Of course the lives we lead is natural. It's in human nature to build cities and factories, therefore we live natural human lives. It is no less natural for a human to eat processed food (which is made by humans of course) than it is for a beaver to live in a lodge.
I eat an omnivorous diet which occasionally involves a rare steak amongst other things. Since I joined the workforce in 1990, I have had exactly one day off sick, and it wasn't serious enough to warrant seeing the doctor. The only time I've seen a doctor in this time were for routine medicals (aviation, employment etc). and for an injury.
The compiler would have to be able to read minds to throw a warning on that (it won't, even with -Wall). It's a perfectly sane thing to compare a function pointer to NULL in most code. Perhaps we need -Wdowhatimeannotwhatisay.
A few machines? All of the most popular distros (Ubuntu, Fedora Core, Debian, CentOS) have free update services using yum or apt, and have had these mechanisms for years.
The truth sometimes hurts. Theo de Raadt just doesn't dress it up. I wouldn't hire Theo as a diplomat (well, not unless I wanted to actually start a war), but I would hire him as someone who can write secure code. I don't care if he has no social skills; I'm after secure code. That's why we use OpenBSD for security critical things.
Fat chance. Back in the day, when 2D acceleration was the in thing and Linux just had X support, S3 were one to avoid because unlike Tseng Labs and co. they didn't publish the specs for their card so no X support. I suspect they are still one to avoid today. The chances of them publishing the specs so someone else (i.e. not using up their time) can do the drivers are tiny.
Just one question - surely if you're using an Ethernet switch, and the switch is at one end of the wire, and the computer at the other - where do the collisions come from? How would a token-based network be superior to a switch based network? You don't lose the cost of the switch - IBM Token Ring still had a 'concentrator' in the middle so you wound up with a physical star, just like with switched ethernet (and was significantly more prone to breakage should a card go on the fritz and stop passing the token).
On a server I have that sees plenty of network traffic over switched ethernet, I still have '0' in the interface's collision column.
It's good 'ol Rip-Off Britain again. Don't forget, in the US, sales tax is also much lower than 17.5% (typically 4-6% - that's if Apple are even paying sales tax to the various states, I have a feeling that they are not).
The 'rip off' effect in Britain is such that the exchange rate for normal purposes is pretty much irrelevant - really, £1 is only worth about $1 (or less) for most things in actual 'purchasing power' - music especially (often, you'll find a CD in Britain is £14 and only $12 in the US).
But my Sierra was made in Germany. I also owned a 1986 Dodge Ram when I lived in the United States - it also suffered from the same kind of reliability problems as the Sierra (hard cold+damp starting/needing frequent tune-ups). It also rusted, but because it was made out of an awful lot more metal, it would take longer to rust through. Most 1970s cars were junkers by the time they were 10 years old; most mid 90s cars are still perfectly good and ran like they did when they came out of the factory.
Ah, the 'they don't make cars like they used to' fallacy.
I've owned older cars. They need a lot of TLC to keep running reliably - my old Mini needed new points/condenser every 6 months (with the oil change), frequent spark plug/HT cable replacements, and I had to have the cylinder head off twice in the time I owned it. Although it was a fun car to drive (and had lots of character), it needed a LOT of maintenance to have any hope of reliability. There was actually a very noticable performance difference after each 6 monthly maintenance - those old cars really did deteriorate that fast. It also suffered from rust.
My next car was a 1984 Sierra. Not much in the way of computer control, but unlike the Mini, I *never ever* had ignition problems with it, and only broke down a couple of times (the clutch cables were very bad and had a tendency to suddenly go). I had much fewer reliability problems with the Sierra. I still had them though. It could be hard starting on a cold damp morning, and would run a bit rough until it was warmed up (automatic chokes were never a good idea). It also rusted.
My current car, a 1995 Audi A4, just runs consistently, day in, day out. It always starts easily whether it's hot or cold, damp or dry. It is galvanized, so despite living on a small windswept island with a very salty atmosphere, it doesn't rust. After I've cleaned it, it looks just as good as the day it came out of the factory - the paint is so much better on it than older cars (most 1970s cars, when 11 years old would be unreliable, faded and rusty). It also drives just as well today as when it came from the factory - it still performs like the book says it should. It never needs tune-ups like carburetted cars like the Sierra or Mini did because the computer keeps it tuned all the time.
Sure, older cars are much easier to work on - but the point is, I don't have to work on my Audi like I had to work on my Sierra (or even more so, the Mini). And my Audi doesn't rust.
Most people hardly use the computing power of a late-90s Mac (typing letters and reading email and browsing the web is hardly CPU intensive) so the lower overall wattage of the machine will mean it's more energy efficient. CPU instructions per watt is pretty irrelevant unless you really are using the processing power, and not having the computer sitting waiting on I/O all the time.
That's always been the case. I was in a school with about 600 people. We had (very advanced for the time - it was in the period 1985 to 1990) a network of BBC Microcomputers - which had a built-in BASIC interpreter (reputedly, one of the best of its era) as well as a built in 6502 assembler (it was the about the only machine of the era which came with an assembler).
We had a large room in the bottom of the science block with about 15 of these machines, plus a Mac Plus. There were a few other BBCs on the network - the geography block had three (and a Mac).
So a school with 600 students. There was no problem getting on to one of these machines at lunch or after school - demand was simply not very high. There were perhaps half a dozen of us who'd actually use these machines to program. The only time the computer room was full and every machine was in use was when a teacher had a class in there, and more often than not the computer was being used as a teaching aid rather than teaching something about the computer itself.
Programming wasn't rewarded for me at HS level (in the 8-bit days), so much to my teachers chagrin, instead of working on their classes, I wrote 6502 asm instead. You must get your priorities right:-)
Of course, my school reports were very, very bad and I failed most of my exams. But in the end it didn't really matter.
Airplanes fly on the principles described by Newton and Bernoulli, *not* Marconi and Turing! Planes would quite happily fly if the FAA didn't exist at all. Sure, airlines would be impractical, but the vast majority of air traffic in the United States (general aviation) spends most of its time flying on nice VFR days not talking to ATC at all.
Oh, I'm not trying to excuse Microsoft for this - in fact, it's really a pointer to why the quality of Microsoft software is so variable, and why Windows and IE has so many serious security problems and other bugs - despite highly talented developers, the suspected lack of any kind of software engineering process almost inevitably results in the kinds of problems we see with their software -- and the difficulty in getting decent support for more obscure parts of Windows even when you're paying for a very expensive support contract.
I suspect some divisions of MS have a decent software process and others don't - much alluded to by the piece the other day about how developers get so much freedom (and developers DO NOT choose to have a software process by their own volition - generally, software process is castor oil to programmers).
Worse still was the talk a while back that Bill Gates wanted to get his hands dirty with Vista (then Longhorn) and spend 50% of his time back on technical stuff. Bill Gates has a reputation for being a bit of a 'prima donna' about any programming he thinks is wrong, and he's been a manager for a while and is probably un-current. Imagine you're a developer on Windows, and Gates has decided to work on a code module you're involved in. Would you have the guts to correct programming mistakes that Gates makes, knowing full well his reputation for hauling developers into his office and haranguing them over their code? Knowing you could get summarily fired? I think the problems with XP will just continue in Vista because of things like this.
I get the distinct impression that Microsoft being tardy to release the documentation for their protocols isn't purely due to malice - or even not malice at all. I get the impression that things like this simply aren't documented INSIDE of Microsoft, and the original developers of the code have left, and MS staff are now busily trying to document a gigantic mass of source code.
I have direct experience of Microsoft having none or inadequate internal documentation (see my latest JE for the full discussion - http://slashdot.org/~Alioth/journal/133996). A quick precis is that we were working with the GINA in the NT 4.0 days, and we had an expensive support contract with Microsoft (IIRC, numbers bandied about were US $40K) because of what we were doing with the GINA. We actually ended up speaking to Microsoft developers - who couldn't answer our questions. We ended up reverse engineering the MS GINA to find out how to set everything up correctly. It was interesting to note that the publically available GINA documentation improved substantially a couple of years later when Windows 2000 came out. Perhaps the developers felt ashamed that their customers had to resort to reverse engineering because this expensive support contract fell flat.
Currently, most of the power (where you live) might be generated by burning coal, but large scale power generation is:
- vastly more efficient at turning chemical energy into something usable compared to small scale (small scale being a car engine) - vastly easier to make clean via scrubbing 'bad stuff' out the exhaust than small scale catalytic converters on cars
An electric car can also get better efficiencies by using things like regenerative braking - instead of just turning kinetic energy into heat when you want to slow down, you turn the KE back into electricity which you put back into the car's battery instead. Electric cars also don't use energy while stopped at a traffic light or stuck in a traffic jam unlike an idling car engine.
The other thing about electric cars is if you start migrating your power generation to nuclear, wind, solar (etc). no one has to go and buy a new car. No infrastructure has to be changed. However, to move from (say) gasoline cars to hydrogen cars, everyone has to buy new cars and all the infrastructure has to be changed.
I guess they must have; not only did I set it up within 10 minutes, I instructed someone how to set it up who had never set up a VPN before in around 10 minutes. The example that comes with OpenVPN is just about ready to go for a simple preshared key setup - just substitute your own information where necessary.
I don't know when you tried it, but when I did (recently) there was a 'quick setup guide' and it took me less than 10 minutes to set up with a simple pre-shared key.
Dvorak is also the idiot who whined about how the Windows Idle Process was using 98% of CPU time and causing his computer to thrash. He simply doesn't have a clue. Why Slashdot isn't just rejecting submissions out flat that contain the word 'Dvorak' is a mystery (I mean they could even do it with a regular expression).
If Apple is also a failure and being beat upon by Gates like a rented mule...then give me some of that failure! It must be wonderful to be beat like a mule and to be an utter failure, after all, your stock price quadruples over the period of a couple of years. I think more of us can do with failure like that!
It's hardly worthless. The majority of what you might do with Solaris from an admin point of view is pretty much the same whether you're running Linux, BSD, AIX or whatever.
Slashdot Mirror
Of course the lives we lead is natural. It's in human nature to build cities and factories, therefore we live natural human lives. It is no less natural for a human to eat processed food (which is made by humans of course) than it is for a beaver to live in a lodge.
The human body is omnivorous, NOT herbivorous.
I eat an omnivorous diet which occasionally involves a rare steak amongst other things. Since I joined the workforce in 1990, I have had exactly one day off sick, and it wasn't serious enough to warrant seeing the doctor. The only time I've seen a doctor in this time were for routine medicals (aviation, employment etc). and for an injury.
The compiler would have to be able to read minds to throw a warning on that (it won't, even with -Wall). It's a perfectly sane thing to compare a function pointer to NULL in most code. Perhaps we need -Wdowhatimeannotwhatisay.
A few machines? All of the most popular distros (Ubuntu, Fedora Core, Debian, CentOS) have free update services using yum or apt, and have had these mechanisms for years.
The truth sometimes hurts. Theo de Raadt just doesn't dress it up. I wouldn't hire Theo as a diplomat (well, not unless I wanted to actually start a war), but I would hire him as someone who can write secure code. I don't care if he has no social skills; I'm after secure code. That's why we use OpenBSD for security critical things.
From the facility's website:
http://www.sns.gov/aboutsns/benefits.htm
They don't have to open source the drivers. They just need to publish the card's specs and someone else will write the drivers.
Fat chance. Back in the day, when 2D acceleration was the in thing and Linux just had X support, S3 were one to avoid because unlike Tseng Labs and co. they didn't publish the specs for their card so no X support. I suspect they are still one to avoid today. The chances of them publishing the specs so someone else (i.e. not using up their time) can do the drivers are tiny.
Just one question - surely if you're using an Ethernet switch, and the switch is at one end of the wire, and the computer at the other - where do the collisions come from? How would a token-based network be superior to a switch based network? You don't lose the cost of the switch - IBM Token Ring still had a 'concentrator' in the middle so you wound up with a physical star, just like with switched ethernet (and was significantly more prone to breakage should a card go on the fritz and stop passing the token).
On a server I have that sees plenty of network traffic over switched ethernet, I still have '0' in the interface's collision column.
It's good 'ol Rip-Off Britain again. Don't forget, in the US, sales tax is also much lower than 17.5% (typically 4-6% - that's if Apple are even paying sales tax to the various states, I have a feeling that they are not).
The 'rip off' effect in Britain is such that the exchange rate for normal purposes is pretty much irrelevant - really, £1 is only worth about $1 (or less) for most things in actual 'purchasing power' - music especially (often, you'll find a CD in Britain is £14 and only $12 in the US).
But my Sierra was made in Germany. I also owned a 1986 Dodge Ram when I lived in the United States - it also suffered from the same kind of reliability problems as the Sierra (hard cold+damp starting/needing frequent tune-ups). It also rusted, but because it was made out of an awful lot more metal, it would take longer to rust through. Most 1970s cars were junkers by the time they were 10 years old; most mid 90s cars are still perfectly good and ran like they did when they came out of the factory.
Ah, the 'they don't make cars like they used to' fallacy.
I've owned older cars. They need a lot of TLC to keep running reliably - my old Mini needed new points/condenser every 6 months (with the oil change), frequent spark plug/HT cable replacements, and I had to have the cylinder head off twice in the time I owned it. Although it was a fun car to drive (and had lots of character), it needed a LOT of maintenance to have any hope of reliability. There was actually a very noticable performance difference after each 6 monthly maintenance - those old cars really did deteriorate that fast. It also suffered from rust.
My next car was a 1984 Sierra. Not much in the way of computer control, but unlike the Mini, I *never ever* had ignition problems with it, and only broke down a couple of times (the clutch cables were very bad and had a tendency to suddenly go). I had much fewer reliability problems with the Sierra. I still had them though. It could be hard starting on a cold damp morning, and would run a bit rough until it was warmed up (automatic chokes were never a good idea). It also rusted.
My current car, a 1995 Audi A4, just runs consistently, day in, day out. It always starts easily whether it's hot or cold, damp or dry. It is galvanized, so despite living on a small windswept island with a very salty atmosphere, it doesn't rust. After I've cleaned it, it looks just as good as the day it came out of the factory - the paint is so much better on it than older cars (most 1970s cars, when 11 years old would be unreliable, faded and rusty). It also drives just as well today as when it came from the factory - it still performs like the book says it should. It never needs tune-ups like carburetted cars like the Sierra or Mini did because the computer keeps it tuned all the time.
Sure, older cars are much easier to work on - but the point is, I don't have to work on my Audi like I had to work on my Sierra (or even more so, the Mini). And my Audi doesn't rust.
Most people hardly use the computing power of a late-90s Mac (typing letters and reading email and browsing the web is hardly CPU intensive) so the lower overall wattage of the machine will mean it's more energy efficient. CPU instructions per watt is pretty irrelevant unless you really are using the processing power, and not having the computer sitting waiting on I/O all the time.
That's always been the case. I was in a school with about 600 people. We had (very advanced for the time - it was in the period 1985 to 1990) a network of BBC Microcomputers - which had a built-in BASIC interpreter (reputedly, one of the best of its era) as well as a built in 6502 assembler (it was the about the only machine of the era which came with an assembler).
We had a large room in the bottom of the science block with about 15 of these machines, plus a Mac Plus. There were a few other BBCs on the network - the geography block had three (and a Mac).
So a school with 600 students. There was no problem getting on to one of these machines at lunch or after school - demand was simply not very high. There were perhaps half a dozen of us who'd actually use these machines to program. The only time the computer room was full and every machine was in use was when a teacher had a class in there, and more often than not the computer was being used as a teaching aid rather than teaching something about the computer itself.
Programming wasn't rewarded for me at HS level (in the 8-bit days), so much to my teachers chagrin, instead of working on their classes, I wrote 6502 asm instead. You must get your priorities right :-)
Of course, my school reports were very, very bad and I failed most of my exams. But in the end it didn't really matter.
Airplanes fly on the principles described by Newton and Bernoulli, *not* Marconi and Turing! Planes would quite happily fly if the FAA didn't exist at all. Sure, airlines would be impractical, but the vast majority of air traffic in the United States (general aviation) spends most of its time flying on nice VFR days not talking to ATC at all.
Oh, I'm not trying to excuse Microsoft for this - in fact, it's really a pointer to why the quality of Microsoft software is so variable, and why Windows and IE has so many serious security problems and other bugs - despite highly talented developers, the suspected lack of any kind of software engineering process almost inevitably results in the kinds of problems we see with their software -- and the difficulty in getting decent support for more obscure parts of Windows even when you're paying for a very expensive support contract.
I suspect some divisions of MS have a decent software process and others don't - much alluded to by the piece the other day about how developers get so much freedom (and developers DO NOT choose to have a software process by their own volition - generally, software process is castor oil to programmers).
Worse still was the talk a while back that Bill Gates wanted to get his hands dirty with Vista (then Longhorn) and spend 50% of his time back on technical stuff. Bill Gates has a reputation for being a bit of a 'prima donna' about any programming he thinks is wrong, and he's been a manager for a while and is probably un-current. Imagine you're a developer on Windows, and Gates has decided to work on a code module you're involved in. Would you have the guts to correct programming mistakes that Gates makes, knowing full well his reputation for hauling developers into his office and haranguing them over their code? Knowing you could get summarily fired? I think the problems with XP will just continue in Vista because of things like this.
I get the distinct impression that Microsoft being tardy to release the documentation for their protocols isn't purely due to malice - or even not malice at all. I get the impression that things like this simply aren't documented INSIDE of Microsoft, and the original developers of the code have left, and MS staff are now busily trying to document a gigantic mass of source code.
I have direct experience of Microsoft having none or inadequate internal documentation (see my latest JE for the full discussion - http://slashdot.org/~Alioth/journal/133996). A quick precis is that we were working with the GINA in the NT 4.0 days, and we had an expensive support contract with Microsoft (IIRC, numbers bandied about were US $40K) because of what we were doing with the GINA. We actually ended up speaking to Microsoft developers - who couldn't answer our questions. We ended up reverse engineering the MS GINA to find out how to set everything up correctly. It was interesting to note that the publically available GINA documentation improved substantially a couple of years later when Windows 2000 came out. Perhaps the developers felt ashamed that their customers had to resort to reverse engineering because this expensive support contract fell flat.
We do. It's called biodiesel. Or in a more simple form, a wood burning fireplace. Or a steam train.
Electricity is like an abstraction layer.
Currently, most of the power (where you live) might be generated by burning coal, but large scale power generation is:
- vastly more efficient at turning chemical energy into something usable compared to small scale (small scale being a car engine)
- vastly easier to make clean via scrubbing 'bad stuff' out the exhaust than small scale catalytic converters on cars
An electric car can also get better efficiencies by using things like regenerative braking - instead of just turning kinetic energy into heat when you want to slow down, you turn the KE back into electricity which you put back into the car's battery instead. Electric cars also don't use energy while stopped at a traffic light or stuck in a traffic jam unlike an idling car engine.
The other thing about electric cars is if you start migrating your power generation to nuclear, wind, solar (etc). no one has to go and buy a new car. No infrastructure has to be changed. However, to move from (say) gasoline cars to hydrogen cars, everyone has to buy new cars and all the infrastructure has to be changed.
I guess they must have; not only did I set it up within 10 minutes, I instructed someone how to set it up who had never set up a VPN before in around 10 minutes. The example that comes with OpenVPN is just about ready to go for a simple preshared key setup - just substitute your own information where necessary.
I don't know when you tried it, but when I did (recently) there was a 'quick setup guide' and it took me less than 10 minutes to set up with a simple pre-shared key.
Dvorak is also the idiot who whined about how the Windows Idle Process was using 98% of CPU time and causing his computer to thrash. He simply doesn't have a clue. Why Slashdot isn't just rejecting submissions out flat that contain the word 'Dvorak' is a mystery (I mean they could even do it with a regular expression).
If Apple is also a failure and being beat upon by Gates like a rented mule...then give me some of that failure! It must be wonderful to be beat like a mule and to be an utter failure, after all, your stock price quadruples over the period of a couple of years. I think more of us can do with failure like that!
It's hardly worthless. The majority of what you might do with Solaris from an admin point of view is pretty much the same whether you're running Linux, BSD, AIX or whatever.