When said that way, it sounds pretty unappealing.
However, when said another way, it works: this is the internet radio equivalent of cable (which, at many schools, you get whether you want it or not). I am now too old to remember whether that was the deal when I was at Penn State or not (they brought in cable in my dorm during my senior year if I remember correctly; but I can't remember if it was an opt-in with extra payment or was just stuck on the room&board bill). Obviously once demand exceeds some percentage X, it gets a lot cheaper (for the university) to just charge everybody, so I'd be surprised if they don't do that now.
Certainly the "real world" ain't much better; I'm stuck with full cable TV if I want high-speed internet access despite the fact that I only want to watch a few of the channels. And even if I wanted to watch 0 of the channels; there ain't another option in town for non-dialup which doesn't involve incessant yodelling.
my old employer, acronym generated from Sight Sound Suck, had basically two modes in most drivers: one (WHQL-mode) where the thing was run in the best possible image-quality mode, working around various sundry hardware bugs; and then a set of speed-modes if game-EXEs were found.
With benchmarks, it was a bit trickier. One thing they did in particular was to actually attempt to figure out what app was running by doing an analysis of the first N texture loads; which bit us in the ass later on when the same value ended up coincidentally being generated for one of the WHQL tests (where output MUST be exactly correct).
NVidia used to be better than this; we had respect for them for at least just implementing the reference rasterizer (fast) in hardware rather than trying to implement new features in hardware without MS's blessing. Those HW features would be implemented well in 3dfx's case, crappy in ours; but the result was identical - if it's not part of the Direct3D API, it's completely useless.
If there was ever time for engineering software, you might see different results. However, the disciplines most people consider "engineering" have levels of process built in that software has never found acceptable.
For instance, you can't stop building a bridge halfway through and then decide to reengineer it. You can with software, even when it's a bad idea. You can't start using new concrete in the second half of your bridge. You can with software, even when it's a bad idea.
The flexibility of software is the reason why software engineering usually fails - the managers never hold strong enough against the feature creep brigade from sales. I don't care how persuasive the MBA wankers are at BridgeCo; they're not going to be able to convince the civil engineers' boss to do something stupid just to make the bridge a bit prettier halfway through the project.
-You can know NFS,AFS, and Samba
-You can know Apache
-You can know X11
-You can know sendmail/postfix
-You can know telnet/ssh/rsh
-You can know how to install security updates
Dude, when I used a mainframe for school back in the early 90s, none of that stuff except sendmail would have even REMOTELY been applicable. I think perhaps you don't know what mainframes _are_.
Damn straight; the first thing superstars need to learn out of college is that most of the time, there's a guy with pointy hair in the way of what they'd REALLY like to do with the code. That's why only companies run by idiots hire people who wouldn't be happy with jobs that only occasionally challenge them.
And the market size represented by all the PCs at all the governments in the world is not, I believe, big enough to make Microsoft lose more than a minute or two of sleep at night.
Except that libertarians like ESR don't want the government to intervene via antitrust law, so it doesn't matter how many governments MS pisses off.
The only thing that stopped IBM was the internal organizational changes brought about by the ongoing antitrust suit. Otherwise, you'd probably still be using a dumb terminal on your desk right now.
Some of the tests are very exacting - you had better not be even one pixel off.
Others are more lenient - some had a "passing" grade set at 85%, if I recall, for that very reason.
The basic geometry, incl. culling, is supposed to be 100% perfect.
Actually, the software can, in most cases, check the rendering to the screen. The "correctness" benchmarks do this. The problem is that it is slow.
WHQL DCT tests do this - you get two windows (in most tests); one of which was drawn using the reference rasterizer; and one of which was drawn using the graphics card; and believe me, they do test pixel-by-pixel.
PC Magazine's benchmark did something similar; but again, it's not factored into the benchmark score. And obviously they don't test enough things if they got fooled here; but that's an argument for expanding the "correctness" suite.
I worked for a pretty crappy graphics card company (with hardware that was either architected poorly for Direct3D or simply buggy), and it was standard operating procedure to cheat any way we could.
One thing we'd do was do the fastest but lowest quality path UNLESS we found that the EXE was one of a list of games where quality was important. We used to test for the benchmarks instead but got caught at least once when one of the benchmark makers figured out they could randomly change their EXE name right before launch and catch people that way (not just us, by the way).
Even the good companies like 3dfx and nvidia cheated - they just did it less because their hardware was generally better.
People wonder why drivers are buggy; it's because of crappy hardware design first (NVidia used to just design to the reference rasterizer; which helped BIG-TIME) and then second the fact that 90% of the driver-writers' time was spent either cheating or eking out another 1% of performance.
x86 is completely different from PPC - recompiling isn't even the half of it; and I shudder to think of attempting to run an OS on x86 which doesn't have a ton of its kernel written in assembler.
And meanwhile, your lack of will to engage in the collective process might mean that the homestead you deed to your children will no longer be able to be sustainable due to the climate change that could not be stopped because too many people abdicated their common responsibilities.
It all clarifies that you try to live with a small footprint. That's wonderful. I try to do that too.
But two people living with a small footprint in a regulatory regime which favors people who live with large footprints doesn't do anything but penalize the people who want to do the right thing and join those two.
It's the old libertarian blind-spot again: the tragedy of the commons. Without emissions taxes, it economically benefits me to be as dirty as possible (because it's almost always cheaper than being clean); because I get 100% of the benefit and 1/6,000,000,000th of the cost.
Only evil big government can fix the tragedy of the commons.
I'd be curious, then, to hear how you propose to do anything about emissions of CO2 without the participation of the corrupt, inefficient, big governments; (which, by the way, somehow managed to do it right when it came to CFCs).
This type of moaning about how we can't trust government to fix a problem is basically a red herring - they are the only ones who CAN. Saying that we shouldn't use them as tools here is giving in to those who actually prefer to do nothing.
The big problem here is that Kyoto is ALREADY a compromise between your two extremes. A cut of like 10% in emissions is hardly returning to a pre-industrial state; you've been spending too much time watching FoxNews.
Note that I was specifically talking about cases where your organization already has N workers, and just hired worker N+1, and need to get a new machine for that worker. Not running old stuff for current workers.
But even in that subset, there's two separate issues:
1. Preloading (small/medium businesses generally prefer to have the OS preloaded) - this precludes NT4 in all practical manner already.
2. Non-preloaded (large businesses and a few others who would generally prefer their own image be put on the machine) - this will be less practical every day - it IS pretty difficult to find brand new machines where everything will be supported well by NT4 today. Not impossible; but definitely more trouble than a lot of organizations would consider it worth.
Yes, Linux is easy to use, *for what it does*. But one of the things it *doesn't* do well is replicate Windows. And, for better or for worse, "Windows" is what most people seem to want, not "an OS that is easy to use".
Or, their definition of "easy to use" means "runs Windows apps" or "runs all my apps" or even "runs a bunch of apps in categories X, Y, and Z".
Of course I've been through this with OS/2. If you think that your definition of "easy to use" is the canonical one, it gets very difficult to effectively communicate with others.
The hard part is buying new hardware in an organization which doesn't keep up with Microsoft's OS schedule.
For instance, I dare you to try to buy a new machine now from a Dell or a HP which runs NT4; so that your employee N+1 will have the same environment as employee N.
The difference between your KDE example and Windows is that Netscape had been making a profitable business out of selling a browser (and thus had an incentive to respond to user feedback; at least, large corporate user feedback).
This is the thing that nearly everybody forgets: netscape was making money SELLING BROWSERS. Yes, they were making more money on the server side; but the fact that for the last N years, the main browser has been made by somebody who has no incentive whatsoever to satisfy browser users of any type can't be good.
You suggested that somebody with a Music or PoliSci degree would be as good of a programmer as somebody with a CS or CE degree. The implication was, obviously, that a liberal arts major who knew how to program would be better off.
Sorry for calling you on what you actually said.
" You seem to be assuming that the only way to prepare for a career as a programmer is to study computer science. And that you can only judge a CS curriculum by how well it prepares you for such a career. Neither is true. Computer science and computer engineering are completely separate disciplines -- and neither is a prerequisite to a career as a programer. Of course, some companies won't even look at you if you don't have a CS degree. In some cases, that makes sense. But not always. And not all companies think that way. My last full time job was at a major software company where half the programmers didn't even have a computer degree -- and a lot of them majored in in such high tech fields as Poli Sci or Music."
Why bother making such a statement if you didn't think having a more "broad" background was a good thing for one's career?
Hint: Right now, if you try to get a job without word-of-mouth, and your educational qualifications consist of a Music or PoliSci major with some programming classes, you'll be flipping burgers, at best. Giving advice in a forum like slashdot where a lot of the audience is in college carries a certain responsibility along with it not to be such a pretentious twat.
Arguing that a broader education is necessary to be a good programmer in your case boiled down to supporting people learning part-time programming while getting a liberal arts degree. That's career suicide even in a normal economy (i.e. better than the one we have now).
We get 100 resumes a day for each available programming position. Since I and my coworkers all know that the correlation between (lacks a very technical degree) and (bad programmer) is so high, it would be business suicide for us to fail to use that as a screening tool.
As I've said twice now, word-of-mouth can overcome this. But you can't always get a job that way.
I question whether you've ever been involved in screening a large number of people for a position, if you can afford to be so cavalier about education. Yes, as I said at the beginning, there are some people who will be good programmers without the degree; but there is such a high correlation between good programmers and a CS or EE degree that I would never risk hiring one of those people without a word-of-mouth recommendation.
One huge company, one medium company, and one small company; two of the three sold software as a product; two of the three sold hardware as a product.
Most of the people we worked with who lacked CS or EE degrees were people who could string scripts together, but if you put them on a big project, they were still trying to string scripts together, except they were doing it in C or C++.
Slashdot Mirror
When said that way, it sounds pretty unappealing. However, when said another way, it works: this is the internet radio equivalent of cable (which, at many schools, you get whether you want it or not). I am now too old to remember whether that was the deal when I was at Penn State or not (they brought in cable in my dorm during my senior year if I remember correctly; but I can't remember if it was an opt-in with extra payment or was just stuck on the room&board bill). Obviously once demand exceeds some percentage X, it gets a lot cheaper (for the university) to just charge everybody, so I'd be surprised if they don't do that now. Certainly the "real world" ain't much better; I'm stuck with full cable TV if I want high-speed internet access despite the fact that I only want to watch a few of the channels. And even if I wanted to watch 0 of the channels; there ain't another option in town for non-dialup which doesn't involve incessant yodelling.
my old employer, acronym generated from Sight Sound Suck, had basically two modes in most drivers: one (WHQL-mode) where the thing was run in the best possible image-quality mode, working around various sundry hardware bugs; and then a set of speed-modes if game-EXEs were found. With benchmarks, it was a bit trickier. One thing they did in particular was to actually attempt to figure out what app was running by doing an analysis of the first N texture loads; which bit us in the ass later on when the same value ended up coincidentally being generated for one of the WHQL tests (where output MUST be exactly correct). NVidia used to be better than this; we had respect for them for at least just implementing the reference rasterizer (fast) in hardware rather than trying to implement new features in hardware without MS's blessing. Those HW features would be implemented well in 3dfx's case, crappy in ours; but the result was identical - if it's not part of the Direct3D API, it's completely useless.
If there was ever time for engineering software, you might see different results. However, the disciplines most people consider "engineering" have levels of process built in that software has never found acceptable. For instance, you can't stop building a bridge halfway through and then decide to reengineer it. You can with software, even when it's a bad idea. You can't start using new concrete in the second half of your bridge. You can with software, even when it's a bad idea. The flexibility of software is the reason why software engineering usually fails - the managers never hold strong enough against the feature creep brigade from sales. I don't care how persuasive the MBA wankers are at BridgeCo; they're not going to be able to convince the civil engineers' boss to do something stupid just to make the bridge a bit prettier halfway through the project.
Damn straight; the first thing superstars need to learn out of college is that most of the time, there's a guy with pointy hair in the way of what they'd REALLY like to do with the code. That's why only companies run by idiots hire people who wouldn't be happy with jobs that only occasionally challenge them.
Tens of millions of computers, perhaps. Tens of millions of new PCs? Not a chance.
And the market size represented by all the PCs at all the governments in the world is not, I believe, big enough to make Microsoft lose more than a minute or two of sleep at night.
Except that libertarians like ESR don't want the government to intervene via antitrust law, so it doesn't matter how many governments MS pisses off. The only thing that stopped IBM was the internal organizational changes brought about by the ongoing antitrust suit. Otherwise, you'd probably still be using a dumb terminal on your desk right now.
Some of the tests are very exacting - you had better not be even one pixel off. Others are more lenient - some had a "passing" grade set at 85%, if I recall, for that very reason. The basic geometry, incl. culling, is supposed to be 100% perfect.
Actually, the software can, in most cases, check the rendering to the screen. The "correctness" benchmarks do this. The problem is that it is slow. WHQL DCT tests do this - you get two windows (in most tests); one of which was drawn using the reference rasterizer; and one of which was drawn using the graphics card; and believe me, they do test pixel-by-pixel. PC Magazine's benchmark did something similar; but again, it's not factored into the benchmark score. And obviously they don't test enough things if they got fooled here; but that's an argument for expanding the "correctness" suite.
I worked for a pretty crappy graphics card company (with hardware that was either architected poorly for Direct3D or simply buggy), and it was standard operating procedure to cheat any way we could.
One thing we'd do was do the fastest but lowest quality path UNLESS we found that the EXE was one of a list of games where quality was important. We used to test for the benchmarks instead but got caught at least once when one of the benchmark makers figured out they could randomly change their EXE name right before launch and catch people that way (not just us, by the way).
Even the good companies like 3dfx and nvidia cheated - they just did it less because their hardware was generally better.
People wonder why drivers are buggy; it's because of crappy hardware design first (NVidia used to just design to the reference rasterizer; which helped BIG-TIME) and then second the fact that 90% of the driver-writers' time was spent either cheating or eking out another 1% of performance.
x86 is completely different from PPC - recompiling isn't even the half of it; and I shudder to think of attempting to run an OS on x86 which doesn't have a ton of its kernel written in assembler.
I already mentioned one: the CFC treaty.
Not much of a reward, by my way of thinking.
But two people living with a small footprint in a regulatory regime which favors people who live with large footprints doesn't do anything but penalize the people who want to do the right thing and join those two.
It's the old libertarian blind-spot again: the tragedy of the commons. Without emissions taxes, it economically benefits me to be as dirty as possible (because it's almost always cheaper than being clean); because I get 100% of the benefit and 1/6,000,000,000th of the cost.
Only evil big government can fix the tragedy of the commons.
This type of moaning about how we can't trust government to fix a problem is basically a red herring - they are the only ones who CAN. Saying that we shouldn't use them as tools here is giving in to those who actually prefer to do nothing.
The big problem here is that Kyoto is ALREADY a compromise between your two extremes. A cut of like 10% in emissions is hardly returning to a pre-industrial state; you've been spending too much time watching FoxNews.
Note that I was specifically talking about cases where your organization already has N workers, and just hired worker N+1, and need to get a new machine for that worker. Not running old stuff for current workers.
But even in that subset, there's two separate issues:
1. Preloading (small/medium businesses generally prefer to have the OS preloaded) - this precludes NT4 in all practical manner already.
2. Non-preloaded (large businesses and a few others who would generally prefer their own image be put on the machine) - this will be less practical every day - it IS pretty difficult to find brand new machines where everything will be supported well by NT4 today. Not impossible; but definitely more trouble than a lot of organizations would consider it worth.
Or, their definition of "easy to use" means "runs Windows apps" or "runs all my apps" or even "runs a bunch of apps in categories X, Y, and Z".
Of course I've been through this with OS/2. If you think that your definition of "easy to use" is the canonical one, it gets very difficult to effectively communicate with others.
The hard part is buying new hardware in an organization which doesn't keep up with Microsoft's OS schedule. For instance, I dare you to try to buy a new machine now from a Dell or a HP which runs NT4; so that your employee N+1 will have the same environment as employee N.
The difference between your KDE example and Windows is that Netscape had been making a profitable business out of selling a browser (and thus had an incentive to respond to user feedback; at least, large corporate user feedback). This is the thing that nearly everybody forgets: netscape was making money SELLING BROWSERS. Yes, they were making more money on the server side; but the fact that for the last N years, the main browser has been made by somebody who has no incentive whatsoever to satisfy browser users of any type can't be good.
You suggested that somebody with a Music or PoliSci degree would be as good of a programmer as somebody with a CS or CE degree. The implication was, obviously, that a liberal arts major who knew how to program would be better off. Sorry for calling you on what you actually said. " You seem to be assuming that the only way to prepare for a career as a programmer is to study computer science. And that you can only judge a CS curriculum by how well it prepares you for such a career. Neither is true. Computer science and computer engineering are completely separate disciplines -- and neither is a prerequisite to a career as a programer. Of course, some companies won't even look at you if you don't have a CS degree. In some cases, that makes sense. But not always. And not all companies think that way. My last full time job was at a major software company where half the programmers didn't even have a computer degree -- and a lot of them majored in in such high tech fields as Poli Sci or Music." Why bother making such a statement if you didn't think having a more "broad" background was a good thing for one's career? Hint: Right now, if you try to get a job without word-of-mouth, and your educational qualifications consist of a Music or PoliSci major with some programming classes, you'll be flipping burgers, at best. Giving advice in a forum like slashdot where a lot of the audience is in college carries a certain responsibility along with it not to be such a pretentious twat.
Arguing that a broader education is necessary to be a good programmer in your case boiled down to supporting people learning part-time programming while getting a liberal arts degree. That's career suicide even in a normal economy (i.e. better than the one we have now). We get 100 resumes a day for each available programming position. Since I and my coworkers all know that the correlation between (lacks a very technical degree) and (bad programmer) is so high, it would be business suicide for us to fail to use that as a screening tool. As I've said twice now, word-of-mouth can overcome this. But you can't always get a job that way.
I question whether you've ever been involved in screening a large number of people for a position, if you can afford to be so cavalier about education. Yes, as I said at the beginning, there are some people who will be good programmers without the degree; but there is such a high correlation between good programmers and a CS or EE degree that I would never risk hiring one of those people without a word-of-mouth recommendation.
One huge company, one medium company, and one small company; two of the three sold software as a product; two of the three sold hardware as a product. Most of the people we worked with who lacked CS or EE degrees were people who could string scripts together, but if you put them on a big project, they were still trying to string scripts together, except they were doing it in C or C++.