I am always a bit wary when people talk about usage of some particular system "doubling." Linux usage will undoubtedly increase on the desktop, but doubling? That's a very generous estimate of the situation. I would be happy if it doubled, if only because people would slowly stop assuming that everybody in the world is either a Windows or Mac user. Let's be serious though: it will not double, unless someone big like Dell or HP actually tries marketing their Linux systems. Dell's Linux systems go un-advertised, and therefore unnoticed by anyone who doesn't know what Linux is or why it is worth their time.
Again, it depends on the OS. Like I said, there is binary compatibility between RHEL4, RHEL5, Fedora core 3, Fedora core 4, Fedora core 5, Fedora core 6, Fedora 7, Fedora 8, CentOS 4, and CentOS 5. There are compatibility libraries for that, and they are in the repositories for anyone who happens to need them.
Also note that Microsoft ships compatibility libraries with each version of Windows. The Visual Studio 6 runtime libraries are still shipped with Vista, and they date back all the way to 1998. You mentioned different library versions; try running the original Doom95 binary in an out of the box Vista install, and you will encounter a similar problem, because Doom95 relies on a very out of date graphics library. In Windows, this is referred to as "DLL hell," but it is really a problem that transcends OSes.
...or if you are trying to manage more than 1 computer. My university pays more per yer to manage Windows on several hundred desktops than it spends on new books for the library. Ironically, many of those desktops are IN the library and used to find books.
Suppose they rolled out CentOS or RHEL (since they like spending money) on those systems. The systems could be managed, in their entirety, from one location, without requiring the $100000+ in other commercial software they use right now; SSH and a few quick scripts would keep everything in order (in fact, that is how things used to be managed). They would also be paying less for the antivirus, antispyware, anti* software they currently need to use to prevent students from accidentally screwing things up -- SELinux in enforcing mode would take care of that. Despite the hundreds of thousands of dollars the university pours into keeping the desktops up and running, people still manage to screw them up; again, Linux systems have proved their worth before, and as a case study, Columbia universities public terminals are all running Linux and are a lot more reliable than ours (a smart person will note that I have not named my university).
In total, I would estimate that about $500000 could be saved, right off the bat, if our desktops were all migrated to CentOS, RHEL, or a similar EL system. That's just our desktops, and that's assuming that we didn't have to get rid of all the expensive end user software that they've installed, like Matlab (which is available on a rather beefy Solaris server anyway, so why are we paying for a site license? Because the Windows desktops don't handle X11 tunneling very well). Counting the various servers and other systems that we are running, most of which could be replaced fairly easily with an open source system, the savings climb way up into the millions.
"Linux is terrible when it comes to binary portability."
Which is, while true, a misleading statement. Yes, there is poor binary portability between different distros, but that is because they are different operating systems, and just happen to have extremely similar (or sometimes identical) kernels. There is actually very good binary compatibility between my computer running Fedora 8, my girlfriend's computer running Fedora 8, and my uncle's computer running RHEL 4. In fact, binary packages for RHEL 4 can be run in the latest Fedora, RHEL, or CentOS, as long as you have the compat libraries installed (e.g. libstdc++5).
The moral of the story? Unless you expect binary compatibility between Windows and FreeBSD, don't expect binary compatibility between different Linux distros. Different distros are different OS's, with different goals and different target markets or applications, and very often different philosophies.
Redundancy has its limits, and a major link going down could potentially have a cascading effect that causes redundant links to go down. Consider this example: A and B are connected by two redundant links, each operating at 75% of capacity. If one link goes down, and the other takes over, it will now be operating at 150%, and will quickly become clogged; it could become so busy, in fact, that it goes down entirely.
My guess is that the situation in the middle east is similar. This was a major link, and when it went down, the traffic spilled over to other links and overloaded them, causing the disruption in service. Then again, I don't work in the middle east, so I don't really know how they set things up...
If big corporate companies need backward compatibility, they can get it through virtualization or through an optional legacy support package of some sort (akin to the way many Linux distros have a GCC3 package for older apps). The article's point is that Vista shouldn't try to be a single package that satisfies everyone. Home users don't usually need the same level of backward compatibility as enterprise customers, so why should they be burdened with it?
The music thing also makes sense; let OEMs figure out how they want the system to work. Maybe a workstation producer won't want to include all the music and movie stuff, but an entertainment system should have it. Again, different needs in different key markets. Fedora is doing that with its "spins," there is no reason Microsoft can't do it.
The goal is to defeat the RIAA, not concede on the DRM issue. If the RIAA is unwilling to sell DRM-free music, then we just need to stop buying their music, and start listening to the music of independent artists who use open formats like OGG. The danger of DRM is underscored by this sort of behavior -- hard-coding maintenance programs to ignore certain processes? We've been down this road before.
What they should do is require the meta tag for nonstandard rendering, and make standard rendering the default. Web developers shouldn't be encouraged to ignore standards.
Unfortunately, trying to make encryption both secure and transparent is not an easy task. A good encryption system encrypts the private key itself; that requires the user to enter a password to use the encryption at all. Even that is more than most end users are willing to put up with.
Please, most people don't have any knowledge of email encryption, and probably have no idea about proposals like these. Worse, even people who are aware of email encryption don't want to be hassled with it, because it involves typing in an extra password (and heaven forbid we should do anything like that). Sadly, privacy is at the bottom of most people's priority lists, just under "free speech" and "due process" (yes, most of the people I know think it is wrong for a lawyer to defend someone who is "obviously guilty"). This won't scare anyone to action. Encryption should have become popular after several policies were enacted, and it still hasn't. The problem is very simple: most people do not care enough to go to the trouble of encryption.
Except that this isn't parents trying to provide a good environment, it is just parents trying to bar their children from access to a certain website. Parents trying to provide a good environment would sit down and talk about the dangers of sexual predators on MySpace and similar websites, and instruct their children to immediately contact mommy or daddy if someone starts propositioning them for sex (not that we live in a culture where parents are encouraged to discuss anything pertaining to sex with their children). Growing up, the Internet was just starting to reach its current level of popularity, and my mother was very clear with me when we got our first computer about what to do if someone asked to meet me or started talking about sex, I listened, and there was never a problem with me using the computer, even if I was unsupervised.
Oh well, we haven't encouraged parents to actually speak to their kids about this stuff for a long time, opting to shield children from anything deemed harmful by anyone.
This is just another attempt by some politicians to claim that they are fighting to protect our children. Later on, when nobody actually remembers any of this, these politicians can tell a cheering crowd, "I worked hard to give parents the ability to limit their child's MySpace access, and help shield their children from sexual predators online." Of course it is idiotic, and children will find a way around it in less than a minute, but if this were really about protecting our children, it would be an educational program, not another pathetic attempt at technical measures to block their access.
Parents already have the ability to employ technical measures to block their children's access to MySpace, so all this really is is a way for politicians to claim that they are fighting for our children. Yeah, I know, kids can already get around other measures, but hey, kids can get around this too, as everyone has already pointed out.
What these 49 states should be doing (as should the 1 state not participating) is starting a program to educate children and their parents about the risks of social networking websites, and how to handle situations where a predator might contact a child on such websites.
It depends on how you phrase it. For example, this is the explanation I commonly give when I tell people about Red Hat and how they give things like JBOSS out for free:
"The business model is centered around charging businesses to get help from a Red Hat employee with setting up or troubleshooting their software, so it is in Red Hat's best interests to give the software out for free. Of course, for people like us, that just means getting free, legal software to use."
People are very open to that idea, because it makes perfect sense to them. Granted, RHEL itself is not free, but that's the sort of explanation people are responsive to (and it is true) and it helps to build trust in free (as in beer) software. The idea that open source software is inherently of a lower quality still exists, but the way to show that that is not the case is to mention companies like Red Hat, Mandriva, and Canonical, who have made a lot of money giving all this stuff away.
RTFA. The document in question was a collage, with a picture of the university president, a bulldozer, and some text referencing the "memorial garage."
Amazingly, that is exactly how to generate interest. People like things that are shiny, translucent, 3D, etc.
We actually had a discussion here at FUDCon about this very issue yesterday, and one of the real problems most people encountered wasn't generated interest as much as it was keeping people interested when they encounter little bugs or usability issues (there are quite a few in the Fedora utils which will be resolved this year). Unfortunately, polishing off apps to make them more usable (or even just having them update the UI during a complex operation) is not a high salience issue, and so such things often get neglected in open source projects, even though they can be a deciding factor in keeping non-programmers interested.
That's a very big stretch. That statement could be read in another, more likely and more innocuous way, that the president of the university wanted the garage named for him (I guess there weren't any other buildings left). It hardly seems to be a threat, and you would need counseling yourself if you started walking around with plain-clothed policemen because you thought that the collage was a threatening document. The threat was to this president's plan to build the garage, and so he just found a clever way to rid himself of that "problem."
School shootings seem to be used as a pretext for schools to accomplish their non-academic goals these days. At my university, for example, the dining halls recently received several large, flat panel TV's each, which provide us with vital information about the price of food and upcoming "dining hall events" (food that isn't normally served but is just as bad). When I noted to a friend that this all seemed like a waste of electricity, especially since we have a coal-fired power plant right on campus, one of the dining hall supervisors overheard me and said, "Yeah, but these can also be used as an emergency communications system,..." and went on to talk about how students need to be informed.
It was easy to call bullshit, since we already had a system for that. More to the point, using people's fear of a lunatic going on a shooting rampage to justify ludicrous measures like my school's TV's or this George school expelling this student is a disgrace.
People who say that a problem facing Linux is that there are too many distributions and too many different ways to do things have the wrong perception of what an OS is. Different distributions are different operating systems, and expecting binary compatibility across different OS's is folly regardless of what kernel is in use. That's why systems like automake/autoconf exist, and standard API's like POSIX, exist, so that source code can be recompiled on different platforms without too much pain.
"Linux" is not a single operating system, it is just a kernel. The kernel can be run without GNU utils, without X11, etc.
LANCOR is claiming to have a patent on a keyboard that allows the user to type in Nigerian. How can you possibly call that an innovation, and how can you possibly accuse anyone of stealing that? Do you think that Nigerians should be required to pay LANCOR every time they write anything down, or just when they decide to type in their native language? How can anyone claim a patent on a system that follows the already existing rules of any language, or anything at all? They didn't invent anything, they just put into code a long list of rules that already existed.
Here's a problem for you: patent a keyboard that can render all the strange facets of written English: upper case letters, lower case letters, diacritical markings, punctuation, etc. Come back to me when coding language rules is considered to be innovation.
Red Hat and Canonical are competing corporations. They have very similar business models and they are targeting the same markets (Canonical recently announced a push to enter the enterprise market, Red Hat's stronghold).
Ubuntu's main selling point was that it was easy to set up and use, that it "just worked," and that it had some fun media programs. Fedora 8 has all those strong points, and a few features that are of interest to people who know about computing (the way daemons and applets interact, for example).
Ultimately, however, the issue is that nobody except the developers of this product actually knows what the purpose of those connects is. Is it really just for quality tracking and product news/offers? Is it possibly also for the purpose of stopping copyright infringement (that is, illegal software use)? Is it sending your actual activities to Adobe?
This is probably innocuous, but who knows? This is what happens when software is distributed only in binary form -- users pick up on something suspicious, and start to assume the worst, and then that whole mob mentality enters the picture. Suppose you discovered that GIMP was connecting to a strange looking host -- what would you do then?
Slashdot Mirror
I am always a bit wary when people talk about usage of some particular system "doubling." Linux usage will undoubtedly increase on the desktop, but doubling? That's a very generous estimate of the situation. I would be happy if it doubled, if only because people would slowly stop assuming that everybody in the world is either a Windows or Mac user. Let's be serious though: it will not double, unless someone big like Dell or HP actually tries marketing their Linux systems. Dell's Linux systems go un-advertised, and therefore unnoticed by anyone who doesn't know what Linux is or why it is worth their time.
Also note that Microsoft ships compatibility libraries with each version of Windows. The Visual Studio 6 runtime libraries are still shipped with Vista, and they date back all the way to 1998. You mentioned different library versions; try running the original Doom95 binary in an out of the box Vista install, and you will encounter a similar problem, because Doom95 relies on a very out of date graphics library. In Windows, this is referred to as "DLL hell," but it is really a problem that transcends OSes.
Suppose they rolled out CentOS or RHEL (since they like spending money) on those systems. The systems could be managed, in their entirety, from one location, without requiring the $100000+ in other commercial software they use right now; SSH and a few quick scripts would keep everything in order (in fact, that is how things used to be managed). They would also be paying less for the antivirus, antispyware, anti* software they currently need to use to prevent students from accidentally screwing things up -- SELinux in enforcing mode would take care of that. Despite the hundreds of thousands of dollars the university pours into keeping the desktops up and running, people still manage to screw them up; again, Linux systems have proved their worth before, and as a case study, Columbia universities public terminals are all running Linux and are a lot more reliable than ours (a smart person will note that I have not named my university).
In total, I would estimate that about $500000 could be saved, right off the bat, if our desktops were all migrated to CentOS, RHEL, or a similar EL system. That's just our desktops, and that's assuming that we didn't have to get rid of all the expensive end user software that they've installed, like Matlab (which is available on a rather beefy Solaris server anyway, so why are we paying for a site license? Because the Windows desktops don't handle X11 tunneling very well). Counting the various servers and other systems that we are running, most of which could be replaced fairly easily with an open source system, the savings climb way up into the millions.
Which is, while true, a misleading statement. Yes, there is poor binary portability between different distros, but that is because they are different operating systems, and just happen to have extremely similar (or sometimes identical) kernels. There is actually very good binary compatibility between my computer running Fedora 8, my girlfriend's computer running Fedora 8, and my uncle's computer running RHEL 4. In fact, binary packages for RHEL 4 can be run in the latest Fedora, RHEL, or CentOS, as long as you have the compat libraries installed (e.g. libstdc++5).
The moral of the story? Unless you expect binary compatibility between Windows and FreeBSD, don't expect binary compatibility between different Linux distros. Different distros are different OS's, with different goals and different target markets or applications, and very often different philosophies.
My guess is that the situation in the middle east is similar. This was a major link, and when it went down, the traffic spilled over to other links and overloaded them, causing the disruption in service. Then again, I don't work in the middle east, so I don't really know how they set things up...
The music thing also makes sense; let OEMs figure out how they want the system to work. Maybe a workstation producer won't want to include all the music and movie stuff, but an entertainment system should have it. Again, different needs in different key markets. Fedora is doing that with its "spins," there is no reason Microsoft can't do it.
The goal is to defeat the RIAA, not concede on the DRM issue. If the RIAA is unwilling to sell DRM-free music, then we just need to stop buying their music, and start listening to the music of independent artists who use open formats like OGG. The danger of DRM is underscored by this sort of behavior -- hard-coding maintenance programs to ignore certain processes? We've been down this road before.
What they should do is require the meta tag for nonstandard rendering, and make standard rendering the default. Web developers shouldn't be encouraged to ignore standards.
Unfortunately, trying to make encryption both secure and transparent is not an easy task. A good encryption system encrypts the private key itself; that requires the user to enter a password to use the encryption at all. Even that is more than most end users are willing to put up with.
Please, most people don't have any knowledge of email encryption, and probably have no idea about proposals like these. Worse, even people who are aware of email encryption don't want to be hassled with it, because it involves typing in an extra password (and heaven forbid we should do anything like that). Sadly, privacy is at the bottom of most people's priority lists, just under "free speech" and "due process" (yes, most of the people I know think it is wrong for a lawyer to defend someone who is "obviously guilty"). This won't scare anyone to action. Encryption should have become popular after several policies were enacted, and it still hasn't. The problem is very simple: most people do not care enough to go to the trouble of encryption.
Oh well, we haven't encouraged parents to actually speak to their kids about this stuff for a long time, opting to shield children from anything deemed harmful by anyone.
This is just another attempt by some politicians to claim that they are fighting to protect our children. Later on, when nobody actually remembers any of this, these politicians can tell a cheering crowd, "I worked hard to give parents the ability to limit their child's MySpace access, and help shield their children from sexual predators online." Of course it is idiotic, and children will find a way around it in less than a minute, but if this were really about protecting our children, it would be an educational program, not another pathetic attempt at technical measures to block their access.
What these 49 states should be doing (as should the 1 state not participating) is starting a program to educate children and their parents about the risks of social networking websites, and how to handle situations where a predator might contact a child on such websites.
"The business model is centered around charging businesses to get help from a Red Hat employee with setting up or troubleshooting their software, so it is in Red Hat's best interests to give the software out for free. Of course, for people like us, that just means getting free, legal software to use."
People are very open to that idea, because it makes perfect sense to them. Granted, RHEL itself is not free, but that's the sort of explanation people are responsive to (and it is true) and it helps to build trust in free (as in beer) software. The idea that open source software is inherently of a lower quality still exists, but the way to show that that is not the case is to mention companies like Red Hat, Mandriva, and Canonical, who have made a lot of money giving all this stuff away.
RTFA. The document in question was a collage, with a picture of the university president, a bulldozer, and some text referencing the "memorial garage."
We actually had a discussion here at FUDCon about this very issue yesterday, and one of the real problems most people encountered wasn't generated interest as much as it was keeping people interested when they encounter little bugs or usability issues (there are quite a few in the Fedora utils which will be resolved this year). Unfortunately, polishing off apps to make them more usable (or even just having them update the UI during a complex operation) is not a high salience issue, and so such things often get neglected in open source projects, even though they can be a deciding factor in keeping non-programmers interested.
That's a very big stretch. That statement could be read in another, more likely and more innocuous way, that the president of the university wanted the garage named for him (I guess there weren't any other buildings left). It hardly seems to be a threat, and you would need counseling yourself if you started walking around with plain-clothed policemen because you thought that the collage was a threatening document. The threat was to this president's plan to build the garage, and so he just found a clever way to rid himself of that "problem."
It was easy to call bullshit, since we already had a system for that. More to the point, using people's fear of a lunatic going on a shooting rampage to justify ludicrous measures like my school's TV's or this George school expelling this student is a disgrace.
...that just about says it all.
That depends on what you are doing. If you are doing CPU-bound stuff, Linux wins for wasting fewer CPU cycles on bouncy icons...
"Linux" is not a single operating system, it is just a kernel. The kernel can be run without GNU utils, without X11, etc.
Here's a problem for you: patent a keyboard that can render all the strange facets of written English: upper case letters, lower case letters, diacritical markings, punctuation, etc. Come back to me when coding language rules is considered to be innovation.
Must I pull a Stallman and point out that Macs are Personal Computers (PCs)?
Ubuntu's main selling point was that it was easy to set up and use, that it "just worked," and that it had some fun media programs. Fedora 8 has all those strong points, and a few features that are of interest to people who know about computing (the way daemons and applets interact, for example).
This is probably innocuous, but who knows? This is what happens when software is distributed only in binary form -- users pick up on something suspicious, and start to assume the worst, and then that whole mob mentality enters the picture. Suppose you discovered that GIMP was connecting to a strange looking host -- what would you do then?