Actually, he can use BitKeeper. He'll just have to pay for it.
This new EULA only affects the free version of BitKeeper.
And it makes perfect sense: why should a company develop a piece of software, and give a limited (???) version away for free in hopes of people paying for the full version, only to allow people to use the free version to create competing software?
There are movies with a $200 million US box office that have "lost money."
Let's say you make a movie that cost $100 million. It grosses $200 million. You just made a mint, right? Wrong. The theaters keep 50% of that, automatically bringing you down to $100 million. If there are any production companies involved, split the $100 million like so (assuming that a)the studio is acting as the distributor and isn't taking an extra cut because of it): 50% goes to the studio, everybody else divides up the rest. If any big name actors are getting money "off the back end", that means their percentage is coming out of the $200 million gross, not the $100 million net. And it comes out of YOUR money, not the studio's.
The problem is just that there are SO MANY MIDDLEMEN. The theater gets 50%. The distributor takes the other 50%, gives themselves 50% of that (unless the distributor is also the studio that made the movie), and gives the rest to whoever is left. Nevermind actors getting money off the back end.
I once read that the only thing holding up Indianna Jones #4 was that Paramount couldn't figure out how they were supposed to make money off the film. All the actors in major roles were demanding huge salaries and a huge percentage of the movie's gross, as well as Spielberg's percentage of the gross, as well as the actual cost of the movie, and there was almost no money left over for Paramount. Luckily they seemed to have solved the problem, and Indianna Jones #4 is in the pipeline.
When a DVD is released most if not all of the production cost has all ready been made back.
That is so not true.
Your average movie needs to rake in at the box office much more than what it cost to make just to break even. How do I come to this conclusion?
Let's say, for example, you make a movie, and shop it around, and some company agrees to distribute the movie. Now, unless the company is run by idiots, they will 1)demand 50% of the money you make, and 2)demand that their costs get covered FIRST (i.e. they get paid back the cost of making the prints, promotional materials and other advertising, and getting it all out to theaters, before you even get so much as a dime).
Now, let's pretend your movie's budget was $10 million, and that the distributor likes it so much (or you're such a good negotiator) that they've agreed to roll the costs of prints, advertising, etc., into their 50%. When released, your movie makes $40 million. You're a huge success now, right? Maybe.
$40 million - 50% for the exhibitor (aka the movie theater) = $20 million. $20 million - 50% for the distributor = $10 million left for you. Congrats, you broke even.
Let's not forget that more movies are being made now than 10 years ago, and that for most movies the box office gross decreases by approximately 50% each week, and the result is that very few movies do much more than break even, let alone become huge successes.
Actually, I've always liked RedHat (the other distros I've used are Mandrake, Slackware, and TurboLinux, as well as all 3 BSDs). But I really like KDE. Not just the apps, but the whole shebang, environment and all.
I am definately thinking of giving Mandrake 9.0 a try. A unified desktop is NOT what I want. I want KDE including the KDE environment. I want a choice. And it looks like my choice from now on is going to be Mandrake.
I bow down to your superior knowledge of high-end SGI systems.
But still, one must admit, that if this system the article talks about had cool graphics hardware to match, it would be interesting to know just how many FPSes one could get from Quake 3 (optimized to take maximum advantage of the hardware, of course).
I'm sorry, but you don't buy an old SGI so that you can run Linux. If you want an old UNIX workstation to run Linux on, get an old Sun.
I use Linux on my PC, but run IRIX on my Indy. Unlike Linux, IRIX has been specifically designed for SGI hardware. The only way this could be any more of a travesty is if you ran Linux on an O2 or an Octane, because there's no way in hell Linux takes advantage of the custom hardware in those machines.
I'm not trying to flame or troll or whatever, it just disappoints me that there are people who'd go to the trouble of getting an SGI, only to put Linux on it. The "there was no OS installed" argument falls flat, because for every system on eBay that comes without a hard disk or no OS media, there's one for a bit more that comes with the OS media (which is the only way you can get IRIX and still have SGI approve; SGI's approval means you don't have to pay $600 to get a new license, because there will already be a license to run IRIX on that particular machine; you need a license to get access to the functional updates, as only the mainenance updates are provided for free).
Half the fun of having an old SGI is running IRIX on it. Sure, IRIX might be quirky, but it's still part of the "experience".
Actually, 3D rendering is not the kind of thing you'd use a system like this for. In fact, 3D rendering is better (aka faster) when done on a lot of separate, but networked, systems. You hand a dual-CPU box a frame (or part of a frame) and say, "Render this." and the box chomps away on that frame for however long it takes.
On a system like this, it'd be much better for something where you needed to process a lot of data (tens of gigabytes, all needing to be stored in memory if you want the processing to be completed this century). Things like weather prediction, nuclear test simulation, mechanical engineering, genetics, etc., for all of which the speed of the memory system is just as important as the speed of the CPU. Add to that the fact that there are a lot of CPUs, and the speed of the memory system becomes vital.
SGI's banishing of the traditional system bus is a godsend. Each CPU can get the same memory bandwidth it would get if it was the only CPU in the box. Even SGI's workstations have done this, starting with the Octane (via a crossbar switch), where each subsystem (graphics, memory, whatever) could independently access the other systems at maximum speed, and were guaranteed that bandwidth. Video capture cards could go straight from the video capture card to the hard disk if both devices had enough intelligence, bypassing the CPU and not even touching RAM, meanwhile the CPU could be processing some stuff in RAM, and both would take place at maximum speed.
Actually, the STREAM TRIAD test is not synthetic or irrelevant. The target audience for a machine like this is one where the data to be operated on is probably several tens or even hundreds of gigabytes in size (weather prediction, nuclear test simulation, etc.). For this type of application, CPU performance is important, but memory performance is equally important, because having a blazingly fast CPU means nothing if that CPU can't access the data it is supposed to process at a reasonable speed.
As far as system bus (aka memory bus) goes, the system bus is very often the biggest bottleneck, especially on multiprocessor systems, because the CPUs are all competing to use it. By eliminating this bottleneck, the CPUs can all operate at maximum speed (i.e. no CPU time wasted waiting for more data to come in from the memory subsystem, no competing for access to the system bus, etc.). And if you say "Well that's what cache is for!" we'll all know how big of an idiot you are, because you can't fit gigabytes upon gigabytes of data in cache.
Actually, SGI has seen a bit of a rebound lately. While the visual FX industry may be desperate to get rid of their SGI hardware, the visualization (CAD, CAM, medical imaging, etc.), simulation, and to some extent the manufacturing industries (as well as a few others) have shown increasing interest in SGI hardware.
Like another poster said, SGI went from $230 million in losses to $10 million in positive income. That means from -$230 million to +$10 million.
Yeah, it's the same guy. He took two companies with very strong UNIX systems, tried to develop "NT strategies" for each of them, and when those failed he left. I don't think it was much of a coincidence that he later went to work for MS.
"Gee, I'll go work for some of MS's competitors (or potential competitors), and try to get them to use NT. The result will be one of the following: 1)their core business will be trashed and the company will go under, 2)the whole NT thing will work out very well for them, and they will become dependent on Microsoft. Either way, Mr Gates, Microsoft wins." must surely have been his line of thinking, or somebody's (ahem *[insert MS executive here]* ahem) idea.
Problem is, these things don't have any graphics hardware.
If you want cool graphics, along with Origin architecture and speed, you need an Onyx or an Onyx2 (which is really just an Origin w/graphics hardware).
The number of frames per second that the Onyx2 equivalent of this thing would be scary.
To say that you've never heard of it, and because of that it is therefore worthless, is awfully presumptuous.
You can get Plan 9 from CheapBytes.
It was supposed to be the next evolution of UNIX, even created by the guys who came up with UNIX in the first place. But UNIX was too popular, and Plan 9 never really caught on.
But this article seems a bit outdated, or maybe the author has been living in the stone age. Solaris 2.9? 3.0? Unless I'm gravely mistaken, we're at Solaris 9 right now, and I don't see a lot of shops running Plan 9.
I believe (though I could be wrong; I haven't used Redhat 6 in a long time) that the linuxconf program had an option to show you what it was going to do before it went and did it.
Alas, Redhat no longer includes this program, but most recent distros are easier to config than Redhat 6.
Keeping in mind that most config stuff goes in/etc, just read the docs for whatever you want to configure. If your distro is any good, it should include the Linux FAQs and HOWTOs, which greatly help with learning how to configure stuff by hand.
Luckily, one doesn't need to configure their system by hand nearly as much as one used to:)
That's probably part of your problem. In fact, I find that the most prevalent problem among people who switch from MS operating systems to UNIX is that much (if not most) of their DOS/Windows knowledge just doesn't apply. UNIX was around before DOS or Windows, and the result is that the UNIX way is very different from the Windows/DOS way.
My first exposure to UNIX was FreeBSD (in the long ago time of 1998). When I first started using it (all my friends told me I'd have an easier time if I tried Linux first; oh well), I was totally lost. I knew that things were going to be different from Windows and DOS, but I had one major hangup (that I didn't even realize was a hangup): all my old DOS knowledge. Then one day somebody told me "Forget what you know about DOS or Windows. Unlearn it." I haven't had any trouble since then.
Oh, BTW, RedHat 6 had the "linuxconf" (or was it "linuxconfig"?) program, which used ncurses for text menus, or a GUI if X Windows was running. If you really did RTFM, or even browse through the default menus in GNOME or KDE, you would've found this program. With it you could configure almost anything (everything from LILO to Apache).
Well, if you're buying machines for a render farm (which is what he said they were for), they don't need a DVD-burner, Indesign, DVD-authoring software, video editing software, an office package, and loads of other goodies.
You need a system with a barebones install of your chosen OS + whatever you need for networking, and the rendering software you're using.
Every machine in a renderfarm will be headless (i.e. no keyboard, mouse, or monitor attached), except for maybe the machine in charge of queing render jobs and the like. Maybe.
As for finding a dual-athlon system for cheap, remember that machines to be used in renderfarms don't need expensive 3D accelerators, expensive sound cards, massive hard disks (or any, depending on the speed of your ethernet), expensive sound cards. All you need is 1)one or two fast processors, 2)as much RAM as you can afford to put in each box, and 3)a fast ethernet card (gigabit, whatever). And you most certainly do not need 64-bit PCI, or any PCI, since the computers making up your renderfarm will most likely be rackmount, and therefore have the network card integrated into the motherboard.
It is an unfounded myth that developers have to develop for both.
- I use KDE - I run GNOME apps every day
How do I do this? Simple: I have the GNOME libraries and GTK+ installed, along with KDE. This means that I can use GnoRPM instead of Kpackage, The GIMP, etc. I use KDevelop or Kate + an Xterm window for programming, and Konqueror for web browsing.
And I never write programs specifically for either desktop. I use GTK+ or FLTK for all my GUI programming needs (yes, that means my programs could theoretically run on Windows; I haven't tried it, though).
As to why RedHat seems to blindly hate KDE, I think it is because it doesn't have "GNU" anywhere in the title;)
...is so much crap. They're just trying to sell their book ("We discuss this matter further in our book" blah blah blah).
As for the whole CSS/DHTML/XHTML/SuperHTML/HTML2000, get over it. While I agree that some designers go overboard when a simple <h2> tag would do, using CSS or whatever does not automagically guarantee correct code. For every site that uses CSS properly, there are 2 that use it incorrectly.
But that is the thing about this article: on the one hand it argues that people should strive for compliance, then on the other argues that people should only write webpages for the latest whiz-bang browser.
In my oppinion, the best way to eliminate nonstandard HTML: if you see it, e-mail the site's webmaster, preferrably with the "correct" HTML to fix it.
But that leads my rant to another point: what is so wrong with using plain-old vanilla HTML? If all browsers being standards-compliant is so important, why does the article argue that we should use CSS and "correct" JavaScript to detect for individual browsers?
It orbits the sun in a horseshoe-shaped orbit that goes quite a distance above and below the elliptical plane. The horseshoe orbits overlap, but don't take 770 years. However, because these horseshoe-shaped orbits overlap, Cruithne eventually goes all the way around the sun. It is this that takes 770 years.
It is affected by Earth's gravity (indeed, that is why its orbit is shaped like a horseshoe), but that doesn't make it a satellite of Earth.
Therefore, this "third" moon is actually the second.:b
Also remember that on the moon, it would be very bright. To compensate, the astronauts would have had to close down the aperture (the little part of the lens that controls how much light enters the camera). Thus, the stars are less likely to show up.
Actually, they did it because international treaties state that if a company from Country A launches something into space, regardless of where it is launched from, the government of Country A is financially responsible for any damage caused, i.e. the rocket goes off course and lands on somebody's house over in Country B.
Also, people want to control the amount of space debris, because early on, nobody really worried about it, and it has now become a problem. There are, in fact, people whose job is to track space debris, and if it is going to collide with something important (space shuttle/ISS/satellite/top secret death ray) make sure whatever it was that was so important gets out of the way. A company would not be very happy if they spent millions to launch a satellite, only to have it get knocked out of existence by space junk striking it at a few thousand meters per second.
And let's not forget the export controls. The company would need the US gov'ts approval to get the damn thing outside of US borders in the first place. Just imagine the heads that would roll if some company's multimillion dollar space vehicle was impounded by customs on its way to the launch site.
This is because the parts have to be certified for space. You can't just go to RadioShack and get some 7400s off the shelf, because if cosmic rays flipped a few bits or tripped a few lines, the result could kill the half-dozen astronauts on the shuttle.
Slashdot Mirror
No. Pure 16-bit addressing means you can have, at most, 64 KB of RAM (65536 bytes). 24-bit addressing results in a max. of 16 MB of RAM.
2^16 = 65536 (64 KB)
2^24 = 16777216 (16 MB)
As for 8-bit CPUs, most used 16-bit addressing because having a maximum of 256 bytes of RAM would be quite limiting.
Actually, he can use BitKeeper. He'll just have to pay for it.
This new EULA only affects the free version of BitKeeper.
And it makes perfect sense: why should a company develop a piece of software, and give a limited (???) version away for free in hopes of people paying for the full version, only to allow people to use the free version to create competing software?
All the salt does is lower the freezing temperature a few degrees, not eliminate it.
IIRC, the water at the bottom of the ocean is well below surface freezing temperature.
Except that when you put water under enough pressure, it won't freeze. That is why the bottom of the ocean isn't frozen.
There are movies with a $200 million US box office that have "lost money."
Let's say you make a movie that cost $100 million. It grosses $200 million. You just made a mint, right? Wrong. The theaters keep 50% of that, automatically bringing you down to $100 million. If there are any production companies involved, split the $100 million like so (assuming that a)the studio is acting as the distributor and isn't taking an extra cut because of it): 50% goes to the studio, everybody else divides up the rest. If any big name actors are getting money "off the back end", that means their percentage is coming out of the $200 million gross, not the $100 million net. And it comes out of YOUR money, not the studio's.
The problem is just that there are SO MANY MIDDLEMEN. The theater gets 50%. The distributor takes the other 50%, gives themselves 50% of that (unless the distributor is also the studio that made the movie), and gives the rest to whoever is left. Nevermind actors getting money off the back end.
I once read that the only thing holding up Indianna Jones #4 was that Paramount couldn't figure out how they were supposed to make money off the film. All the actors in major roles were demanding huge salaries and a huge percentage of the movie's gross, as well as Spielberg's percentage of the gross, as well as the actual cost of the movie, and there was almost no money left over for Paramount. Luckily they seemed to have solved the problem, and Indianna Jones #4 is in the pipeline.
When a DVD is released most if not all of the production cost has all ready been made back.
That is so not true.
Your average movie needs to rake in at the box office much more than what it cost to make just to break even. How do I come to this conclusion?
Let's say, for example, you make a movie, and shop it around, and some company agrees to distribute the movie. Now, unless the company is run by idiots, they will 1)demand 50% of the money you make, and 2)demand that their costs get covered FIRST (i.e. they get paid back the cost of making the prints, promotional materials and other advertising, and getting it all out to theaters, before you even get so much as a dime).
Now, let's pretend your movie's budget was $10 million, and that the distributor likes it so much (or you're such a good negotiator) that they've agreed to roll the costs of prints, advertising, etc., into their 50%. When released, your movie makes $40 million. You're a huge success now, right? Maybe.
$40 million - 50% for the exhibitor (aka the movie theater) = $20 million.
$20 million - 50% for the distributor = $10 million left for you. Congrats, you broke even.
Let's not forget that more movies are being made now than 10 years ago, and that for most movies the box office gross decreases by approximately 50% each week, and the result is that very few movies do much more than break even, let alone become huge successes.
Actually, I've always liked RedHat (the other distros I've used are Mandrake, Slackware, and TurboLinux, as well as all 3 BSDs). But I really like KDE. Not just the apps, but the whole shebang, environment and all.
I am definately thinking of giving Mandrake 9.0 a try. A unified desktop is NOT what I want. I want KDE including the KDE environment. I want a choice. And it looks like my choice from now on is going to be Mandrake.
I bow down to your superior knowledge of high-end SGI systems.
But still, one must admit, that if this system the article talks about had cool graphics hardware to match, it would be interesting to know just how many FPSes one could get from Quake 3 (optimized to take maximum advantage of the hardware, of course).
I'm sorry, but you don't buy an old SGI so that you can run Linux. If you want an old UNIX workstation to run Linux on, get an old Sun.
I use Linux on my PC, but run IRIX on my Indy. Unlike Linux, IRIX has been specifically designed for SGI hardware. The only way this could be any more of a travesty is if you ran Linux on an O2 or an Octane, because there's no way in hell Linux takes advantage of the custom hardware in those machines.
I'm not trying to flame or troll or whatever, it just disappoints me that there are people who'd go to the trouble of getting an SGI, only to put Linux on it. The "there was no OS installed" argument falls flat, because for every system on eBay that comes without a hard disk or no OS media, there's one for a bit more that comes with the OS media (which is the only way you can get IRIX and still have SGI approve; SGI's approval means you don't have to pay $600 to get a new license, because there will already be a license to run IRIX on that particular machine; you need a license to get access to the functional updates, as only the mainenance updates are provided for free).
Half the fun of having an old SGI is running IRIX on it. Sure, IRIX might be quirky, but it's still part of the "experience".
Actually, 3D rendering is not the kind of thing you'd use a system like this for. In fact, 3D rendering is better (aka faster) when done on a lot of separate, but networked, systems. You hand a dual-CPU box a frame (or part of a frame) and say, "Render this." and the box chomps away on that frame for however long it takes.
On a system like this, it'd be much better for something where you needed to process a lot of data (tens of gigabytes, all needing to be stored in memory if you want the processing to be completed this century). Things like weather prediction, nuclear test simulation, mechanical engineering, genetics, etc., for all of which the speed of the memory system is just as important as the speed of the CPU. Add to that the fact that there are a lot of CPUs, and the speed of the memory system becomes vital.
SGI's banishing of the traditional system bus is a godsend. Each CPU can get the same memory bandwidth it would get if it was the only CPU in the box. Even SGI's workstations have done this, starting with the Octane (via a crossbar switch), where each subsystem (graphics, memory, whatever) could independently access the other systems at maximum speed, and were guaranteed that bandwidth. Video capture cards could go straight from the video capture card to the hard disk if both devices had enough intelligence, bypassing the CPU and not even touching RAM, meanwhile the CPU could be processing some stuff in RAM, and both would take place at maximum speed.
Actually, the STREAM TRIAD test is not synthetic or irrelevant. The target audience for a machine like this is one where the data to be operated on is probably several tens or even hundreds of gigabytes in size (weather prediction, nuclear test simulation, etc.). For this type of application, CPU performance is important, but memory performance is equally important, because having a blazingly fast CPU means nothing if that CPU can't access the data it is supposed to process at a reasonable speed.
As far as system bus (aka memory bus) goes, the system bus is very often the biggest bottleneck, especially on multiprocessor systems, because the CPUs are all competing to use it. By eliminating this bottleneck, the CPUs can all operate at maximum speed (i.e. no CPU time wasted waiting for more data to come in from the memory subsystem, no competing for access to the system bus, etc.). And if you say "Well that's what cache is for!" we'll all know how big of an idiot you are, because you can't fit gigabytes upon gigabytes of data in cache.
Actually, SGI has seen a bit of a rebound lately. While the visual FX industry may be desperate to get rid of their SGI hardware, the visualization (CAD, CAM, medical imaging, etc.), simulation, and to some extent the manufacturing industries (as well as a few others) have shown increasing interest in SGI hardware.
Like another poster said, SGI went from $230 million in losses to $10 million in positive income. That means from -$230 million to +$10 million.
Yeah, it's the same guy. He took two companies with very strong UNIX systems, tried to develop "NT strategies" for each of them, and when those failed he left. I don't think it was much of a coincidence that he later went to work for MS.
"Gee, I'll go work for some of MS's competitors (or potential competitors), and try to get them to use NT. The result will be one of the following: 1)their core business will be trashed and the company will go under, 2)the whole NT thing will work out very well for them, and they will become dependent on Microsoft. Either way, Mr Gates, Microsoft wins." must surely have been his line of thinking, or somebody's (ahem *[insert MS executive here]* ahem) idea.
It's only too funny that MS canned him.
Problem is, these things don't have any graphics hardware.
If you want cool graphics, along with Origin architecture and speed, you need an Onyx or an Onyx2 (which is really just an Origin w/graphics hardware).
The number of frames per second that the Onyx2 equivalent of this thing would be scary.
Plan 9 is an operating system.
To say that you've never heard of it, and because of that it is therefore worthless, is awfully presumptuous.
You can get Plan 9 from CheapBytes.
It was supposed to be the next evolution of UNIX, even created by the guys who came up with UNIX in the first place. But UNIX was too popular, and Plan 9 never really caught on.
But this article seems a bit outdated, or maybe the author has been living in the stone age. Solaris 2.9? 3.0? Unless I'm gravely mistaken, we're at Solaris 9 right now, and I don't see a lot of shops running Plan 9.
I believe (though I could be wrong; I haven't used Redhat 6 in a long time) that the linuxconf program had an option to show you what it was going to do before it went and did it.
/etc, just read the docs for whatever you want to configure. If your distro is any good, it should include the Linux FAQs and HOWTOs, which greatly help with learning how to configure stuff by hand.
:)
Alas, Redhat no longer includes this program, but most recent distros are easier to config than Redhat 6.
Keeping in mind that most config stuff goes in
Luckily, one doesn't need to configure their system by hand nearly as much as one used to
I'm originally a DOS type...
That's probably part of your problem. In fact, I find that the most prevalent problem among people who switch from MS operating systems to UNIX is that much (if not most) of their DOS/Windows knowledge just doesn't apply. UNIX was around before DOS or Windows, and the result is that the UNIX way is very different from the Windows/DOS way.
My first exposure to UNIX was FreeBSD (in the long ago time of 1998). When I first started using it (all my friends told me I'd have an easier time if I tried Linux first; oh well), I was totally lost. I knew that things were going to be different from Windows and DOS, but I had one major hangup (that I didn't even realize was a hangup): all my old DOS knowledge. Then one day somebody told me "Forget what you know about DOS or Windows. Unlearn it." I haven't had any trouble since then.
Oh, BTW, RedHat 6 had the "linuxconf" (or was it "linuxconfig"?) program, which used ncurses for text menus, or a GUI if X Windows was running. If you really did RTFM, or even browse through the default menus in GNOME or KDE, you would've found this program. With it you could configure almost anything (everything from LILO to Apache).
Well, if you're buying machines for a render farm (which is what he said they were for), they don't need a DVD-burner, Indesign, DVD-authoring software, video editing software, an office package, and loads of other goodies.
You need a system with a barebones install of your chosen OS + whatever you need for networking, and the rendering software you're using.
Every machine in a renderfarm will be headless (i.e. no keyboard, mouse, or monitor attached), except for maybe the machine in charge of queing render jobs and the like. Maybe.
As for finding a dual-athlon system for cheap, remember that machines to be used in renderfarms don't need expensive 3D accelerators, expensive sound cards, massive hard disks (or any, depending on the speed of your ethernet), expensive sound cards. All you need is 1)one or two fast processors, 2)as much RAM as you can afford to put in each box, and 3)a fast ethernet card (gigabit, whatever). And you most certainly do not need 64-bit PCI, or any PCI, since the computers making up your renderfarm will most likely be rackmount, and therefore have the network card integrated into the motherboard.
Actually, Photoshop is widely known to be more heavily optimized for Macs than for PCs, nevermind the plugins.
It is an unfounded myth that developers have to develop for both.
;)
- I use KDE
- I run GNOME apps every day
How do I do this? Simple: I have the GNOME libraries and GTK+ installed, along with KDE. This means that I can use GnoRPM instead of Kpackage, The GIMP, etc. I use KDevelop or Kate + an Xterm window for programming, and Konqueror for web browsing.
And I never write programs specifically for either desktop. I use GTK+ or FLTK for all my GUI programming needs (yes, that means my programs could theoretically run on Windows; I haven't tried it, though).
As to why RedHat seems to blindly hate KDE, I think it is because it doesn't have "GNU" anywhere in the title
...is so much crap. They're just trying to sell their book ("We discuss this matter further in our book" blah blah blah).
As for the whole CSS/DHTML/XHTML/SuperHTML/HTML2000, get over it. While I agree that some designers go overboard when a simple <h2> tag would do, using CSS or whatever does not automagically guarantee correct code. For every site that uses CSS properly, there are 2 that use it incorrectly.
But that is the thing about this article: on the one hand it argues that people should strive for compliance, then on the other argues that people should only write webpages for the latest whiz-bang browser.
In my oppinion, the best way to eliminate nonstandard HTML: if you see it, e-mail the site's webmaster, preferrably with the "correct" HTML to fix it.
But that leads my rant to another point: what is so wrong with using plain-old vanilla HTML? If all browsers being standards-compliant is so important, why does the article argue that we should use CSS and "correct" JavaScript to detect for individual browsers?
As for our second alleged moon: it is not a moon!
:b
It orbits the sun in a horseshoe-shaped orbit that goes quite a distance above and below the elliptical plane. The horseshoe orbits overlap, but don't take 770 years. However, because these horseshoe-shaped orbits overlap, Cruithne eventually goes all the way around the sun. It is this that takes 770 years.
It is affected by Earth's gravity (indeed, that is why its orbit is shaped like a horseshoe), but that doesn't make it a satellite of Earth.
Therefore, this "third" moon is actually the second.
Also remember that on the moon, it would be very bright. To compensate, the astronauts would have had to close down the aperture (the little part of the lens that controls how much light enters the camera). Thus, the stars are less likely to show up.
Actually, they did it because international treaties state that if a company from Country A launches something into space, regardless of where it is launched from, the government of Country A is financially responsible for any damage caused, i.e. the rocket goes off course and lands on somebody's house over in Country B.
Also, people want to control the amount of space debris, because early on, nobody really worried about it, and it has now become a problem. There are, in fact, people whose job is to track space debris, and if it is going to collide with something important (space shuttle/ISS/satellite/top secret death ray) make sure whatever it was that was so important gets out of the way. A company would not be very happy if they spent millions to launch a satellite, only to have it get knocked out of existence by space junk striking it at a few thousand meters per second.
And let's not forget the export controls. The company would need the US gov'ts approval to get the damn thing outside of US borders in the first place. Just imagine the heads that would roll if some company's multimillion dollar space vehicle was impounded by customs on its way to the launch site.
This is because the parts have to be certified for space. You can't just go to RadioShack and get some 7400s off the shelf, because if cosmic rays flipped a few bits or tripped a few lines, the result could kill the half-dozen astronauts on the shuttle.