Open source is not a panacea against bad coding and failure to follow security practices Sendmail is another excellent example of an open source project with a horrible security record.
Note that a more professional re-creation of PHPNuke would not neccesarily be more secure - an open source redesign by people experienced in and familiar with security best practices would be at least as secure as a closed-source project by the guy who wrote all the flaws in PHPNuke to begin with.
I'll just sum up by saying that the problem with PHPNuke is not that its OSS, it's that it's just shitty. The kind of programmers that're attracted to quick & dirty languages like PHP don't help, either (no offense meant here, but the anal retentive security freaks that have done so much for things like BSD generally aren't that interested in web scripting languages).
Thats a great story but for every case like that theres 20 where a guy has all the drawings and patterns and whatnot and it's NOT exactly how people end up doing it - but it's close enough that he can hassle people 10 years down the road.
On top of that, if making a prototype was technically impossible, then, imo the true innovation involved is in overcoming those technical barriers - conceptualizing is important but it's better protected by trade secret and/or copyrights than by patents.
Apple redirected G5s that were originially scheduled to fulfull pre-orders to VT, resulting in (at least) 1100 people, who'd placed preorders and been promised a ship date, being out of a G5 and having to wait till a second ship (which may very well have been 6 weeks, I dunno).
It grates on people being told that they're less important than a PR move.
It's for exactly the reason you state - source code is a single implementation. Process patents protect the whole concept. Thats why shit like the Eolas patent (which contains zero, zip, nada implementation details) pisses people like me off - you're supposed to be able to recreate an invention from the patent, but it's just a list of ideas. These days it's perfectly allowable to patent things that you not only haven't done, but don't have any idea how to do and have no intention of doing (theres assloads of patents on bio and nanotech "processes" that have never been implemented).
This is a much better post than your original one:P For one thing, you avoid the generic "threading is bad" argument.
The main reason why people (at least on Linux & other unixy platforms) have avoided threading in really high end/scalable servers is because threads don't scale well. Well, now we've got a scheduler and thread libraries that scale very nearly 1:1. You're talking about designs that work around a problem that no longer exists (well, at least isn't as extreme - I haven't done the sort of testing and benchmarking I'd need to do to say that with real confidence). Theoretically, a threaded implementation will outperorm a non-threaded one - when one thread is blocking on i/o, another one can step in and do some work. The reality of that is less perfect, of course, but we're getting alot closer. Event driven design is great(I'm a big fan of this myself). I/O completion ports are great, but they're just the kernel cooperating to make your threads more efficent. Most of the architectures you point to as an alternative to threading are implemented using threads under the hood - given good thread libraries (which Linux has, now) there's no reason you can't just use threads directly and save the overhead of the framework (there may be other reasons to use the framework, but if your main concern is top-end scalability and I/O you won't want to do that).
Back when stileproject.com was amusing and sometimes interesting, before the porn sellout, it had a gallery called "linux loving sluts". It's probably still there if you google for it, although I'm at work and can't provide you with a link.
If underwear models with Tux on thier briefs isn't enough for you, I don't know what will be.
The obvious case is where someone with an agenda controls the media through which one can speak.
I consider non-governmental restrictions on speech to still be restricted. My belief is pretty simple - forums that provide uncensored and unrestricted speech (as a practical matter, this pretty much limits it to the internet) should have near-blanket immunity to claims of libel. Anonymous speech absolutely MUST be protected as well - it's imperitive for our political process if nothing else. You're 100% correct that its incredibly difficult to get a message out that isn't approved by someone in a mass media conglomerate and thats one reason why the people who own those conglomerats can be held responsible for thier content - this is basically identical to the common carrier exemptions in place today.
Thats not informative, thats just dumb. How do you think things like I/O completion are implemented? Crappy threading performance on Linux (and in unix in general) has historically been because of crappy threading libraries, and because process creation is relatively cheap so people tended to just fork children instead of spinning threads. Just because you're doing it in kernel threads instead of userspace threads doesn't mean that it's not threaded.
And I'm _really_ not sure why you're showing us a frigging Python framework as some sort of example of super-performant network programming. Pythons great and all but a performance monster it is not. "Yeah, boss, we use a runtime-compiled interperter for all our performance-critical code, but by God we avoid context switching!"
No - the immunity would only apply if you allow ALL comments without editorial or censorship. If everyone has free reign to speak, then nobody has any greater credibility. Newspapers do NOT print every letter recieved (unless they're very small) - it's a practical impossibility.
Note that I'm not talking about immunity for the people posting the comments, neccesarily - I'm talking about immunity for the people HOSTING the comments, which is where people attack.
I also think that you tend to over-estimate the credibility of an attack, particularly in a forum where the accused has an equal voice.
That said, I also don't have all that much against repealling or at least limiting our current libel and slander laws - truth should (of course) be an absoulte defense, plausible truch should be near-absolute, opinion rather than stated fact should have a very high burden of proof of both intent to and actual harm before the speech is even removed, much less damages awarded. The more free your speech, the more free your freedoms - there is no way that unrestricted speech can make people less free. Only restricted speech can do that.
Theres nothing wrong with manifest destiny, despite the bad rap it gets in history classes. Sure, it was used as an excuse for genocide and some of the most horrific and inhumane treatement of other humans in modern history (well, lets say in the top 10). That doesn't mean that the principle, which was more or less as the parent described - that we should always look outward and expand ourselves over the next horizon - is unsound or "bad" or anything else. There's nothing there that says we HAVE to kill all the people/whatever we find out there.
Now, while I'm talking about history classes I'd just like to bitch that never, not once, even when I was really into space and did alot of reading on it (at a relatively low level, to be sure, but still lots) did I read about a Russian moon landing. If you'd asked me before today if the Russian space program had ever landed ANYTHING on the moon I would have said no, and been very sure I was correct. Damn the cold war and it's annoying revisionist history nonsense.
I should point out that the only reason there haven't been manned trips inside volcanos, the bottom of the ocean, and the surface of the sun is the technological barriers to doing so. The first two are pretty far beyond what we can do now, but getting humans into the deep ocean is something that people are actively working on now, and have been for a long time.
How do we allow, for instance, a police officer to post anonymously that his superior is corrupt, in the hope of justifiably sinking the superior's career, without providing a method for crooks to anonymously sink the career of a good police officer who is a thorn in their sides?
You can't. This is why freedom of speech is important, because the ability to say things like that is neccesary. Contrary to public opinion, it's really hard to ruin someones reputation just by saying nasty things about them unless you've got a highly controlled distribution channel - websites like these don't meet that criteria. The editors of the New York Post, on the other hand, can very easily ruin (or at least mar) someones reputation because it's NOT a public forum where anyone can be heard - they don't have to provide alternative viewpoints.
Public forums that allow for the free posting of all viewpoints (like Slashdot) should be inherently protected from libel claims based on those posts.
MS will happily sell you SQL Server for that sort of market. You're supposed to cluster instead of using big iron. I imagine that this would work fine for the vast majority of even high end oracle installations (and I seriously doubt most oracle isntallations are actually that high end - certainly none of the ones I've come across). I can't speak to the price/performance ratio of clustering vrs big iron, but I would tend to assume that clustering would be cheaper but with a somewhat lower cap on absolute performance.
You've got a good point but if you're going to swallow your pride, maybe Dennys or Safeway is hiring... you've got to remember the long term too, and unless you've got alot of faith that this company is going to be around (and keep you around) for the long term, making sure that the work you're doing in your own remains yours is a good idea.
Well, you may very well argue with the logic but note that none of these actions have anything to do with the GPL - it's a politically motivated decision by the project leaders that seems to be based largely on a dislike for the APSL and apples DMCA-based shutdown of iDVD extensions.
Say I write a device driver for Linux. Thats a kernel module, so it needs to be GPLed as it links to the kernel (assume for the sake of argument that Linus never said binary modules were okay). This device driver is for some funky driver that I make, say a video card. The linux device driver is released under the GPL, as is my obligation under the license on the kernel. However, the copyright on that implementation remains my own, and I can port it to BSD and release it as a binary module under whatever license I want if I feel like it. I can even port it to Windows and package it up with a nice pretty EULA that forbids benchmarking. YES, the implementation of JFS on AIX is part of AIX, and since AIX is a derivitive of SysV they can't give it away [this is actually arguable and depends on a certain legal view of derived works that is by no means guaranteed]. That doesn't mean that JFS can't be ported to Linux (or any other operating sytem) - IBM owns the copyright on AIX.
One of the guidelines when determining copyright infringment is how essential the copied works are to the alleged infringing work - if it's a minor detail, for example, and can be removed without changing the nature of the work or (in the case of software) without unduly affecting the operation then it's much harder to claim any sort of damages or compensation for the infringment, and much easier to make a claim of fair use. The speed and ease with which the code could be replaced might very well influence the judges opinion - note that SCO attempted to address this when it listed the files, saying that the code couldn't be easily removed or altered even though the actual number of lines was small (less than 1% of the kernel even by a very generous estimate).
The 911 is reasonable - I ditched my landline for Vonage anyway, though. My reasoning was that if it's important enough to call 911, someone can run to the bar next door. Or kick in the neighbors door, or something.
The only issue I've had with Vonage has been people refusing to believe where I live because I had a non-local area code, and one issue with a snarky Dominos who wouldn't deliver for the same reason.
Theres a new breed of linkfarm thats pulluting results. Using the ever usefull "-" to filter some of them out - I've also noticed the frequency of them going down the last couple days so maybe Google has made some tweaks recently.
This isn't a hard concept - it's been spelled out repeatedly by the FSF and Stallman and basically everyone associated with the GPL.
One of the freedoms espoused (and enforced) by the GPL is the freedom to have the source code for the applications that run on your system. The GPL protects and enforces that freedom. The BSD or public domain or a wide variety of other licenses do not - they don't consider it as important. Thats fine, it's a personal belief and a personal decision to make.
Theres a further goal, which is to keep ones own code visibly in the public, and to require improvements to it to be public - that generally comes from the work ethic of the developers, who want to create a good product and want other people who work on it to contribute those improvements back. BSD/public domain/etc don't address that desire either.
I'm not the least bit interested in getting into a GPL/BSD flamewar (it's all about your definitions of free anyway, not about anything substantive), and since I'll happily ignore anyone who tells me I should release under BSD it's not going to get anywhere.
I'll note in closing that you've done an excellent job of turning a statement by the poster that had nothing to do with your pet peeve - "...GPL code should STAY free for all like it was..." into a nice troll thread. Note that releasing under a BSD license or the public domain wouldn't keep the code free as it ues when it was published under the GPL.
Slashdot Mirror
Note that a more professional re-creation of PHPNuke would not neccesarily be more secure - an open source redesign by people experienced in and familiar with security best practices would be at least as secure as a closed-source project by the guy who wrote all the flaws in PHPNuke to begin with.
I'll just sum up by saying that the problem with PHPNuke is not that its OSS, it's that it's just shitty. The kind of programmers that're attracted to quick & dirty languages like PHP don't help, either (no offense meant here, but the anal retentive security freaks that have done so much for things like BSD generally aren't that interested in web scripting languages).
On top of that, if making a prototype was technically impossible, then, imo the true innovation involved is in overcoming those technical barriers - conceptualizing is important but it's better protected by trade secret and/or copyrights than by patents.
It grates on people being told that they're less important than a PR move.
It's for exactly the reason you state - source code is a single implementation. Process patents protect the whole concept. Thats why shit like the Eolas patent (which contains zero, zip, nada implementation details) pisses people like me off - you're supposed to be able to recreate an invention from the patent, but it's just a list of ideas. These days it's perfectly allowable to patent things that you not only haven't done, but don't have any idea how to do and have no intention of doing (theres assloads of patents on bio and nanotech "processes" that have never been implemented).
The main reason why people (at least on Linux & other unixy platforms) have avoided threading in really high end/scalable servers is because threads don't scale well. Well, now we've got a scheduler and thread libraries that scale very nearly 1:1. You're talking about designs that work around a problem that no longer exists (well, at least isn't as extreme - I haven't done the sort of testing and benchmarking I'd need to do to say that with real confidence). Theoretically, a threaded implementation will outperorm a non-threaded one - when one thread is blocking on i/o, another one can step in and do some work. The reality of that is less perfect, of course, but we're getting alot closer. Event driven design is great(I'm a big fan of this myself). I/O completion ports are great, but they're just the kernel cooperating to make your threads more efficent. Most of the architectures you point to as an alternative to threading are implemented using threads under the hood - given good thread libraries (which Linux has, now) there's no reason you can't just use threads directly and save the overhead of the framework (there may be other reasons to use the framework, but if your main concern is top-end scalability and I/O you won't want to do that).
If underwear models with Tux on thier briefs isn't enough for you, I don't know what will be.
*sob* She's going to use IE when she grows up!
It's not especially hard to roll your own rich text editor entirely in DHTML using only standards compliant DOM. It'll work in both IE and moz.
I consider non-governmental restrictions on speech to still be restricted. My belief is pretty simple - forums that provide uncensored and unrestricted speech (as a practical matter, this pretty much limits it to the internet) should have near-blanket immunity to claims of libel. Anonymous speech absolutely MUST be protected as well - it's imperitive for our political process if nothing else. You're 100% correct that its incredibly difficult to get a message out that isn't approved by someone in a mass media conglomerate and thats one reason why the people who own those conglomerats can be held responsible for thier content - this is basically identical to the common carrier exemptions in place today.
And I'm _really_ not sure why you're showing us a frigging Python framework as some sort of example of super-performant network programming. Pythons great and all but a performance monster it is not. "Yeah, boss, we use a runtime-compiled interperter for all our performance-critical code, but by God we avoid context switching!"
Note that I'm not talking about immunity for the people posting the comments, neccesarily - I'm talking about immunity for the people HOSTING the comments, which is where people attack.
I also think that you tend to over-estimate the credibility of an attack, particularly in a forum where the accused has an equal voice.
That said, I also don't have all that much against repealling or at least limiting our current libel and slander laws - truth should (of course) be an absoulte defense, plausible truch should be near-absolute, opinion rather than stated fact should have a very high burden of proof of both intent to and actual harm before the speech is even removed, much less damages awarded. The more free your speech, the more free your freedoms - there is no way that unrestricted speech can make people less free. Only restricted speech can do that.
Now, while I'm talking about history classes I'd just like to bitch that never, not once, even when I was really into space and did alot of reading on it (at a relatively low level, to be sure, but still lots) did I read about a Russian moon landing. If you'd asked me before today if the Russian space program had ever landed ANYTHING on the moon I would have said no, and been very sure I was correct. Damn the cold war and it's annoying revisionist history nonsense.
I should point out that the only reason there haven't been manned trips inside volcanos, the bottom of the ocean, and the surface of the sun is the technological barriers to doing so. The first two are pretty far beyond what we can do now, but getting humans into the deep ocean is something that people are actively working on now, and have been for a long time.
You can't. This is why freedom of speech is important, because the ability to say things like that is neccesary. Contrary to public opinion, it's really hard to ruin someones reputation just by saying nasty things about them unless you've got a highly controlled distribution channel - websites like these don't meet that criteria. The editors of the New York Post, on the other hand, can very easily ruin (or at least mar) someones reputation because it's NOT a public forum where anyone can be heard - they don't have to provide alternative viewpoints.
Public forums that allow for the free posting of all viewpoints (like Slashdot) should be inherently protected from libel claims based on those posts.
MS will happily sell you SQL Server for that sort of market. You're supposed to cluster instead of using big iron. I imagine that this would work fine for the vast majority of even high end oracle installations (and I seriously doubt most oracle isntallations are actually that high end - certainly none of the ones I've come across). I can't speak to the price/performance ratio of clustering vrs big iron, but I would tend to assume that clustering would be cheaper but with a somewhat lower cap on absolute performance.
You've got a good point but if you're going to swallow your pride, maybe Dennys or Safeway is hiring... you've got to remember the long term too, and unless you've got alot of faith that this company is going to be around (and keep you around) for the long term, making sure that the work you're doing in your own remains yours is a good idea.
Well, you may very well argue with the logic but note that none of these actions have anything to do with the GPL - it's a politically motivated decision by the project leaders that seems to be based largely on a dislike for the APSL and apples DMCA-based shutdown of iDVD extensions.
Say I write a device driver for Linux. Thats a kernel module, so it needs to be GPLed as it links to the kernel (assume for the sake of argument that Linus never said binary modules were okay). This device driver is for some funky driver that I make, say a video card. The linux device driver is released under the GPL, as is my obligation under the license on the kernel. However, the copyright on that implementation remains my own, and I can port it to BSD and release it as a binary module under whatever license I want if I feel like it. I can even port it to Windows and package it up with a nice pretty EULA that forbids benchmarking. YES, the implementation of JFS on AIX is part of AIX, and since AIX is a derivitive of SysV they can't give it away [this is actually arguable and depends on a certain legal view of derived works that is by no means guaranteed]. That doesn't mean that JFS can't be ported to Linux (or any other operating sytem) - IBM owns the copyright on AIX.
One of the guidelines when determining copyright infringment is how essential the copied works are to the alleged infringing work - if it's a minor detail, for example, and can be removed without changing the nature of the work or (in the case of software) without unduly affecting the operation then it's much harder to claim any sort of damages or compensation for the infringment, and much easier to make a claim of fair use. The speed and ease with which the code could be replaced might very well influence the judges opinion - note that SCO attempted to address this when it listed the files, saying that the code couldn't be easily removed or altered even though the actual number of lines was small (less than 1% of the kernel even by a very generous estimate).
X is roughly 2000. Y is roughly 2,000,000. Therefore, the kernel is worth 699 * 1000 = roughly 700 thousand dollars.
You forgot 401 "Authentication Required"
The only issue I've had with Vonage has been people refusing to believe where I live because I had a non-local area code, and one issue with a snarky Dominos who wouldn't deliver for the same reason.
I do not, and never have, gotten a dialtone without paying for service on a landline.
Theres a new breed of linkfarm thats pulluting results. Using the ever usefull "-" to filter some of them out - I've also noticed the frequency of them going down the last couple days so maybe Google has made some tweaks recently.
Maybe he should have just advertised with Google the correct way instead of paying linkfarm slimeballs to artificially up his rating. Fuck him, I say.
One of the freedoms espoused (and enforced) by the GPL is the freedom to have the source code for the applications that run on your system. The GPL protects and enforces that freedom. The BSD or public domain or a wide variety of other licenses do not - they don't consider it as important. Thats fine, it's a personal belief and a personal decision to make.
Theres a further goal, which is to keep ones own code visibly in the public, and to require improvements to it to be public - that generally comes from the work ethic of the developers, who want to create a good product and want other people who work on it to contribute those improvements back. BSD/public domain/etc don't address that desire either.
I'm not the least bit interested in getting into a GPL/BSD flamewar (it's all about your definitions of free anyway, not about anything substantive), and since I'll happily ignore anyone who tells me I should release under BSD it's not going to get anywhere.
I'll note in closing that you've done an excellent job of turning a statement by the poster that had nothing to do with your pet peeve - "...GPL code should STAY free for all like it was..." into a nice troll thread. Note that releasing under a BSD license or the public domain wouldn't keep the code free as it ues when it was published under the GPL.