Home systems (like mine) DO need bind. I can cache lookups here and browse quickly, or wait forever for my @home name server to respond. BIG difference.
But you don't need BIND for this.
Check out dnscache which is part of the djbdns package.
Notice that the "server" and "workstation" configurations themselves are Microsoft-isms.:-)
No they're not.
Long before Microsoft entered into the scene with NT, Vendors such as Sun were selling UNIX servers and workstations. True, this mostly referred to hardware configurations rather that OS configurations, but that was simply reflective of the fact that they were hardware vendors rather than software vendors.
That's the first (and hopefully only) time I ever hope to see the words "elegant", "minimalistic", and "Sendmail" together in the same sentence.
I guess I wasn't clear that Sendmail and BIND weren't minimalistic, and that qmail and djbdns were (at least by comparison), and therefore more secure.
I think that the real reason that MS systems were hit so hard by Code Red and it's descendents is that there is a real difference in the culture of the respective developer communities.
There is no reason why all those home systems and corporate desktops should have IIS running in the first place. There is also no reason (generally) for a home linux system to be running, say, BIND or wu-ftpd.
So why does Microsoft encourage the installation of unneccessary software on it's systems, and why doesn't it make it easier to not install those services in the first place?
It comes down to culture. Unix-like operating systems are minimalist and modular, because the development communities appreciate elegant code (not neccessarily elegant interfaces).
Whereas Microsoft prizes a DWIM (Do What I Mean) approach, which encourages adding functionality 'just-in-case', as Microsoft seems to think that actually asking a user to install a component is a failure on their part.
In the long run, elegant, minimalistic code is easier to understand, and therefore easier to secure (examples are Sendmail vs. qmail, or BIND vs. djbdns).
TRW Swap Meet is held the last Saturday of each month on the northwest corner of Marine Ave. (Compton Blvd.) and Aviation Blvd. in Manhattan Beach. 7 A.M. to 11:30 A.M.
As another example to manufacturing scarcity (albeit an entirely legal one), consider Ty Beanie Babies: This is a company that developed a specific expertise in creating an artificial scarcity, and making sure that supply was always just a step behind demand.
For a more sinister example, consider the coal mining and oil industries which have been funding the anti-nuclear movement behind the scenes for years, and scuttling space-based power sattelite plans to maintain the energy scarcity that keeps people using fossil fuels.
Or General Motors, which bought up the Los Angeles Red-Line trolley system, only in order to dismantle it, thereby creating a transportation scarcity in Los Angeles, which their cars helped fill.
I'm sure you can think of other examples, both historical and recent.
why not say this: "If you modify the program and distribute either the program or its output, then you must provide your modifications to all recipients of the modified program or its output."
The problem with attaching the requirement to the program's output, is that it could 'contaminate' what would otherwise essentially be a static data dump.
This is exactly why the GPL was modified with respect to gcc and Bison to allow the output (even though it was in part derived from the software itself) to be exempted from the GPL.
The line must be drawn somewhere in the vicinity of exposing the functionality of the software interactively. 'Public performance of the software' is the best attempt at tying this into existing case law that I'm aware of.
What I'd like is for exaqmple a modified gcc that was exposed as a compilation service to be required to distribute it's modifications, but the compiled output (the result of using the service) to still be exempted. And if the modified gcc was being used internally as a service, then the resulting compiled binaries should still be able to be distributed publicly, without the gcc modifications needing to be distributed.
This case reminds me of the 1994 case US vs. Thomas, where a California couple operated a BBS (called Amatuer Action) whose content was legal within CA, but a DA in Tennessee dialed up from within his jurisdiction, declared that they had violated his local community standards, and had them picked up in California, and transported to Tennessee, where they were convicted of 11 counts of obscenity. They lost their appeal.
As usual, I should point out that if the root were run properly, allowing any TLD to be added, this squabbling over an artificially-limited resource would be eliminated.
Indeed.
It is becomming increasingly apparent to me that as we move from a scarcity economy to one of abundance, attention is shifting from control of scarce resources to control of the means of creating scarcity.
In other words, in an abundance economy, the only thing that is scarce is scarcity itself.
Therefore, ICANN can be viewed as nothing more than a tool for manufacturing and maintaining scarcity, and after that scarcity has been created, a tool for controlling it.
[stuff about problems with forking the license deleted]
I guess this is why it's taking so long...:-)
Maybe I'm the only one who thinks this is a bad idea. But the thought that merely setting up a listen socket, and having someone connect to that socket can force me to give them the code is scary.
I think that this is an overreaction. There is still a difference between a public performance and a private one. If you set up a reasonable security measure that has to be circumvented in order to gain access to your service, or even if your service had no security but was never publicly announced, then you most likely aren't publicly performing the software, and whoever circumvented your security (or searched for the port at random) would not qualify for any consideration regardi[0our source. But then again, IANAL.
How do you make an optional clause out of this? Is it that the author or authors get to pick if they want this requirement or not? If it's never possible to require this extra feature because it's truly optional, then why bother? If it isn't optional, you fork the license into GPL and NGPL (network GPL) and you have to pick one of them. Based on how the license would have to work, GPL + NGPL = NGPL, so eventually everything could get drawn into NGPL. One of the reasons that the GPL has worked so well is that there's only one of it, so having two licenses cannot be good. This would not be a trivial fork of the license. The FSF has argued over much smaller points before, so I doubt that the clause can be made optional.
I'm not so sure that you're right in this regard. Not all LGPL'd software gets sucked into the GPL. However, IANAL, so I'm perfectly willing to say that I dunno.
Hmm. Actually, to take your argument to the extreme, wouldn't there have to be an NLGPL as well? I guess that's four licenses that could be combined in 16 different ways...
If the GPLv3 has this clause in it, then it cannot be a Free Software license, because the definition for Free Software says "the freedom to run the program, for any purpose". But a restriction on public performance is a restriction on running the program.
Ah, I see where the misunderstanding is coming from. I apologise for not being more clear in the first place.
Closing the 'web-app loophole' doesn't involve disallowing public performances at all. It simply means that if you *do* publicly perform a modified version of the software (that is, you make it available as a service), then you must provide the source to your modifications, just as if you had distributed a copy of the modified software.
I simply meant that publicly performing the modified software without making the changes available would be disallowed.
The purpose of Free Software is to give freedom to the user, not a big stick to the developer. I fail to see how a user has less freedom when they excercise their free will by freely choosing to use a web application. If they freely choose to rely on someone else's copy of the software, then that is their business, and not yours.
The GPL and other copylefted free software licenses aim to ensure user freedoms by restricting developer's ability to restrict users' freedoms.
The GPL does not restrict how one may use the software at all. If I release code under the GPL, then I do not want users' freedoms restricted in regards to derivative redistributions of my code. I also do not want users' freedoms restricted in regards to derivative public performances of my code.
I do not want a big stick. I specifically want to reduce the size of other developers sticks when they are using them to beat users away from code derived from my code.
Limiting public performance goes against the very nature of Free Software. Please reread the Free Software definition and then explain how restricting the right of the user to perform his own private modifications of the software can possibly fit that definition.
Well, basically what I (and others) are calling for is an optional restriction on public performances of private modificactions, but no restriction on private performances of private modifications.
Thus, private modifications that a company made for their own internal use would be treated no differently than now, but exposing a service to the world in general would constitute a public performance, and require that the modified source code be made available.
If such an option is not made available to developers seeking to create Free-as-in-speech implementations of key web services, then users who increasingly rely on web services will face a marked decline in their Freedom, as proprietary interests deploy non-free versions of the code.
So, are you saying that if I write a MUD under the GPL, and someone telnets to it, I will have to give them the sourcecode for my MUD?
If you don't intend on making the source code available, what is the point of using the GPL?
Let's say that you write an interesting codebase for creating MUD games, and you release it under the GPL. obviously, your intention is to have derivative versions of the code remain Free, so that no-one is locked in to a proprietary vendor.
However, an unscrupulous company can set up a proprietary service that uses your code without having to distribute their modifications, since they don't actually distribute their modifications in binary form either. They merely expose a service. Result: Reduced freedom for users. This is very similar in result to a BSD license.
Now, since the GPL allows private modifications with no requirement to distribute the modifications (for a companies internal use, say), then I think a good extension of this principle would be to allow 'private performances' of modified code, with no requirement to distribute modifications.
Also, I have a feeling that any 'public performance' distribution requirement will be an optional clause in the GPL, rather than an across-the-board requirement.
As both Bruce Perens and Tim O'Reilly have pointed out, it is possible to publicly deploy a web-app that is derived from GPL'd software without having to distribute your modifications.
While I certainly feel that it should be possible to do this for applications that are deployed internally without having the deployment count as 'distribution', I am less happy about deployments on public websites. I would want web-applications that I create to have an additional 'public-performance' clause in their license that would require modifications that are publicly deployed to be made available in source form.
This is the so called 'web-app loophole', and I was wondering what your thoughts on the matter were?
Why Kodak? because they were always about bringing the ability to create to the average joe. I don't think that it's too much of a strech to imagine them extending this philosophy to multimedia as well, especially after Microsoft started screwing them.
Imagine BeOS based kiosks, digital cameras, digital videocameras... Not to mention a BeOS based set-top box that shows all those pictures and video clips...
While making the network 'smarter' would enable businesses to control 'quality of service' in a more direct way, this will also fossilize the network just as it is today, and eliminate future development of innovative uses for it.
Many large businesses have been using dedicated connections to EDI providers such as AT&T which has always used a 'smarter' store-and-forward protocol. The result, has been a charge-per-transaction revenue model that has locked these businesses in over a very long period of time.
Rather than trying to turn the Internet into a 'smarter' network, I suggest that efforts should focus on building a smarter network in parallel to the Internet, using the same wires, routers, firewalls, etc. That way, when the new network fails to provide some needed flexibility (and it will), developers can fall back on the flexible-but-dumb internet. This will also ensure an upgrade path exists when the smart-but-brittle network needs to be replaced with something else.
Call this the 'pluggable protocol' model. It also allows for multiple special purpose protocols to operate at the same time. There could be a protocol designed specifically for MMORPGs, for example, that was optimized for minimizing latency.
The alternatives for future development otherwise are rather poor. Imagine trying to build a peer-to-peer network over EDI. Not pretty.
Contracts, licenses, and other legal documents do this all the time. Terms can be redefined as neccessary within the context of the document.
There is nothing sinister about this, it simply saves the trouble of having to use a made up word and defining that, instead.
Let me stress again, that if the terms of the GPL are violated, then the license terminates, and you only have those rights that you get under ordinary copyright law. The redefinition of the term 'derive' does not affect the definition WRT copyright law. If you were doing anything that would not be permitted under copyright law (such as redistribution), then you cannot continue to do so, since your license was terminated.
This is not a bug, it is a feature, and without it the GPL would not be enforceable.
If you are doing things that would be permitted under copyright, then you have no problems, even though your license under the GPL has terminated.
Remember, if you don't agree with the GPL's terms, conditions, and definitions, then you don't have to accept the license in the first place. --
Slashdot Mirror
Check out dnscache which is part of the djbdns package.
Long before Microsoft entered into the scene with NT, Vendors such as Sun were selling UNIX servers and workstations. True, this mostly referred to hardware configurations rather that OS configurations, but that was simply reflective of the fact that they were hardware vendors rather than software vendors.
I think that the real reason that MS systems were hit so hard by Code Red and it's descendents is that there is a real difference in the culture of the respective developer communities.
There is no reason why all those home systems and corporate desktops should have IIS running in the first place. There is also no reason (generally) for a home linux system to be running, say, BIND or wu-ftpd.
So why does Microsoft encourage the installation of unneccessary software on it's systems, and why doesn't it make it easier to not install those services in the first place?
It comes down to culture. Unix-like operating systems are minimalist and modular, because the development communities appreciate elegant code (not neccessarily elegant interfaces).
Whereas Microsoft prizes a DWIM (Do What I Mean) approach, which encourages adding functionality 'just-in-case', as Microsoft seems to think that actually asking a user to install a component is a failure on their part.
In the long run, elegant, minimalistic code is easier to understand, and therefore easier to secure (examples are Sendmail vs. qmail, or BIND vs. djbdns).
Microsodt has published a transcript of the panel discussion commemorating the 20th anniversary.
...But some are more Equal than others.
- George Orwell, Animal Farm.
TRW Swap Meet is held the last Saturday of each month on the northwest corner of Marine Ave. (Compton Blvd.) and Aviation Blvd. in Manhattan Beach. 7 A.M. to 11:30 A.M.
Buyers free, sellers $5.00.
As another example to manufacturing scarcity (albeit an entirely legal one), consider Ty Beanie Babies: This is a company that developed a specific expertise in creating an artificial scarcity, and making sure that supply was always just a step behind demand.
For a more sinister example, consider the coal mining and oil industries which have been funding the anti-nuclear movement behind the scenes for years, and scuttling space-based power sattelite plans to maintain the energy scarcity that keeps people using fossil fuels.
Or General Motors, which bought up the Los Angeles Red-Line trolley system, only in order to dismantle it, thereby creating a transportation scarcity in Los Angeles, which their cars helped fill.
I'm sure you can think of other examples, both historical and recent.
This is exactly why the GPL was modified with respect to gcc and Bison to allow the output (even though it was in part derived from the software itself) to be exempted from the GPL.
The line must be drawn somewhere in the vicinity of exposing the functionality of the software interactively. 'Public performance of the software' is the best attempt at tying this into existing case law that I'm aware of.
What I'd like is for exaqmple a modified gcc that was exposed as a compilation service to be required to distribute it's modifications, but the compiled output (the result of using the service) to still be exempted. And if the modified gcc was being used internally as a service, then the resulting compiled binaries should still be able to be distributed publicly, without the gcc modifications needing to be distributed.
This case reminds me of the 1994 case US vs. Thomas, where a California couple operated a BBS (called Amatuer Action) whose content was legal within CA, but a DA in Tennessee dialed up from within his jurisdiction, declared that they had violated his local community standards, and had them picked up in California, and transported to Tennessee, where they were convicted of 11 counts of obscenity. They lost their appeal.
Does anyone know what happnned after that?
...but someone has finally re-discovered the principle behinfd the repelatron!
Tom Swift Jr. invented all the cool stuff.
It is becomming increasingly apparent to me that as we move from a scarcity economy to one of abundance, attention is shifting from control of scarce resources to control of the means of creating scarcity.
In other words, in an abundance economy, the only thing that is scarce is scarcity itself.
Therefore, ICANN can be viewed as nothing more than a tool for manufacturing and maintaining scarcity, and after that scarcity has been created, a tool for controlling it.
Well, the fact that someone could take my GPL'd web-service code and turn it into a proprietary web-service is scary to me.
If I wasn't concerned about proprietary free-riders, then I'd be using a BSD style license, and not the GPL.
I guess this is why it's taking so long...
Hmm. Actually, to take your argument to the extreme, wouldn't there have to be an NLGPL as well? I guess that's four licenses that could be combined in 16 different ways...
I'm still saying: "I dunno".
Closing the 'web-app loophole' doesn't involve disallowing public performances at all. It simply means that if you *do* publicly perform a modified version of the software (that is, you make it available as a service), then you must provide the source to your modifications, just as if you had distributed a copy of the modified software.
I simply meant that publicly performing the modified software without making the changes available would be disallowed.
Again, my apologies for my imprecision.
The GPL does not restrict how one may use the software at all. If I release code under the GPL, then I do not want users' freedoms restricted in regards to derivative redistributions of my code. I also do not want users' freedoms restricted in regards to derivative public performances of my code.
I do not want a big stick. I specifically want to reduce the size of other developers sticks when they are using them to beat users away from code derived from my code.
This is all about Freedom for the user.
Thus, private modifications that a company made for their own internal use would be treated no differently than now, but exposing a service to the world in general would constitute a public performance, and require that the modified source code be made available.
If such an option is not made available to developers seeking to create Free-as-in-speech implementations of key web services, then users who increasingly rely on web services will face a marked decline in their Freedom, as proprietary interests deploy non-free versions of the code.
Let's say that you write an interesting codebase for creating MUD games, and you release it under the GPL. obviously, your intention is to have derivative versions of the code remain Free, so that no-one is locked in to a proprietary vendor.
However, an unscrupulous company can set up a proprietary service that uses your code without having to distribute their modifications, since they don't actually distribute their modifications in binary form either. They merely expose a service. Result: Reduced freedom for users. This is very similar in result to a BSD license.
Now, since the GPL allows private modifications with no requirement to distribute the modifications (for a companies internal use, say), then I think a good extension of this principle would be to allow 'private performances' of modified code, with no requirement to distribute modifications.
Also, I have a feeling that any 'public performance' distribution requirement will be an optional clause in the GPL, rather than an across-the-board requirement.
As both Bruce Perens and Tim O'Reilly have pointed out, it is possible to publicly deploy a web-app that is derived from GPL'd software without having to distribute your modifications.
While I certainly feel that it should be possible to do this for applications that are deployed internally without having the deployment count as 'distribution', I am less happy about deployments on public websites. I would want web-applications that I create to have an additional 'public-performance' clause in their license that would require modifications that are publicly deployed to be made available in source form.
This is the so called 'web-app loophole', and I was wondering what your thoughts on the matter were?
is that it's Kodak.
Why Kodak? because they were always about bringing the ability to create to the average joe. I don't think that it's too much of a strech to imagine them extending this philosophy to multimedia as well, especially after Microsoft started screwing them.
Imagine BeOS based kiosks, digital cameras, digital videocameras... Not to mention a BeOS based set-top box that shows all those pictures and video clips...
The possibilities are interesting.
While making the network 'smarter' would enable businesses to control 'quality of service' in a more direct way, this will also fossilize the network just as it is today, and eliminate future development of innovative uses for it.
Many large businesses have been using dedicated connections to EDI providers such as AT&T which has always used a 'smarter' store-and-forward protocol. The result, has been a charge-per-transaction revenue model that has locked these businesses in over a very long period of time.
Rather than trying to turn the Internet into a 'smarter' network, I suggest that efforts should focus on building a smarter network in parallel to the Internet, using the same wires, routers, firewalls, etc. That way, when the new network fails to provide some needed flexibility (and it will), developers can fall back on the flexible-but-dumb internet. This will also ensure an upgrade path exists when the smart-but-brittle network needs to be replaced with something else.
Call this the 'pluggable protocol' model. It also allows for multiple special purpose protocols to operate at the same time. There could be a protocol designed specifically for MMORPGs, for example, that was optimized for minimizing latency.
The alternatives for future development otherwise are rather poor. Imagine trying to build a peer-to-peer network over EDI. Not pretty.
There is nothing sinister about this, it simply saves the trouble of having to use a made up word and defining that, instead.
Let me stress again, that if the terms of the GPL are violated, then the license terminates, and you only have those rights that you get under ordinary copyright law. The redefinition of the term 'derive' does not affect the definition WRT copyright law. If you were doing anything that would not be permitted under copyright law (such as redistribution), then you cannot continue to do so, since your license was terminated.
This is not a bug, it is a feature, and without it the GPL would not be enforceable.
If you are doing things that would be permitted under copyright, then you have no problems, even though your license under the GPL has terminated.
Remember, if you don't agree with the GPL's terms, conditions, and definitions, then you don't have to accept the license in the first place.
--