That's a great analogy, but it doesn't fully answer the question. The TOC is going to be higher.
I assume you meant the TCO, the Total Cost of Ownership.
It's not necessarily going to be higher. Direct support and training costs should be higher, although not necessarily by all that much. However, if each OS is being used for a legitimate reason, and the employees that are going to be affected are properly trained, there may be gains in other areas to balance it back out, or even make the net affect to lower the TCO. So, no, the TCO is not always going to be higher because of diversity.
Isn't there some sort of torque problem that needs to be addressed with these hard drive-based devices? Have they engineered the things so that they don't twitch in your backpack?
Actually if you understood elementary physics you'd not have asked that. The platters are gyros, the effect is not a twitch, but actually a slight resistance to being moved. They do, of course, 'twitch' very slightly when the power is turned on.
The real problem is that it's difficult to make the other parts move as close to those platters as they need to, without being so close they 'crash' into the platters when you bump something.
Thanks for the link. Nice to know those binaries are still out there - I like Opera 6.03 great, but it could still be very handy to have access to 3.62 - an HTML 3.2 compliant browser that will fit on a floppy and run on win16 could come in very handy.
Win 16 means it's compatible with not just win3.1 but also even very old versions of OS/2, WINE will run it perfectly (I know, I was running it on a dual boot Win 3.11/Slackware box for quite awhile, it was very handy, the same binaries running from the same directory, with the same settings files regardless of OS.)
It was NOT introduced in version 4. I started using Opera at version 3 point something, 3.6 I believe it was, and it had it then. It's had it since the earliest alpha version if I'm not mistaken. What was added recently was the ability to turn it off, for the whiners out there that complained about it endlessly.
Not quite true. Xenix was based on the AT&T source - the SysIII source. AT&T didn't license the Unix name at the time, just the code - that's why it was called Xenix instead of MS Unix to begin with. SCO updated it with Berkeley components, SysV compatibility, and so forth.
The part you are correct on, however, is that SCO Unix reflects a new (in '89) license from AT&T, which included rights to the name Unix as well as the actual SysV code.
Just in case you didn't realise it, that MicroSoft Xenix is actually the same thing as SCO Unix (well, if it carried the Xenix name it was an earlier version, but the same codeline, the same unix.) SCO was Microsofts outsourcer - MS didn't actually write Xenix, they just bought the license and hired SCO to do it. Eventually MS lost interest in Xenix, so SCO bought the license and renamed it SCO Unix...
I doubt it very much. People that buy multiprocessor Sun systems are used to paying, and most probably won't blink at those prices. They have to give away the single processor version to compete at all - Linux and BSD are very capable competitors on that hardware, and they're free after all. But Solaris still has the advantage on their multiproc boxen, so people that need that kind of performance will pay for it.
Well the first time I installed linux myself it was RedHat. The init system was so bizzarely convoluted I suspected it was invented simply to torture me;) I actually liked rpm at the time, it's certainly a lot better than windows install shield crap. But it's needlessly complicated, and when it goes haywire (normally because of poor packaging I am told by rpm advocates - I don't doubt that's true, but then again it's needlessly difficult to package things properly under it too) it can be really ugly.
After installing slack and setting it up I could never go back to an rpm distro. The init system makes sense! The package system makes sense!
With redhat when something went wrong I could never be sure why until I spent the time tracking it down, and it always turned out to be something different and strange each time, sometimes my fault, but often not. When slack screws up, it's always one of two things - either there is a hardware problem or I screwed something up myself. After about a year I started having confidence in that, and that confidence has always turned out to be well founded. It makes troubleshooting so much easier.
Throw in the towel on this one, trust me. Slackware's package system is archaic and nearly useless compared to something like RPM.
"Archaic" perhaps,.tgz is a basic standard that can be relied upon nonetheless, and more than adequate for most uses. What it lacks in fancy features it makes up for simplicity and reliability.
None of the big players release their stuff in pkg format anymore. You're looking at either source, a.tar.gz format that has an installer inside, or an RPM.
Or waiting a day or two until another slacker gets the source and generates a binary pkg against the current stable release. Between slackware.com and the various user sites like linuxmafia.com and linuxpackages.net most any needed software can be installed from packages, and there is a very high level of quality maintained in the compilation.
To change configuration in any popular Linux distro you can edit text files.
Exactly! And even after stating it you never understand it. Slackware does things in very standard ways, so instead of learning the peculiarities of a particular distribution, you just learn solid standard *nix.
The other thing I'd argue is that if people were to learn about Linux they should learn about the Linux Standards Base - things like SysV init scripts, the FHS, and the RPM packaging system (although the FHS is still waiting on the RPM 4 version of Maximum RPM so v3 is still the written standard).
The LSB is just wind. Slackware is the living standard that you can install on your machine, run, use, tinker with.
SysVinit is just annoying. RPM *shrug* you can install it on slack if you need it, knock yourself out.
I just deselected the spyware and it installed happily without it, I hardly see why including it for those that don't mind it means they deserve to be skewered by the RIAA. It's been a great service, I'll be very sorry to see it go.
If you really think bothering to check and see when a program is on instead of choosing it by name at the UI level you probably run windows and take your car to the shop whenever you have a problem. Some of us are comfortable with the command line, own a wrench, and are perfectly capable of working a VCR without a special interface. For us, the 'service' is very little help and a major source of annoyance. I think he knows exactly what he wants, and the last thing he should do is listen to you.
(And no, I don't approve of this whole business of recording special programs without your consent and not letting you delete them, but that's a separate issue [from] the reserve space.)
Actually it isn't. The recording is done using the reserve space, so it's the same issue.
Frankly I think you're both half right - he's right that the space is his to use as he wants, and you are right that he shouldn't be surprised that they use it the way they designed it to be used. He needs to unplug it and start figuring out how to hack it.
I don't think it's so crazy either, except that RMS is going after developers and geeks who already *know* the relationship that GNU has with Linux. He's not trying to persuade the distribution creators to change their names (to Red Hat GNU/Linux, for example). Instead he chastises a LUG... his own fans.
He's hardly 'going after them' - he is declining their invitation and stating his reason, and sticking to it. He has no shortage of things to do, and this is one of his well known requirements and has been for a long time. If they are indeed his fans as you say, why do they not simply use GNU/Linux as the written long form as he asks?
As I say, he has no lack for things to do, and his policy is well known, if they want him to speak they know what they have to do. He has no obligation to speak there.
On this page [gnu.org], Stallman would have everyone believe that the Linux kernel would never have even been written if it weren't for GNU. This is patently false. Torvalds designed the first stages of the Linux kernel on top of Minix, largely using Minix tools. But he knew all along that he wanted to get away from Minix and at the time, GNU had the most complete set of libraries and utilities available for free. (In fact, if I recall, it wasn't even Linus who started porting the GNU stuff over to Linux, it was other developers in it just for the hacking.)
You are partially correct, it's possible that Linus would have created a kernel anyway, but what we call Linux today certainly would not exist. Minix was not free, and it was quite frankly the most useless unix-inspired OS ever made. Without GCC it seems quite probable the whole project would never have flown in any form. Even BSD uses GCC.
Just because GNU tools have been ported to almost every OS doesn't really mean that they're being used on almost every OS.
Of course it does! Who the hell do you think ports Free Software to new platforms? People who need to use that software on that platform. It's nonsense to think that people port programs when there is no demand for them.
BSD offers GNU tools in ports, but they don't come with any BSD by default. Most BSD users just stick with the BSD tools, I think.
So if the GNU users aren't the majority on the platform they don't exist? Huh?
Again, who do you think ported the programs and why? Do you have any clue how Free Software works?
The arguement I like best is - if it's GNU/Linux, why isn't it GNU/BSD?
Why the heck would it be? You're confused. GNU's Not Unix. Linux Is Not UniX.
BSD is Unix.
That help?
Is that perhaps because BSD came out before GNU?? Hmmm - inquiring minds want to know?
BSD did exist before GNU/Linux, but it was not Free back then.
How much of GNU was based on BSD if any?
Quite little actually. Remember, BSD wasn't Free in the beginning. Had it been, a lot of stuff would never have been written, almost certainly including Linux itself. People don't rewrite an entire production quality OS from the ground up just for kicks.
We know RMS doesn't like the one true editor VI so he had to come up with something else, but beyond that???
My personal opinion is that RMS is nuts. With the crusade he's on, I'm amazed that newer versions of the GPL don't include a clause that say something to the effect of "Any derivitives or bundled software included with this product must be prefixed by the recursive acronym GNU."
I don't think you understand the man at all. He doesn't want anything that isn't GNU to be called GNU. He does think that when you have the various distributions of 'Linux' which are bundles of the linux kernel and a bunch of other Free Software, a good portion GNU, you should call it GNU/LINUX. I don't think that's so crazy.
GNU did start the ball rolling, and they did a lot of crucial and necessary work along the way, and they're still doing important work today. It was Stallman's dream alone not so long ago that there would be an entire Free Operating System (and not just one!) that has not only caught but in many ways beaten the commercial Unixes that had only a little earlier accustomed the world to commercial software. He was laughed at, but instead of worrying he wrote code, and pursuaded people to write code, or to donate money to pay people to write code. Crucial tools without which a Free kernel could not be written - shells, fileutilities, compilers, text editors. The infrastructure of the Operating System.
GNU's Not Unix. Linux Is Not UniX.
No, RMS, it takes strong bullheadedness to criticize so strongly the only reason your GNU tools are still alive today.
Ignorant nonsense. GNU tools are used on every Unix, on Apple systems, on Microsoft systems, on VMS even. The GNU tools literally made Linux possible, and Linux is not the only kernel on which the GNU tools and other Free Software can be combined to produce a complete functional and completely Free system.
Although I agree with almost everything you said, I disagree very much with your conclusion. Slashdot rocks. Why? Because of posters like you, who take the time to write long, well-phrased articles about things like 'why slashdot sucks' or, well, why 'slashdot sucks' perhaps.
It's not because of the goatse.cz guy, or the penis bird, or Natalie Porter Nude and Petrified even (although I'll admit all of those were funny once, and the last thought can still send shivers down my spine if it catches me unawares.)
Anyway, I don't think the point to the site was ever to be some sort of hallowed highbrow thing. Yes, it's pitiful the average intelligence they're catering to now, but it's still a little better than many 'mainstream' sites, and a lot of interesting stuff does get posted. A terribly large dose of bullshit too, of course, but ese es el vive, no?
For anything more than a trivial program, any sort of proof of correctness, for all possible inputs, is pretty impossible.
Which is why the key to simple code is small procedures or objects.
Do one thing.
Do it well.
Prove that you do it well.
Move on to the next phase.
It can be done. It does require more planning than some are capable of. And it may not come together as quick as if you just hack what you need. But when you can get away with it, it's very nice.
Re:Sorry, but ARM == No floating point hardware!
on
Transmeta Meets Blades
·
· Score: 2, Informative
Actually I found some specs, SA@600MHz uses 450mW, so you could power 4 for the same price as one Crusoe. Perhaps it's a political thing, I'd certainly rather Transmeta get the business than Intel, but I still don't see the technical justification.
Low power consumption. A 90% power savings is sort of irrelevant with a single CPU, but talk about saving 90% of electricity across 20 CPUs, and that's a decent savings.
Get a clue. The Crusoe consumes about 2 watts. Very nice compared to Pentium-class room heaters, yes, but I asked why they choice Crusoe over StrongARM, not Crusoe over IA-32. A 600mhz SA uses 450mW, so you can run roughly 4 of them for the same power and heat as one Crusoe.
The advantages that Crusoe has are two - first, as I mentioned originally - x86 compatibility. This is not a help for a supercomputer - you're going to be compiling everything from source anyway. The other advantage, that I forgot, is that the SA doesn't have an FPU. That, at least, is a legitimate reason to consider the Crusoe, but I'm still not sure the decision actually makes sense - the SA is a very nice chip and if programmed right it should have no problem keeping up with the Crusoe even on FP, figuring that you can use 4 times as many SAs for the same heat and power requirements.
I've seen the two used almost interchangbly when referring to modern large systems.
I have as well, it's a common mistake for nontechnical types (particularly reporters) to make. But they are very different systems. Mainframes have massive redundancy and i/o bandwidth. Supercomputers also have lots of redundancy but they are typically not built for I/O bandwidth at all, but sheer number crunching power. Mainframes are designed to run large databases, supercomputers to do complex mathematics, so you get very different designs for different problems.
The Origin2000 is, if I'm not mistaken, the latest iteration of some of the old Cray designs, and those are definately Supercomputers, not Mainframes. That said, you are of course absolutely correct that ethernet is a major limitation of the sort of cluster we are talking about, and the Crays are still a much better bet for a subset of traditional supercomputer jobs. This is changing, though, as more and more effort and thought goes into improving them.
Re:Sorry, but ARM == No floating point hardware!
on
Transmeta Meets Blades
·
· Score: 3, Informative
Ok, finally that's a legitimate response. It's true ARM doesn't include an FPU. However, the last I checked (and I'm not real up to date on it) using libfloat it had emulation good enough to keep up with IA32 fairly well on FP.
I imagine, though, this is probably the reason. It seems reasonable that Supercomputer work would require some FP, although I don't know for sure.
Slashdot Mirror
I assume you meant the TCO, the Total Cost of Ownership.
It's not necessarily going to be higher. Direct support and training costs should be higher, although not necessarily by all that much. However, if each OS is being used for a legitimate reason, and the employees that are going to be affected are properly trained, there may be gains in other areas to balance it back out, or even make the net affect to lower the TCO. So, no, the TCO is not always going to be higher because of diversity.
Actually if you understood elementary physics you'd not have asked that. The platters are gyros, the effect is not a twitch, but actually a slight resistance to being moved. They do, of course, 'twitch' very slightly when the power is turned on.
The real problem is that it's difficult to make the other parts move as close to those platters as they need to, without being so close they 'crash' into the platters when you bump something.
This sounds wonderful. How much do these babies cost?
Thanks for the link. Nice to know those binaries are still out there - I like Opera 6.03 great, but it could still be very handy to have access to 3.62 - an HTML 3.2 compliant browser that will fit on a floppy and run on win16 could come in very handy.
Win 16 means it's compatible with not just win3.1 but also even very old versions of OS/2, WINE will run it perfectly (I know, I was running it on a dual boot Win 3.11/Slackware box for quite awhile, it was very handy, the same binaries running from the same directory, with the same settings files regardless of OS.)
It was NOT introduced in version 4. I started using Opera at version 3 point something, 3.6 I believe it was, and it had it then. It's had it since the earliest alpha version if I'm not mistaken. What was added recently was the ability to turn it off, for the whiners out there that complained about it endlessly.
Not quite true. Xenix was based on the AT&T source - the SysIII source. AT&T didn't license the Unix name at the time, just the code - that's why it was called Xenix instead of MS Unix to begin with. SCO updated it with Berkeley components, SysV compatibility, and so forth.
The part you are correct on, however, is that SCO Unix reflects a new (in '89) license from AT&T, which included rights to the name Unix as well as the actual SysV code.
Just in case you didn't realise it, that MicroSoft Xenix is actually the same thing as SCO Unix (well, if it carried the Xenix name it was an earlier version, but the same codeline, the same unix.) SCO was Microsofts outsourcer - MS didn't actually write Xenix, they just bought the license and hired SCO to do it. Eventually MS lost interest in Xenix, so SCO bought the license and renamed it SCO Unix...
I doubt it very much. People that buy multiprocessor Sun systems are used to paying, and most probably won't blink at those prices. They have to give away the single processor version to compete at all - Linux and BSD are very capable competitors on that hardware, and they're free after all. But Solaris still has the advantage on their multiproc boxen, so people that need that kind of performance will pay for it.
Well the first time I installed linux myself it was RedHat. The init system was so bizzarely convoluted I suspected it was invented simply to torture me ;) I actually liked rpm at the time, it's certainly a lot better than windows install shield crap. But it's needlessly complicated, and when it goes haywire (normally because of poor packaging I am told by rpm advocates - I don't doubt that's true, but then again it's needlessly difficult to package things properly under it too) it can be really ugly.
After installing slack and setting it up I could never go back to an rpm distro. The init system makes sense! The package system makes sense!
With redhat when something went wrong I could never be sure why until I spent the time tracking it down, and it always turned out to be something different and strange each time, sometimes my fault, but often not. When slack screws up, it's always one of two things - either there is a hardware problem or I screwed something up myself. After about a year I started having confidence in that, and that confidence has always turned out to be well founded. It makes troubleshooting so much easier.
"Archaic" perhaps, .tgz is a basic standard that can be relied upon nonetheless, and more than adequate for most uses. What it lacks in fancy features it makes up for simplicity and reliability.
Or waiting a day or two until another slacker gets the source and generates a binary pkg against the current stable release. Between slackware.com and the various user sites like linuxmafia.com and linuxpackages.net most any needed software can be installed from packages, and there is a very high level of quality maintained in the compilation.
Exactly! And even after stating it you never understand it. Slackware does things in very standard ways, so instead of learning the peculiarities of a particular distribution, you just learn solid standard *nix.
The LSB is just wind. Slackware is the living standard that you can install on your machine, run, use, tinker with.
SysVinit is just annoying. RPM *shrug* you can install it on slack if you need it, knock yourself out.
I just deselected the spyware and it installed happily without it, I hardly see why including it for those that don't mind it means they deserve to be skewered by the RIAA. It's been a great service, I'll be very sorry to see it go.
If you really think bothering to check and see when a program is on instead of choosing it by name at the UI level you probably run windows and take your car to the shop whenever you have a problem. Some of us are comfortable with the command line, own a wrench, and are perfectly capable of working a VCR without a special interface. For us, the 'service' is very little help and a major source of annoyance. I think he knows exactly what he wants, and the last thing he should do is listen to you.
Actually it isn't. The recording is done using the reserve space, so it's the same issue.
Frankly I think you're both half right - he's right that the space is his to use as he wants, and you are right that he shouldn't be surprised that they use it the way they designed it to be used. He needs to unplug it and start figuring out how to hack it.
He's hardly 'going after them' - he is declining their invitation and stating his reason, and sticking to it. He has no shortage of things to do, and this is one of his well known requirements and has been for a long time. If they are indeed his fans as you say, why do they not simply use GNU/Linux as the written long form as he asks?
As I say, he has no lack for things to do, and his policy is well known, if they want him to speak they know what they have to do. He has no obligation to speak there.
You are partially correct, it's possible that Linus would have created a kernel anyway, but what we call Linux today certainly would not exist. Minix was not free, and it was quite frankly the most useless unix-inspired OS ever made. Without GCC it seems quite probable the whole project would never have flown in any form. Even BSD uses GCC.
Of course it does! Who the hell do you think ports Free Software to new platforms? People who need to use that software on that platform. It's nonsense to think that people port programs when there is no demand for them.
So if the GNU users aren't the majority on the platform they don't exist? Huh?
Again, who do you think ported the programs and why? Do you have any clue how Free Software works?
Why the heck would it be? You're confused. GNU's Not Unix. Linux Is Not UniX.
BSD is Unix.
That help?
BSD did exist before GNU/Linux, but it was not Free back then.
Quite little actually. Remember, BSD wasn't Free in the beginning. Had it been, a lot of stuff would never have been written, almost certainly including Linux itself. People don't rewrite an entire production quality OS from the ground up just for kicks.
You just gave yourself away, troll.
Maybe it should be BSD/Linux?I don't think you understand the man at all. He doesn't want anything that isn't GNU to be called GNU. He does think that when you have the various distributions of 'Linux' which are bundles of the linux kernel and a bunch of other Free Software, a good portion GNU, you should call it GNU/LINUX. I don't think that's so crazy.
GNU did start the ball rolling, and they did a lot of crucial and necessary work along the way, and they're still doing important work today. It was Stallman's dream alone not so long ago that there would be an entire Free Operating System (and not just one!) that has not only caught but in many ways beaten the commercial Unixes that had only a little earlier accustomed the world to commercial software. He was laughed at, but instead of worrying he wrote code, and pursuaded people to write code, or to donate money to pay people to write code. Crucial tools without which a Free kernel could not be written - shells, fileutilities, compilers, text editors. The infrastructure of the Operating System.
GNU's Not Unix. Linux Is Not UniX.
Ignorant nonsense. GNU tools are used on every Unix, on Apple systems, on Microsoft systems, on VMS even. The GNU tools literally made Linux possible, and Linux is not the only kernel on which the GNU tools and other Free Software can be combined to produce a complete functional and completely Free system.
Although I agree with almost everything you said, I disagree very much with your conclusion. Slashdot rocks. Why? Because of posters like you, who take the time to write long, well-phrased articles about things like 'why slashdot sucks' or, well, why 'slashdot sucks' perhaps.
It's not because of the goatse.cz guy, or the penis bird, or Natalie Porter Nude and Petrified even (although I'll admit all of those were funny once, and the last thought can still send shivers down my spine if it catches me unawares.)
Anyway, I don't think the point to the site was ever to be some sort of hallowed highbrow thing. Yes, it's pitiful the average intelligence they're catering to now, but it's still a little better than many 'mainstream' sites, and a lot of interesting stuff does get posted. A terribly large dose of bullshit too, of course, but ese es el vive, no?
Which is why the key to simple code is small procedures or objects.
Do one thing.
Do it well.
Prove that you do it well.
Move on to the next phase.
It can be done. It does require more planning than some are capable of. And it may not come together as quick as if you just hack what you need. But when you can get away with it, it's very nice.
Actually I found some specs, SA@600MHz uses 450mW, so you could power 4 for the same price as one Crusoe. Perhaps it's a political thing, I'd certainly rather Transmeta get the business than Intel, but I still don't see the technical justification.
Actually something like this could easily be used in a similar system.
Get a clue. The Crusoe consumes about 2 watts. Very nice compared to Pentium-class room heaters, yes, but I asked why they choice Crusoe over StrongARM, not Crusoe over IA-32. A 600mhz SA uses 450mW, so you can run roughly 4 of them for the same power and heat as one Crusoe.
The advantages that Crusoe has are two - first, as I mentioned originally - x86 compatibility. This is not a help for a supercomputer - you're going to be compiling everything from source anyway. The other advantage, that I forgot, is that the SA doesn't have an FPU. That, at least, is a legitimate reason to consider the Crusoe, but I'm still not sure the decision actually makes sense - the SA is a very nice chip and if programmed right it should have no problem keeping up with the Crusoe even on FP, figuring that you can use 4 times as many SAs for the same heat and power requirements.
I have as well, it's a common mistake for nontechnical types (particularly reporters) to make. But they are very different systems. Mainframes have massive redundancy and i/o bandwidth. Supercomputers also have lots of redundancy but they are typically not built for I/O bandwidth at all, but sheer number crunching power. Mainframes are designed to run large databases, supercomputers to do complex mathematics, so you get very different designs for different problems.
The Origin2000 is, if I'm not mistaken, the latest iteration of some of the old Cray designs, and those are definately Supercomputers, not Mainframes. That said, you are of course absolutely correct that ethernet is a major limitation of the sort of cluster we are talking about, and the Crays are still a much better bet for a subset of traditional supercomputer jobs. This is changing, though, as more and more effort and thought goes into improving them.
Ok, finally that's a legitimate response. It's true ARM doesn't include an FPU. However, the last I checked (and I'm not real up to date on it) using libfloat it had emulation good enough to keep up with IA32 fairly well on FP.
I imagine, though, this is probably the reason. It seems reasonable that Supercomputer work would require some FP, although I don't know for sure.
Where did you see anything about Mainframes?
These clusters are NOT designed to take over from Mainframes, but from Supercomputers. Totally different animals.