The point is, people wanted the "official" guaranteed and tested updates without getting phone and email support, which Redhat are dropping from their range.
Fedora is now the "officially" tested and sanctioned free release made by the company known as Red Hat. The only thing that has changed is that you cannot pay Red Hat for phone support now, but that is no loss for Red Hat as they were losing money on that anyway.
Furthermore, the Fedora packages form the basis of the for-pay Red Hat Enterprise Linux system, and the improvements Red Hat funds for them go back, via the GPL, into Fedora.
The problem is that with Fedora and the yum and apt repositories the maintainer of a package could be on holiday, be busy, just be plain bored, and not update one of the packages he maintains.
As Fedora is sponsored and to some extent driven by Red Hat, this is a non-issue for most of the packages (virtually all of the major ones) in Fedora.
8th Affirmative Defense - The GPL violates the U.S. Constitution, together with copyright, antitrust and export control laws, and IBM's claims based theron, or related thereto, are barred.
Actually, being involved in litigation several times myself, I have come to realize that a standard defensive tactic lawyers use is to bring out every conceivable defense in their responses, no matter how ludicrous. For example, in my defense against a simple traffic violation, my lawyer entered in the pleadings several defenses along the lines of "Section X of Motor Vehicle Code violates the US Constitution" and "Section X of Motor Vehicle Code is selectively enforced by Police such that its enforcement is waived, estopped or otherwise barred as a matter of equity". The lawyer knew that the chances of a judge accepting those defenses were about 1:2^128, but he used them anyway.
So these clauses look like standard litigation boilerplate to me.
This is probably the best "simple" explanation of why the recent American practice of attaching "Terms and Conditions" to everything is wrong. The post author does not assert that it is wrong based on some vaguely defined morals or ethics, but that it is wrong based on plain old economic principles.
Certainly Nvidia's legal department believes that their drivers fall on the legal side of the fence, and I for one am glad, as I enjoy the Nvidia drivers on my GeForce4;-)
However, that's not to say those drivers don't inhabit a grey area of GPL law. While the drivers contain no Linux code per se, I believe they do make use of interfaces of the Linux kernel which are neither published nor intended for normal use. Therefore, it could be said that the code that makes up those drivers is a derivative work of the Linux kernel, as Nvidia has to study kernel source code carefully and adapt their code so as to be able to link directly into it. I suppose they could do some cleanroom reverse engineering to get around that, or something.
But I am not an IP attorney and even if I were, none of the Linux developers seem to be concerned about Nvidia's kernel drivers.
The linux kernel is GPL'd with the added provision that you are allowed to write a binary moduled to be loaded at runtime without having to release your code under the GPL.
Interesting. Where in the kernel distribution does it explicitly state this? I downloaded 2.4.22 from kernel.org and looked in the COPYING file but found nothing authorizing this. Rather, the COPYING simply state that the kernel is released under the terms of GPLv2. I also looked on some of the main pages of kernel.org and could not find anything.
You are probably referring to Linus's interpretation of the GPL that he posted to the lkml. However, as he is only one of many kernel authors, he doesn't have the authority necessary to make interpretations which relax the license for the kernel. He would need explicit permission from every kernel author to make those exemptions to the license. Some of the module owners are firmly in the Free Software camp, so he would probably have to remove or rewrite some parts of the kernel, as they would be unlikely to acquiesce.
If he got all of the permission necessary, he would need to add the information about the deviation from vanilla GPLv2 to the COPYING file. Thus far, I am not aware of his gathering permission from everyone, and he certainly has not changed the COPYING file. Therefore, to the best of my knowledge gathered from the research I have done, the kernel is still technically licensed under the vanilla GPLv2, which means that dynamic, or runtime, linking of a module still requires GPLing the module.
Now, given Linus's stance, a company might feel safe in making a binary-only runtime module. (*cough* Nvidia) However, technically that is still in violation of the license that is attached to current kernels, and it could potentially open the company to legal action in the future, as Linus is free to change his mind, and any of the other developers could also initiate the complaint.
I don't mean to be redundant, but I see this mistake all the time and I don't want people to go away with incorrect ideas. You need to read your licenses carefully, especially if you write or modify code under them.
The Linux kernel is GPLed, not LGPLed. You can link LGPLed code with new code dynamically without releasing source to the new code. However, with the GPL, you *must* release source regardless of whether the code is statically or dynamically linked. So even if Linksys dynamically linked their proprietary modules with the kernel, it would still be a GPL violation.
The only way to avoid distributing source code that works with GPLed code is to write a GPL-compatibly-licensed wrapper program for the GPLed code. Then you write the proprietary program which uses IPC or some other arms-length method of talking to the wrapper.
Is it just me, or does that sound like a good script for a new TV commercial? I am thinking something like the IBM commercials, but IBM doesn't want to antagonize Microsoft. Perhaps a Novell or Sun commercial to go with their new desktop offerings?
The MS proponent can be a stereotypical stressed-out, disheveled, annoying person, and the Linux proponent can be calm, cool, and collected. Just substitute reference to "Peoplesoft" with the sponsor company's name. Rather than "So are we" substitute "We're bigger". Bigger is always better.;-) And rather than the last line being "No, it doesn't" substitute "Not anymore!"
apt-get and Synaptic are absolutely fabulous apps for those of us who are experienced using Linux and installing software for it. However, they fail to help newbies because the process they use to get and install software does not map to the way the newbie thinks about installing software. And anyone who says "The way the newbie thinks is wrong" is simply failing to understand newbies at all.
Now that I've made a generalization, let's have a specific run-through of the problem. I have first-hand experience with switching (some successfully, some not) a number of Windows users to Linux, and here is the problem that they all run into. (NOTE: I ran into this very same problem when I first switched, but I knew no one who could hold my hand through it all. The only reasons I am still using Linux are that I am far more intelligent than most people, so I am better at figuring things out on my own, and I am also incredibly stubborn when it comes to learning something new.)
You are Joe Newbie. You've got your nice shiny Linux system running. You hear about a great app called "FooBar". You like what you hear about it and you decide you want to try it. You search google for it, and go to www.foobar-software.org. You try to download it. But you can only get source or an RPM or DEB package. (Let's assume your friend set you up with Libranet and for some reason you actually know that it's a Debian based distro -- a stretch in its own right) You download the DEB, but you run into dependency conflicts. It wouldn't be so bad if this happened once or twice, but it happens for bloody near every app you try to install.
Now all seasoned Debian users, as well as most users of other distros, will be screaming at you to use apt-get or synaptic, or whatever other package management system. But that's the problem. People accustomed to Windows or Mac are accustomed to going to a store or a vendor's web site and getting the software they want. They have no idea that their computer might somehow "magically" know how to get it for them. Heck, if you didn't know better, why would you think your system would know how to do that? Even when you show them how apt-get works, they still often forget and revert back to the old way. It's a very deeply ingrained habit that only the most persistent learn to break.
To make things worse, even Debian, with a repository probably more exhaustive than any other distro's, still doesn't have all of the packages (and new versions of packages) that a user wants. If that user is a newbie, having to remember multiple methods for acquiring software and knowing when to use each is a further strain.
Granted, if a person sticks with Linux and becomes more accustomed to it, he or she will probably learn how to use apt-get (or insert package management system here) to streamline the package-acquisition process. However, it would be in open source's best interests to try to minimize culture shock so as to further help bring more people in.
So the grandparent poster was dead-on. For example, Mozilla really does get it. Their installer is distro-agnostic and installs everything that the package needs. Even though this may introduce more bloat (redundant packages) for any given distro, it also results in an easy installation for a newbie. Advanced users will know how to get better versions of Mozilla specifically for their distros, but newbies will still be able to participate by getting Mozilla the way they know best.
All end-user focused software packages should follow Mozilla's example of providing a simple executable installer which contains all libraries and files needed to run the software independently of most, if not all, other packages on the system. This certainly isn't the ideal setup from a sysadmin or advanced user standpoint, but it is needed to match the way newbies think about installing software.
Actually yes, telling stories requires effort. If I want to tell a good story I have to get the story in order, get a good audience, set the mood, etc. It takes work..... I find "sharing" to mean that you are going out of your way to give something to someone else.
Even by your twisted logic, I am sharing using P2P. I am sacrificing my bandwidth, for which I pay some sum of money, for someone else's benefit. Regardless of whether I download any files in return, I am giving my upload speed to other people and they don't necessarily have to give anything back to me.
Not that I condone widespread copyright violation. I do wish people would share more copylefted stuff instead of pop crap. However, proles will be proles. But I think that if you are going to blast them, you should at least have a logically consistent viewpoint.
The simple answer is that if you are developing commercial software, you should not be targetting the Fedora distribution. Red Hat and the Fedora project clearly intend this to be like the equivalent of Debian unstable/testing. How many sane commercial software developers would target Debian unstable/testing?
If you want to target an "official" Red Hat distribution with guaranteed stability, Red Hat clearly wants you to target their Enterprise Linux. If you don't want to or cannot afford it, then I would say it's time to stop targetting Red Hat distributions at all, and go with Suse Enterprise, Debian stable, or some other distribution which better meets your needs.
If you really, really want to target Fedora, your best bet would be to bundle all required libraries into one big package, or statically compiling everything, assuming you legally can according to the licenses of the components (ala the Kompany). You might not like the idea of putting all of that "bloat" into your nice, clean little app, but that is what will be necessary to guarantee the level of package-stability you seem to require.
Whether or not you like this new strategery, this is the tack Red Hat appears to be taking. Now you need to decide whether you should keep fishing or move on to some other waters.
At least in mozilla and the mozilla based browsers, you can load a url by pasting it (middle clicking) anywhere on an open page or empty tab.
WARNING: The so-called feature of Mozilla that you are espousing is actually a very old bug (#96972). It was never supposed to be there, and there has been a long-standing discussion about whether to remove it. Many people have filed bug reports and/or requested/voted that it be removed. However, the bug has just enough very vocal supporters among the hacker community to block its removal. Regardless of the debate, this "feature" is classified as a bug, so it is dangerous to just assume it will always be there, since it may one day be stripped out.
Of course, I am biased toward the pro-removal camp. As a software engineer and UI designer, I believe in strictly following specifications. If you want to know more about philosophical or UI design reasons for its removal, just read the pro-removal posts for bug 96972 on bugzilla.mozilla.org. Number 60 is particularly good.
On a more practical note, at work I have to fill out online time cards every two weeks, which involve me filling in lots of duplicate values into a HTML table of text fields. At least once every time, and if I am unlucky, more than once, I will accidentally middle click on the edge of the text field, rather than the field itself, and bam the browser sends me to some web page it interpreted from the numbers or letters I pasted. And when I go back, I will have lost everything I had filled in and have to start over. It occasionally happens to me when I am filling out other web forms as well. Actually, it just happened to me while I was searching for the bug number. Grrr. Needless to say, this bug has driven me crazy.
Don't get me wrong, if Mozilla/Firebird made an icon target for the toolbar, on which you could paste a URL to have it open in [choose one: current window, new window, new tab] I would be all for it. However, as things currently stand, the "functionality" is neither well thought out nor properly implemented; it really is just a bug that happens to have just enough vocal supporters to keep it from being removed, while being a serious annoyance to the rest of us.
In the past, Red Hat only backported security fixes and major bug fixes, but not new features and other new things, in the updates to a given release. So your httpd was not in fact version 2.4.47, but 2.4.40 with security patches for 2.4.47. This practice helped ensure intra-release stability for commercial users. It is considered an acceptable practice and is used by other distros, like Debian stable.
Now, one of the goals of the Fedora Porject is to do more maintenence upstream, from which I imply that they want to end this practice and simply bring new versions of software immediately into the update stream, rather than waiting for the next release cycle. This will be better for home users, but it might in some cases not be good for commercial users, who would rather stability over bleeding edge.
By more clearly splitting the hobbyist OS from the enterprise OS, they can now offer the best to each world.
That sucks, but it is not unexpected. Red Hat does not want to open themselves to a lawsuit from the "owners" of mp3, and now that they are merged with Fedora, Fedora cannot distribute them either.
It's not a big problem though, as you can still easily get such packages from other compatible repositories such as freshrpms.
The irony is that you just described a Tolkien-esque quest of your own, with yourself as the small, weak hero on the side of right and Tolkien's estate as the seemingly-unbeatable villain. Tolkien began many of his books with a personal request to readers to not violate his copyright and give him proper attribution, but the actions you described, taken in his name, would make him turn over in his grave. The Tolkien estate should be ashamed of their acts which have brought dishonor to the name of the man to whom they owe their fortune and fame.
This ought to get modded up as en excellent real-world example of abuse of modern copyright law.
All companies will do anything, ANYTHING, to cut costs. So let me rephrase my original idea. Do you think that companies which are already unscrupulous enough to have fired employees and rehired them as temps would be likely to try to screw Sun using this scheme?
Of course I know that no CEO is going to say "Hey, let's fire all of our employees and rehire them as temps just to screw Sun." But for a company which has already chosen to go the temp route, it would be icing on the cake. And for a company which is contemplating it, it could be another "pro" in favor of firing/rehiring.
Finally, what about companies which outsource a lot of their IT and development infrastructure? They could provide access to Sun's hardware and services to the outsourced employees, but since they are not "real" employees, they don't count toward licensing costs.
Well, I am sure that if I thought of this, then so did someone in Sun legal. So I suppose Sun thinks they will make more money off of this, even despite these apparent loopholes.
"The way I read this, it means only programs which enable the user to both serve and request files. So plain FTP or HTTP are okay because they are really only one way protocols, but Gnutella or Freenet aren't because they allow for two-way communication."
If this is true, then we could simply split the P2P software in independent client and server parts and circumvent the law.
That may actually satisfy the RIAA. The vast majority of P2P users I have seen really are not interested in sharing music, (Of course I know some people are, but hey, I'm talking about the majority) and only share because the integrated client/server (servent?) makes it so darned "convenient" to share or difficult to opt out of sharing. If the P2P software were split in two, most people I know would only install and run the client. I am willing to be that most people in the world would do the same.
If this happened, then massive amounts of content would drop off of P2P networks, and they would become less useful to leech users, who would probably quit using. Fewer providers would make it easier for the *AAs to crack down on genuine copyright infringement, as the safety-in-numbers defense would vanish. Ultimately, the networks would be used primarily by people legitimately interested in sharing copylefted works.
So in effect, this might be a good thing for P2P networks and independent artists. Naturally, I am opposed to legislation mandating it, but P2P makers might want to consider voluntarily doing it to try to improve the currently abysmal signal/noise ratio on networks.
(Of course, Freenet cannot do this, because the entire protocol depends upon not knowing whether a node is a passive client or a server.)
Does anyone know whether contract or temporary workers count as employees on these annual filings? My cousin's company recently fired virtually all of their employees, then rehired about 70 percent back as "temporary contractors" with a 10 percent raise. The company saved big-time, because they no longer had to pay benefits, insurance, and so forth for "temporary" employees. Despite the take-home pay raise, the employees lost out because they had to pay for insurance and retirement out of pocket. However, most accepted the deal because they didn't lose enough to make them want to brave today's tech job market.
So if "temporary" employees do not, in fact, count toward this annual report, it seems that companies might be able to screw Sun by firing employees and then re-hiring them as "temps".
Any accountants out there who can shed some light on this?
Mandrake originally started as Red Hat with KDE and a few other bonuses, ergo you see many more similarities between those two than Debian or Slackware. However, I do not believe that Mandrake tracks Red Hat development anymore (except the RPM package format); they are their own distro now.
Doug Engelbart made an appearance at my university (Drexel) a few years ago (1999?) and one of the undergrads there asked him to sign a mouse. The funny thing was, it was a Mac mouse and Doug sort of poked fun at it for only having one button, as the mouse as he originally designed it had three.
If the only reason you pay for Red Hat Network is to get automatic updates, I strongly suggest you look at apt-get for rpm. It provides the exact same updates as up2date, only they are free. If you don't trust them you can check the digital sigs on the packages; they come unaltered from Red Hat. Optionally, it can also provide additional packages not found on the Red Hat distribution.
Apt-get doesn't explicitly notify you when updates come in, however it is trivial to write a script to automate the process of checking for updates. For the super-lazy, you can even continue to use the free version of Red Hat's up2date notification icon to alert you when updates come in, and then use apt-get to actually fetch them.
Of course, there are probably other reasons you pay for RHN, such as technical support, a desire to give back to Red Hat, etc...
Just thought I'd make sure you know about an excellent free alternative.
Seriously, the very LAST thing we need is to encourage terrorist action in "support" of open source.
Wake up, get a life, and realize that no matter how much you might think you want it, it's not okay to say it. It doesn't matter whether you are joking or not. There are some comments that you need to learn to self-censor; this is one of them.
You don't want to be responsible for encouraging some whacko to actually go out and do it. If you think it seems that the US gov't isn't sympathetic to open source now, just see how they react after some loonie terrorist claims to have "done it all for open source." Just wait until membership in the FSF gets your house raided and your phone tapped...
The way I read this, it means only programs which enable the user to both serve and request files. So plain FTP or HTTP are okay because they are really only one way protocols, but Gnutella or Freenet aren't because they allow for two-way communication. So they aren't trying to blanket outlaw the internet.
That said, this proposed law, if enacted, would be far worse than even the DMCA. It would effectively allow the government to regulate what kinds of Internet protocols are acceptable, and more generally, how people are allowed to communicate over the internet.
I hate to rain on your parade, but you just illegitimately reversed the burden of proof.
Shri stated that companies X, Y, and Z obey the law. You just said "prove it." But it doesn't work that way.
It is natural to assume, legally and intuitively, that entities (people and organizations of people) obey rules, and we only assume otherwise when shown some evidence. For example, an anonymous coward in this thread pointed out some evidence of specific wrongdoing on Berkshire Hathaway's part, so he could then legitimately ask shri for evidence of Berkshire Hathaway's innocence. However, lacking any evidence or even specific allegations of wrongdoing, you cannot ask for nor can you expect shri to provide proof of innocence.
Furthermore, stereotyping is not a legitimate form of proof. You cannot legitimately make a statement like: "All businesses cheat, therefore company X cheats." Although this is logically valid, it is not a sound argument because you haven't provided evidence, nor can you, that the assertion "all companies cheat" is true.
In closing, I could say "geekoid does not collect child pornography." I bet you would get understandably upset if someone then brashly stated "All geeks look at kiddie porn, therefore geekoid also, so prove he doesn't." Stereotyping and absence of proof of innocence are not proofs of guilt.
Slashdot Mirror
Fedora is now the "officially" tested and sanctioned free release made by the company known as Red Hat. The only thing that has changed is that you cannot pay Red Hat for phone support now, but that is no loss for Red Hat as they were losing money on that anyway.
Furthermore, the Fedora packages form the basis of the for-pay Red Hat Enterprise Linux system, and the improvements Red Hat funds for them go back, via the GPL, into Fedora.
As Fedora is sponsored and to some extent driven by Red Hat, this is a non-issue for most of the packages (virtually all of the major ones) in Fedora.
8th Affirmative Defense - The GPL violates the U.S. Constitution, together with copyright, antitrust and export control laws, and IBM's claims based theron, or related thereto, are barred.
Actually, being involved in litigation several times myself, I have come to realize that a standard defensive tactic lawyers use is to bring out every conceivable defense in their responses, no matter how ludicrous. For example, in my defense against a simple traffic violation, my lawyer entered in the pleadings several defenses along the lines of "Section X of Motor Vehicle Code violates the US Constitution" and "Section X of Motor Vehicle Code is selectively enforced by Police such that its enforcement is waived, estopped or otherwise barred as a matter of equity". The lawyer knew that the chances of a judge accepting those defenses were about 1:2^128, but he used them anyway.
So these clauses look like standard litigation boilerplate to me.
This is probably the best "simple" explanation of why the recent American practice of attaching "Terms and Conditions" to everything is wrong. The post author does not assert that it is wrong based on some vaguely defined morals or ethics, but that it is wrong based on plain old economic principles.
So to get the peace prize now you have to add kidnapping and threat of homicide....
Hey, Yassar Arafat won the Peace Prize, so I guess he paved the way.
lordcorusa ducks to avoid the flames.
Certainly Nvidia's legal department believes that their drivers fall on the legal side of the fence, and I for one am glad, as I enjoy the Nvidia drivers on my GeForce4 ;-)
However, that's not to say those drivers don't inhabit a grey area of GPL law. While the drivers contain no Linux code per se, I believe they do make use of interfaces of the Linux kernel which are neither published nor intended for normal use. Therefore, it could be said that the code that makes up those drivers is a derivative work of the Linux kernel, as Nvidia has to study kernel source code carefully and adapt their code so as to be able to link directly into it. I suppose they could do some cleanroom reverse engineering to get around that, or something.
But I am not an IP attorney and even if I were, none of the Linux developers seem to be concerned about Nvidia's kernel drivers.
The linux kernel is GPL'd with the added provision that you are allowed to write a binary moduled to be loaded at runtime without having to release your code under the GPL.
Interesting. Where in the kernel distribution does it explicitly state this? I downloaded 2.4.22 from kernel.org and looked in the COPYING file but found nothing authorizing this. Rather, the COPYING simply state that the kernel is released under the terms of GPLv2. I also looked on some of the main pages of kernel.org and could not find anything.
You are probably referring to Linus's interpretation of the GPL that he posted to the lkml. However, as he is only one of many kernel authors, he doesn't have the authority necessary to make interpretations which relax the license for the kernel. He would need explicit permission from every kernel author to make those exemptions to the license. Some of the module owners are firmly in the Free Software camp, so he would probably have to remove or rewrite some parts of the kernel, as they would be unlikely to acquiesce.
If he got all of the permission necessary, he would need to add the information about the deviation from vanilla GPLv2 to the COPYING file. Thus far, I am not aware of his gathering permission from everyone, and he certainly has not changed the COPYING file. Therefore, to the best of my knowledge gathered from the research I have done, the kernel is still technically licensed under the vanilla GPLv2, which means that dynamic, or runtime, linking of a module still requires GPLing the module.
Now, given Linus's stance, a company might feel safe in making a binary-only runtime module. (*cough* Nvidia) However, technically that is still in violation of the license that is attached to current kernels, and it could potentially open the company to legal action in the future, as Linus is free to change his mind, and any of the other developers could also initiate the complaint.
I don't mean to be redundant, but I see this mistake all the time and I don't want people to go away with incorrect ideas. You need to read your licenses carefully, especially if you write or modify code under them.
The Linux kernel is GPLed, not LGPLed. You can link LGPLed code with new code dynamically without releasing source to the new code. However, with the GPL, you *must* release source regardless of whether the code is statically or dynamically linked. So even if Linksys dynamically linked their proprietary modules with the kernel, it would still be a GPL violation.
The only way to avoid distributing source code that works with GPLed code is to write a GPL-compatibly-licensed wrapper program for the GPLed code. Then you write the proprietary program which uses IPC or some other arms-length method of talking to the wrapper.
Let's call it Youvert! ;-)
Is it just me, or does that sound like a good script for a new TV commercial? I am thinking something like the IBM commercials, but IBM doesn't want to antagonize Microsoft. Perhaps a Novell or Sun commercial to go with their new desktop offerings?
;-) And rather than the last line being "No, it doesn't" substitute "Not anymore!"
The MS proponent can be a stereotypical stressed-out, disheveled, annoying person, and the Linux proponent can be calm, cool, and collected. Just substitute reference to "Peoplesoft" with the sponsor company's name. Rather than "So are we" substitute "We're bigger". Bigger is always better.
apt-get and Synaptic are absolutely fabulous apps for those of us who are experienced using Linux and installing software for it. However, they fail to help newbies because the process they use to get and install software does not map to the way the newbie thinks about installing software. And anyone who says "The way the newbie thinks is wrong" is simply failing to understand newbies at all.
Now that I've made a generalization, let's have a specific run-through of the problem. I have first-hand experience with switching (some successfully, some not) a number of Windows users to Linux, and here is the problem that they all run into. (NOTE: I ran into this very same problem when I first switched, but I knew no one who could hold my hand through it all. The only reasons I am still using Linux are that I am far more intelligent than most people, so I am better at figuring things out on my own, and I am also incredibly stubborn when it comes to learning something new.)
You are Joe Newbie. You've got your nice shiny Linux system running. You hear about a great app called "FooBar". You like what you hear about it and you decide you want to try it. You search google for it, and go to www.foobar-software.org. You try to download it. But you can only get source or an RPM or DEB package. (Let's assume your friend set you up with Libranet and for some reason you actually know that it's a Debian based distro -- a stretch in its own right) You download the DEB, but you run into dependency conflicts. It wouldn't be so bad if this happened once or twice, but it happens for bloody near every app you try to install.
Now all seasoned Debian users, as well as most users of other distros, will be screaming at you to use apt-get or synaptic, or whatever other package management system. But that's the problem. People accustomed to Windows or Mac are accustomed to going to a store or a vendor's web site and getting the software they want. They have no idea that their computer might somehow "magically" know how to get it for them. Heck, if you didn't know better, why would you think your system would know how to do that? Even when you show them how apt-get works, they still often forget and revert back to the old way. It's a very deeply ingrained habit that only the most persistent learn to break.
To make things worse, even Debian, with a repository probably more exhaustive than any other distro's, still doesn't have all of the packages (and new versions of packages) that a user wants. If that user is a newbie, having to remember multiple methods for acquiring software and knowing when to use each is a further strain.
Granted, if a person sticks with Linux and becomes more accustomed to it, he or she will probably learn how to use apt-get (or insert package management system here) to streamline the package-acquisition process. However, it would be in open source's best interests to try to minimize culture shock so as to further help bring more people in.
So the grandparent poster was dead-on. For example, Mozilla really does get it. Their installer is distro-agnostic and installs everything that the package needs. Even though this may introduce more bloat (redundant packages) for any given distro, it also results in an easy installation for a newbie. Advanced users will know how to get better versions of Mozilla specifically for their distros, but newbies will still be able to participate by getting Mozilla the way they know best.
All end-user focused software packages should follow Mozilla's example of providing a simple executable installer which contains all libraries and files needed to run the software independently of most, if not all, other packages on the system. This certainly isn't the ideal setup from a sysadmin or advanced user standpoint, but it is needed to match the way newbies think about installing software.
I know I shouldn't bite on troll bait, but...
.... I find "sharing" to mean that you are going out of your way to give something to someone else.
Actually yes, telling stories requires effort. If I want to tell a good story I have to get the story in order, get a good audience, set the mood, etc. It takes work.
Even by your twisted logic, I am sharing using P2P. I am sacrificing my bandwidth, for which I pay some sum of money, for someone else's benefit. Regardless of whether I download any files in return, I am giving my upload speed to other people and they don't necessarily have to give anything back to me.
Not that I condone widespread copyright violation. I do wish people would share more copylefted stuff instead of pop crap. However, proles will be proles. But I think that if you are going to blast them, you should at least have a logically consistent viewpoint.
The simple answer is that if you are developing commercial software, you should not be targetting the Fedora distribution. Red Hat and the Fedora project clearly intend this to be like the equivalent of Debian unstable/testing. How many sane commercial software developers would target Debian unstable/testing?
If you want to target an "official" Red Hat distribution with guaranteed stability, Red Hat clearly wants you to target their Enterprise Linux. If you don't want to or cannot afford it, then I would say it's time to stop targetting Red Hat distributions at all, and go with Suse Enterprise, Debian stable, or some other distribution which better meets your needs.
If you really, really want to target Fedora, your best bet would be to bundle all required libraries into one big package, or statically compiling everything, assuming you legally can according to the licenses of the components (ala the Kompany). You might not like the idea of putting all of that "bloat" into your nice, clean little app, but that is what will be necessary to guarantee the level of package-stability you seem to require.
Whether or not you like this new strategery, this is the tack Red Hat appears to be taking. Now you need to decide whether you should keep fishing or move on to some other waters.
At least in mozilla and the mozilla based browsers, you can load a url by pasting it (middle clicking) anywhere on an open page or empty tab.
WARNING: The so-called feature of Mozilla that you are espousing is actually a very old bug (#96972). It was never supposed to be there, and there has been a long-standing discussion about whether to remove it. Many people have filed bug reports and/or requested/voted that it be removed. However, the bug has just enough very vocal supporters among the hacker community to block its removal. Regardless of the debate, this "feature" is classified as a bug, so it is dangerous to just assume it will always be there, since it may one day be stripped out.
Of course, I am biased toward the pro-removal camp. As a software engineer and UI designer, I believe in strictly following specifications. If you want to know more about philosophical or UI design reasons for its removal, just read the pro-removal posts for bug 96972 on bugzilla.mozilla.org. Number 60 is particularly good.
On a more practical note, at work I have to fill out online time cards every two weeks, which involve me filling in lots of duplicate values into a HTML table of text fields. At least once every time, and if I am unlucky, more than once, I will accidentally middle click on the edge of the text field, rather than the field itself, and bam the browser sends me to some web page it interpreted from the numbers or letters I pasted. And when I go back, I will have lost everything I had filled in and have to start over. It occasionally happens to me when I am filling out other web forms as well. Actually, it just happened to me while I was searching for the bug number. Grrr. Needless to say, this bug has driven me crazy.
Don't get me wrong, if Mozilla/Firebird made an icon target for the toolbar, on which you could paste a URL to have it open in [choose one: current window, new window, new tab] I would be all for it. However, as things currently stand, the "functionality" is neither well thought out nor properly implemented; it really is just a bug that happens to have just enough vocal supporters to keep it from being removed, while being a serious annoyance to the rest of us.
In the past, Red Hat only backported security fixes and major bug fixes, but not new features and other new things, in the updates to a given release. So your httpd was not in fact version 2.4.47, but 2.4.40 with security patches for 2.4.47. This practice helped ensure intra-release stability for commercial users. It is considered an acceptable practice and is used by other distros, like Debian stable.
Now, one of the goals of the Fedora Porject is to do more maintenence upstream, from which I imply that they want to end this practice and simply bring new versions of software immediately into the update stream, rather than waiting for the next release cycle. This will be better for home users, but it might in some cases not be good for commercial users, who would rather stability over bleeding edge.
By more clearly splitting the hobbyist OS from the enterprise OS, they can now offer the best to each world.
That sucks, but it is not unexpected. Red Hat does not want to open themselves to a lawsuit from the "owners" of mp3, and now that they are merged with Fedora, Fedora cannot distribute them either.
It's not a big problem though, as you can still easily get such packages from other compatible repositories such as freshrpms.
The irony is that you just described a Tolkien-esque quest of your own, with yourself as the small, weak hero on the side of right and Tolkien's estate as the seemingly-unbeatable villain. Tolkien began many of his books with a personal request to readers to not violate his copyright and give him proper attribution, but the actions you described, taken in his name, would make him turn over in his grave. The Tolkien estate should be ashamed of their acts which have brought dishonor to the name of the man to whom they owe their fortune and fame.
This ought to get modded up as en excellent real-world example of abuse of modern copyright law.
All companies will do anything, ANYTHING, to cut costs. So let me rephrase my original idea. Do you think that companies which are already unscrupulous enough to have fired employees and rehired them as temps would be likely to try to screw Sun using this scheme?
Of course I know that no CEO is going to say "Hey, let's fire all of our employees and rehire them as temps just to screw Sun." But for a company which has already chosen to go the temp route, it would be icing on the cake. And for a company which is contemplating it, it could be another "pro" in favor of firing/rehiring.
Finally, what about companies which outsource a lot of their IT and development infrastructure? They could provide access to Sun's hardware and services to the outsourced employees, but since they are not "real" employees, they don't count toward licensing costs.
Well, I am sure that if I thought of this, then so did someone in Sun legal. So I suppose Sun thinks they will make more money off of this, even despite these apparent loopholes.
"The way I read this, it means only programs which enable the user to both serve and request files. So plain FTP or HTTP are okay because they are really only one way protocols, but Gnutella or Freenet aren't because they allow for two-way communication."
If this is true, then we could simply split the P2P software in independent client and server parts and circumvent the law.
That may actually satisfy the RIAA. The vast majority of P2P users I have seen really are not interested in sharing music, (Of course I know some people are, but hey, I'm talking about the majority) and only share because the integrated client/server (servent?) makes it so darned "convenient" to share or difficult to opt out of sharing. If the P2P software were split in two, most people I know would only install and run the client. I am willing to be that most people in the world would do the same.
If this happened, then massive amounts of content would drop off of P2P networks, and they would become less useful to leech users, who would probably quit using. Fewer providers would make it easier for the *AAs to crack down on genuine copyright infringement, as the safety-in-numbers defense would vanish. Ultimately, the networks would be used primarily by people legitimately interested in sharing copylefted works.
So in effect, this might be a good thing for P2P networks and independent artists. Naturally, I am opposed to legislation mandating it, but P2P makers might want to consider voluntarily doing it to try to improve the currently abysmal signal/noise ratio on networks.
(Of course, Freenet cannot do this, because the entire protocol depends upon not knowing whether a node is a passive client or a server.)
Does anyone know whether contract or temporary workers count as employees on these annual filings? My cousin's company recently fired virtually all of their employees, then rehired about 70 percent back as "temporary contractors" with a 10 percent raise. The company saved big-time, because they no longer had to pay benefits, insurance, and so forth for "temporary" employees. Despite the take-home pay raise, the employees lost out because they had to pay for insurance and retirement out of pocket. However, most accepted the deal because they didn't lose enough to make them want to brave today's tech job market.
So if "temporary" employees do not, in fact, count toward this annual report, it seems that companies might be able to screw Sun by firing employees and then re-hiring them as "temps".
Any accountants out there who can shed some light on this?
Mandrake originally started as Red Hat with KDE and a few other bonuses, ergo you see many more similarities between those two than Debian or Slackware. However, I do not believe that Mandrake tracks Red Hat development anymore (except the RPM package format); they are their own distro now.
Doug Engelbart made an appearance at my university (Drexel) a few years ago (1999?) and one of the undergrads there asked him to sign a mouse. The funny thing was, it was a Mac mouse and Doug sort of poked fun at it for only having one button, as the mouse as he originally designed it had three.
If the only reason you pay for Red Hat Network is to get automatic updates, I strongly suggest you look at apt-get for rpm. It provides the exact same updates as up2date, only they are free. If you don't trust them you can check the digital sigs on the packages; they come unaltered from Red Hat. Optionally, it can also provide additional packages not found on the Red Hat distribution.
Apt-get doesn't explicitly notify you when updates come in, however it is trivial to write a script to automate the process of checking for updates. For the super-lazy, you can even continue to use the free version of Red Hat's up2date notification icon to alert you when updates come in, and then use apt-get to actually fetch them.
Of course, there are probably other reasons you pay for RHN, such as technical support, a desire to give back to Red Hat, etc...
Just thought I'd make sure you know about an excellent free alternative.
Seriously, the very LAST thing we need is to encourage terrorist action in "support" of open source.
Wake up, get a life, and realize that no matter how much you might think you want it, it's not okay to say it. It doesn't matter whether you are joking or not. There are some comments that you need to learn to self-censor; this is one of them.
You don't want to be responsible for encouraging some whacko to actually go out and do it. If you think it seems that the US gov't isn't sympathetic to open source now, just see how they react after some loonie terrorist claims to have "done it all for open source." Just wait until membership in the FSF gets your house raided and your phone tapped...
The way I read this, it means only programs which enable the user to both serve and request files. So plain FTP or HTTP are okay because they are really only one way protocols, but Gnutella or Freenet aren't because they allow for two-way communication. So they aren't trying to blanket outlaw the internet.
That said, this proposed law, if enacted, would be far worse than even the DMCA. It would effectively allow the government to regulate what kinds of Internet protocols are acceptable, and more generally, how people are allowed to communicate over the internet.
I hate to rain on your parade, but you just illegitimately reversed the burden of proof.
Shri stated that companies X, Y, and Z obey the law. You just said "prove it." But it doesn't work that way.
It is natural to assume, legally and intuitively, that entities (people and organizations of people) obey rules, and we only assume otherwise when shown some evidence. For example, an anonymous coward in this thread pointed out some evidence of specific wrongdoing on Berkshire Hathaway's part, so he could then legitimately ask shri for evidence of Berkshire Hathaway's innocence. However, lacking any evidence or even specific allegations of wrongdoing, you cannot ask for nor can you expect shri to provide proof of innocence.
Furthermore, stereotyping is not a legitimate form of proof. You cannot legitimately make a statement like: "All businesses cheat, therefore company X cheats." Although this is logically valid, it is not a sound argument because you haven't provided evidence, nor can you, that the assertion "all companies cheat" is true.
In closing, I could say "geekoid does not collect child pornography." I bet you would get understandably upset if someone then brashly stated "All geeks look at kiddie porn, therefore geekoid also, so prove he doesn't." Stereotyping and absence of proof of innocence are not proofs of guilt.