I've been told by Europeans (east and west) that Americans have a reputation for being exceedingly polite. We chalked it Americans having to wonder if the guy they're talking to is carrying a gun he bought at a weekend gunshow (thus not needing the background check that would have revealed his psychotic history).
Just as you suggest distinguishing between Bill the Microsoft Furror and Bill the Gates Foundation guy, one should also distinguish between US$40 billion owned by MS, and whatever assets Bill Gates the US Citizen has.
When I've bought cutting-edge SCSI equipment, I felt gouged. When I've bought previous generation SCSI equipment, I felt like I was getting what I paid for.
I think that the cutting-edge SCSI drives prices are used to offset the R&D costs. By the time IDE drives reach the same point, the research is already paid for, manufacturing details are already solved, and the company is ready for traditional mass-marketing.
Another factor is reliability. I don't believe that SCSI and IDE drives have the same mechanisms, unless you look at old or crappy SCSI drives and cutting-edge IDE drives. Regardless, you get a longer warrenty with SCSI drives than with IDE drives, and you're probably paying something for that. If the drive mechanisms are indeed the same, I expect that the best mechanisms from a lot end up in SCSI drives, and the rest end up in IDE drives. Just as microproccessor speeds are determined by testing.
Do you have a model name or number for these boxes? I would love to read about one, just for the sake of curiousity. What a great idea that has virtually no use today!
Another way to mix SCSI and IDE is to use IDE for the drives, but SCSI for external communication. There are some nice external IDE RAID setups available now, which talk to the server via SCSI.
You can attach a handful of these external boxes to a server, with each box looking like a single big SCSI drive. Because SCSI cables can be strung ridiculously far, there is no problem daisy-chaining these boxes throughout your cabinet. The drives inside the box are IDE, which saves money.
How many drives do I need? At home, not many. In the lab, 9, which is estimated grow by one every half year (each addition is actually several drives in an external cabinet which talks SCSI externally). Just try to get all those ide controllers into a 1U chassis. =-)
Just try to get a single $20 IDE controller past me and into a big box I have to administer.;-)
Possibly true (though I thought the 15Krpm drives used smaller platters to achieve that speed). However, the SCSI drive is probably warranted for 5 years, while the IDE drive is probably warranted for 3 years. Thus there is an advantage with SCSI disks in terms of "reliability", insofar that if a SCSI disk fails in year four or five you get a new one for free. =-)
That said, free drives in years four and five probably don't justify the additional cost.
Aw geez, do you really have to go and ruin my day like that? I've always been tempted to believe this, but ignored it because I can't do anything about it. Maybe I can convince my wife that we should move across the pond.;-)
I love Python, but you are insane. =-) If you decide to do this, be sure that you have a python interpreter and all necessary libs available, as well as any python modules you use, on your root partition. Keep in mind that sysinit runs before non-root partitions are mounted.
I'm not English. I'm an American and have little knowledge of the inner-workings of the BBC. However, I have my own half-baked opinions, which in the tradition of slashdot, I'll share with public.;-)
I think that there *is* competition in the BBC model. The competition occurs between the people who want to provide programming for the BBC. The BBC cannot afford to be complacent, because they're an easy target for politicians. If the BBC screws up, people contact their reps in government and complain, the reps make some national statement about the decline of the BBC, and top BBC management gets fired. Imagine trying to fire Ted Turner because TNT sucks.
Overall, I find I truly prefer BBC sitcoms over American sitcoms. I'll take Monty Python, Black Adder, Red Dwarf, Dr. Who, and a few others whose names I don't know (and Dame Judy rocks) over Archie Bunker, Family Ties, BH 90210, Friends, and Who's the Boss. I can live with the special effect in Dr. Who that aren't up to Hollywood standards. What I can't stand is the feeling that my brain is leaving me for someone more interesting, which happens when I see Friends.
To me, that is evidence which supports the efficacy of the BBC model, and the dysfunction of the TNT model.
A friend in astronomy has had many troubles getting her parents' friends (among many others) to understand that she is not an astrologer. It was bad enough for her (and probably other astro students) as an undergrad that one year the annual Physics and Astronomy Picnic was advertised as the "Psychics and Astrologers picnic".
Yeah, and I think it's great. I was hoping for Liverpool, but anyone besides Man U would be nice. =-) Besides, Man U has a spotty history against Bayern Muenchen -- does Arsenal have a better record in this case? I only get to watch soccer games when my in-laws record them on VHS tape for me (they have cable, including Fox Sports). Unfortunately, sometimes the announcers are Americans with too much and too little to say, all at the same time.
I used Man U, Bayern Muenchen, and the Brazillian team because they have shown such depth (esp. in the face of adversity) over many, many years even though they lose the occasional (possibly important) battle.
Interesting. I was going to offer "The Gimp" as a counterexample. In my case, though, the "victim" was a technical person (a developer) who couldn't afford Photoshop. He loved it, and I think he even went out and bought books for it.
So maybe this ISO should be targetted at technical folks using Windows.
The cosmologists I work with are often less astounded by the "cool" pictures, and more astounded by the "boring" pictures. While I'm looking at two galaxies coliding or some beautiful nebula, they're looking at some mostly empty portion of the photograph. Then they point at a small dot in a black wasteland and say "Wow, can you believe how *red* that is!". =-)
What you are saying isn't entirely true. I believe that various foreign powers were assisting the South (with weapons?) during the Civil War. Without Saudi Arabia's and Egypt's help, our standing in the Middle East would be far different militarily and politically during the Gulf Emergency Presidential Action (our Congress is a bunch of spineless wimps) and our involvement in Afghanistan.
FWIW, NASA == National Aeronautics and Space Administration. The duplicative "National American" would be redundant.;-)
It's not clear to me that we are leading the civilized world, at least not in the sense that others are following us. Instead, I feel as though we are more like (the soccer teams) Manchester United, Bayern Muenchen, or Brazil's national team: we push so hard to win that it's not fun to play against us, and not necessarily fun to play with us, either.
As I pointed out above, Linus works on Linux the kernel and not on a distribution. As you quote, I like it that "GNU/Linux, the operating system" doesn't have a core team. As far as the kernel goes, I'm more optimistic than you that the current "core" structure is adequate.
And I've added three linux machines (ipaq, laptop, and old desktop) that aren't yet registered at the Linux Counter. I imagine that, for instance, embedded linux companies fail to register every embedded machine they use.
Microsoft counts OEM Windows sales as equivalent to Windows usage. They could use registration numbers, but those would be too low. And then there is "piracy" in the case of Windows.
I figure the Linux Counter isn't any worse than any other estimation method -- especially since we're not even being careful to define what we mean by "user". At any rate, they guy's estimate of 18 million linux users falls between the research companies' estimates of 2 million to 60 million. =-)
"Have you ever tried to release a significant piece of software with 101 different target platforms that vary so much?"
No, I haven't. Have you? I have never seen any major project attempt to support every platform with precompiled binaries. Source availability is a major part of things working at all. It seems that developers usually choose to support one or two distros, and leave support for the remaining distros to interested 3rd parties. This is a sensible way to distribute the load.
If one wishes to release one's software binary-only, then I am not very interested in the difficulties encountered. I'll take source code over convenience any day.
It's not clear to me that this guy is in charge of a bunch of secretaries and managers. If the desktops he wants to replace include animators, graphic artists, and/or developers, then an upgrade would likely increase productivity. Of course, not having to reboot Windows six times a day because your MSVC project has > 10,000 files would also improve productivity.
I think that part of the reason I turn in revulsion from KDE is that it stirs horrid memories of CDE. CDE was but ugly, and when I'm reminded of CDE, the latent CDE-trauma that scarred my brain makes the world look ugly. Thankfully, KDE is the only thing that reminds me of CDE.
Similar symptoms can be observed every time I see mwm or a clone.
I thought the installed base statistic with OS X/BSD was for desktops, not machines total. Is there another statistic that relates total BSD machines (I assume we're limiting ourselves to Open/Net/FreeBSD and OS X) to total GNU/Linux machines? The linux counter has 125,000 linux users and 95,000+ registered machines. They estimate 18 million total users (but don't say for machines). I've seen other estimates between, geez, I think it was 2 million and 60 million installed machines? Quite a range.
While we're on that topic, it would be interesting to see something about the embedded market as well. It's always tough to guage these things because GNU/Linux has huge popularity compared to the *BSD operating systems.
One of the things I like best about GNU/Linux, the operating system, is that it doesn't have a core team. It is a less-conservative approach than that of FreeBSD, and is part of the differentiation between GNU/Linux and FreeBSD. I personally prefer the GNU/Linux style, because among those 100+ distributions are some neat, innovative ideas. FreeBSD also has neat, innovative ideas, but I believe that their centralized control reduces the speed at which they can test these ideas. That said, FreeBSD can provide consistency and quality control in a way that GNU/Linux can't. Though an individual vendor can provide quality and consistency within their distrobution, it cannot enforce consistency between distrobutions (though the LSB is a good step toward identifying where it is important to be consistent).
Linux, the kernel, has an informal core body, with people wandering in and out. It is a much smaller project than FreeBSD, since FreeBSD is more than just a kernel. Therefore, one should be careful when reasoning about the benefits FreeBSD development methodology would bring to Linux.
Heh, I guess I thought Win95 was later than it actually was. Maybe it took a while to hit store shelves, or for OEMs to bother preloading it. Doesn't much matter, of course, but I'm curious where the mistake in my brain came from (probably errant neutrinos...;-)
Slashdot Mirror
I've been told by Europeans (east and west) that Americans have a reputation for being exceedingly polite. We chalked it Americans having to wonder if the guy they're talking to is carrying a gun he bought at a weekend gunshow (thus not needing the background check that would have revealed his psychotic history).
-Paul Komarek
Just as you suggest distinguishing between Bill the Microsoft Furror and Bill the Gates Foundation guy, one should also distinguish between US$40 billion owned by MS, and whatever assets Bill Gates the US Citizen has.
-Paul KOmarek
When I've bought cutting-edge SCSI equipment, I felt gouged. When I've bought previous generation SCSI equipment, I felt like I was getting what I paid for.
I think that the cutting-edge SCSI drives prices are used to offset the R&D costs. By the time IDE drives reach the same point, the research is already paid for, manufacturing details are already solved, and the company is ready for traditional mass-marketing.
Another factor is reliability. I don't believe that SCSI and IDE drives have the same mechanisms, unless you look at old or crappy SCSI drives and cutting-edge IDE drives. Regardless, you get a longer warrenty with SCSI drives than with IDE drives, and you're probably paying something for that. If the drive mechanisms are indeed the same, I expect that the best mechanisms from a lot end up in SCSI drives, and the rest end up in IDE drives. Just as microproccessor speeds are determined by testing.
-Paul Komarek
Do you have a model name or number for these boxes? I would love to read about one, just for the sake of curiousity. What a great idea that has virtually no use today!
-Paul Komarek
Another way to mix SCSI and IDE is to use IDE for the drives, but SCSI for external communication. There are some nice external IDE RAID setups available now, which talk to the server via SCSI.
You can attach a handful of these external boxes to a server, with each box looking like a single big SCSI drive. Because SCSI cables can be strung ridiculously far, there is no problem daisy-chaining these boxes throughout your cabinet. The drives inside the box are IDE, which saves money.
-Paul Komarek
How many drives do I need? At home, not many. In the lab, 9, which is estimated grow by one every half year (each addition is actually several drives in an external cabinet which talks SCSI externally). Just try to get all those ide controllers into a 1U chassis. =-)
;-)
Just try to get a single $20 IDE controller past me and into a big box I have to administer.
-Paul Komarek
Possibly true (though I thought the 15Krpm drives used smaller platters to achieve that speed). However, the SCSI drive is probably warranted for 5 years, while the IDE drive is probably warranted for 3 years. Thus there is an advantage with SCSI disks in terms of "reliability", insofar that if a SCSI disk fails in year four or five you get a new one for free. =-)
That said, free drives in years four and five probably don't justify the additional cost.
-Paul Komarek
Aw geez, do you really have to go and ruin my day like that? I've always been tempted to believe this, but ignored it because I can't do anything about it. Maybe I can convince my wife that we should move across the pond. ;-)
-Paul Komarek
I love Python, but you are insane. =-) If you decide to do this, be sure that you have a python interpreter and all necessary libs available, as well as any python modules you use, on your root partition. Keep in mind that sysinit runs before non-root partitions are mounted.
-Paul Komarek
I'm not English. I'm an American and have little knowledge of the inner-workings of the BBC. However, I have my own half-baked opinions, which in the tradition of slashdot, I'll share with public. ;-)
I think that there *is* competition in the BBC model. The competition occurs between the people who want to provide programming for the BBC. The BBC cannot afford to be complacent, because they're an easy target for politicians. If the BBC screws up, people contact their reps in government and complain, the reps make some national statement about the decline of the BBC, and top BBC management gets fired. Imagine trying to fire Ted Turner because TNT sucks.
Overall, I find I truly prefer BBC sitcoms over American sitcoms. I'll take Monty Python, Black Adder, Red Dwarf, Dr. Who, and a few others whose names I don't know (and Dame Judy rocks) over Archie Bunker, Family Ties, BH 90210, Friends, and Who's the Boss. I can live with the special effect in Dr. Who that aren't up to Hollywood standards. What I can't stand is the feeling that my brain is leaving me for someone more interesting, which happens when I see Friends.
To me, that is evidence which supports the efficacy of the BBC model, and the dysfunction of the TNT model.
-Paul Komarek
"Figures. Damned cosmetologists."
A friend in astronomy has had many troubles getting her parents' friends (among many others) to understand that she is not an astrologer. It was bad enough for her (and probably other astro students) as an undergrad that one year the annual Physics and Astronomy Picnic was advertised as the "Psychics and Astrologers picnic".
cosmologist/cosmotologist is a nice addition. =-)
-Paul Komarek
Yeah, and I think it's great. I was hoping for Liverpool, but anyone besides Man U would be nice. =-) Besides, Man U has a spotty history against Bayern Muenchen -- does Arsenal have a better record in this case? I only get to watch soccer games when my in-laws record them on VHS tape for me (they have cable, including Fox Sports). Unfortunately, sometimes the announcers are Americans with too much and too little to say, all at the same time.
I used Man U, Bayern Muenchen, and the Brazillian team because they have shown such depth (esp. in the face of adversity) over many, many years even though they lose the occasional (possibly important) battle.
-Paul Komarek
Interesting. I was going to offer "The Gimp" as a counterexample. In my case, though, the "victim" was a technical person (a developer) who couldn't afford Photoshop. He loved it, and I think he even went out and bought books for it.
So maybe this ISO should be targetted at technical folks using Windows.
-Paul Komarek
The best muppet for to represent the vm system in kernel 2.4 is Animal, the crazy red-haired drum-playing monster.
-Paul Komarek
The cosmologists I work with are often less astounded by the "cool" pictures, and more astounded by the "boring" pictures. While I'm looking at two galaxies coliding or some beautiful nebula, they're looking at some mostly empty portion of the photograph. Then they point at a small dot in a black wasteland and say "Wow, can you believe how *red* that is!". =-)
-Paul Komarek
What you are saying isn't entirely true. I believe that various foreign powers were assisting the South (with weapons?) during the Civil War. Without Saudi Arabia's and Egypt's help, our standing in the Middle East would be far different militarily and politically during the Gulf Emergency Presidential Action (our Congress is a bunch of spineless wimps) and our involvement in Afghanistan.
;-)
FWIW, NASA == National Aeronautics and Space Administration. The duplicative "National American" would be redundant.
It's not clear to me that we are leading the civilized world, at least not in the sense that others are following us. Instead, I feel as though we are more like (the soccer teams) Manchester United, Bayern Muenchen, or Brazil's national team: we push so hard to win that it's not fun to play against us, and not necessarily fun to play with us, either.
-Paul Komarek
As I pointed out above, Linus works on Linux the kernel and not on a distribution. As you quote, I like it that "GNU/Linux, the operating system" doesn't have a core team. As far as the kernel goes, I'm more optimistic than you that the current "core" structure is adequate.
-Paul Komarek
And I've added three linux machines (ipaq, laptop, and old desktop) that aren't yet registered at the Linux Counter. I imagine that, for instance, embedded linux companies fail to register every embedded machine they use.
Microsoft counts OEM Windows sales as equivalent to Windows usage. They could use registration numbers, but those would be too low. And then there is "piracy" in the case of Windows.
I figure the Linux Counter isn't any worse than any other estimation method -- especially since we're not even being careful to define what we mean by "user". At any rate, they guy's estimate of 18 million linux users falls between the research companies' estimates of 2 million to 60 million. =-)
-Paul Komarek
Is it allowed to do something like
#define undef #define HAHAHA \/\/
;-)
-Paul Komarek
"Have you ever tried to release a significant piece of software with 101 different target platforms that vary so much?"
No, I haven't. Have you? I have never seen any major project attempt to support every platform with precompiled binaries. Source availability is a major part of things working at all. It seems that developers usually choose to support one or two distros, and leave support for the remaining distros to interested 3rd parties. This is a sensible way to distribute the load.
If one wishes to release one's software binary-only, then I am not very interested in the difficulties encountered. I'll take source code over convenience any day.
-Paul Komarek
It's not clear to me that this guy is in charge of a bunch of secretaries and managers. If the desktops he wants to replace include animators, graphic artists, and/or developers, then an upgrade would likely increase productivity. Of course, not having to reboot Windows six times a day because your MSVC project has > 10,000 files would also improve productivity.
-Paul Komarek
I think that part of the reason I turn in revulsion from KDE is that it stirs horrid memories of CDE. CDE was but ugly, and when I'm reminded of CDE, the latent CDE-trauma that scarred my brain makes the world look ugly. Thankfully, KDE is the only thing that reminds me of CDE.
Similar symptoms can be observed every time I see mwm or a clone.
-Paul Komarek
I thought the installed base statistic with OS X/BSD was for desktops, not machines total. Is there another statistic that relates total BSD machines (I assume we're limiting ourselves to Open/Net/FreeBSD and OS X) to total GNU/Linux machines? The linux counter has 125,000 linux users and 95,000+ registered machines. They estimate 18 million total users (but don't say for machines). I've seen other estimates between, geez, I think it was 2 million and 60 million installed machines? Quite a range.
While we're on that topic, it would be interesting to see something about the embedded market as well. It's always tough to guage these things because GNU/Linux has huge popularity compared to the *BSD operating systems.
-Paul Komarek
One of the things I like best about GNU/Linux, the operating system, is that it doesn't have a core team. It is a less-conservative approach than that of FreeBSD, and is part of the differentiation between GNU/Linux and FreeBSD. I personally prefer the GNU/Linux style, because among those 100+ distributions are some neat, innovative ideas. FreeBSD also has neat, innovative ideas, but I believe that their centralized control reduces the speed at which they can test these ideas. That said, FreeBSD can provide consistency and quality control in a way that GNU/Linux can't. Though an individual vendor can provide quality and consistency within their distrobution, it cannot enforce consistency between distrobutions (though the LSB is a good step toward identifying where it is important to be consistent).
Linux, the kernel, has an informal core body, with people wandering in and out. It is a much smaller project than FreeBSD, since FreeBSD is more than just a kernel. Therefore, one should be careful when reasoning about the benefits FreeBSD development methodology would bring to Linux.
-Paul Komarek
Heh, I guess I thought Win95 was later than it actually was. Maybe it took a while to hit store shelves, or for OEMs to bother preloading it. Doesn't much matter, of course, but I'm curious where the mistake in my brain came from (probably errant neutrinos... ;-)
-Paul Komarek