patent. If they don't enforce it, maybe they submitted it to protect themselves against a competitor doing the same thing and then trying to enforce it?
We will be able to say that McAfee is "bad" when, and only when, they try to enforce their newly awarded patent against anyone, because only then will we know that their intent was not legitimate protection of their business but rather destruction of other ones, be they for-profit or not.
Let's not forget that this kind of broad patent can be easily circumvented. You only have to do one single improvement over the patented method, and you no longer fall under the patent. It's not the idea of doing Bayesian filterting that can be patented, it is rather HOW it is done. You patent a specific implementation. Ideas cannot be patented, in Europe (don't know about the US system, but I would say it is similar in intent).
That's why people look at prior ART to dispute a patent.
I don't know about SpamAssassin, but I use SpamBayes and I know that their algorithm involves more than just a Bayesian filter. I doubt they would fall under the patent that was just granted, or it would be really bad luck, doing the exact same things the exact same way as it is documented in the patent.