Not to nitpick, but high school was recent enough for me to remember that my life/was/ stressful, and that I/didn't/ have time for other activities -- useful things like self-directed computer science study and piano lessons.
"As a result, they advocate applying the voodoo [a 35-write pass] to PRML and EPRML drives even though it will have no more effect than a simple scrubbing with random data."
No, it's not. Slashdot even did a story on this a while back: data on modern hard disks can't be recovered after being overwritten just once, even with electron microscopes.
I'm going to jump in here because I've been doing a lot of thinking on this exact issue. I'm atheist, by the way.
As an atheist, I don't think there's any sort of God out there making up morals for us, and it's pretty clear that the laws of physics don't directly imply a morality. So, "morality" as a word and concept can only mean some physically arbitrary principle or set of rules defined by humanity.
The moral relativists catch on to this. Where they go wrong is by saying, "therefore, anything I want to say is equally valid to any of your morals". To put it bluntly, that's semantic nonsense. Morality is a word. Useful words mean things. Morality is a useful concept to many people, and it means something to them. But, since it hasn't been examined mathematically, it doesn't mean anything precise.
So I've been working on making it precise. It is inconsistent when imprecise, so I'm looking for the largest consistent subset of moral concepts that doesn't seem to imply anything totally nonsensical. I look at my approach to the task as similar to debugging someone else's computer program. I didn't make up morality, but I'm making it work right. My product so far is a Utilitarian-based scheme that works pretty well for anything I think I'm likely to come up with, but has problems with regard to the rights of animals, fetuses, children, sentient computers, and those in persistent vegetative states. All of those problems could be defined away trivially in the code, but that would be hack. I don't like hacks, so I'm thinking of what the "right" solution to that problem is.
Once my morality is complete, my claim won't be that it is divinely or universally inspired. It will be that it is self-consistent, intuitively correct, and matches my views of how I and others should behave. I won't care whether you agree with me, except to the extent that I might have to change my own actions toward you if you are prone to immoral behaviors. If you don't agree with me and claim that "morality" is something different to you, I'll just say that we're speaking slightly different languages. Whether or not you agree with me, you could get my attention by pointing out inconsistencies in my work or absurd or unexpected results that follow from it. This is much like the approach taken by mathematics today. In fact, my morality will/be/ a form of math, though a form not likely to be interesting mathematically.
This, I think, is the solution to your argument. It is not to speculate the highly unlikely existence of some supreme being, or to (usually unintentionally) claim that the term "morality" has no meaning. It is to define, logically, what you mean by the term, and then to run with that and see where you can get with it.
I've written more about this on my web log (linked to in the home page), and you can randomly comment there if you want to talk to me about it. I may eventually publish my "morality document", but it's really not ready yet.
If your ISP is, in effect, providing you with a hacked web browser, then, yeah, the people who use it are stuck. But, only those people: those who used their own browsers would object, some quite strenuously, to their ISP doing a man-in-the-middle attack on every SSL website, so that scheme wouldn't be workable in practice. Employers can (sometimes) get away with the antics they pull because they're paying you. I wouldn't visit my bank from work unless I were booted from a CD anyway (and that's against policy at a lot of places, so I'd probably just check from home).
Actually, the only significant problem with his proposal is that a high-traffic website would have significantly higher server processing costs if it had to encrypt everything. There are no known breaks for SSL right now, so it's highly unlikely anyone is reading your https traffic except the website on the other end of the connection. An ISP would certainly be able to purchase a certificate for itself, but that certificate would be useless for MITM because a legitimate certificate authority won't knowingly issue a certificate for a DNS domain not under that ISP's control.
Well, that's not what I meant. If doing shift-add sequences were/always/ better, there would be no reason for the imul instruction to exist at all. What I am suggesting is using a shift instruction instead of a multiply in the easy cases, like when you're multiplying by a power of 2.
I did my own test on my laptop just now for multiplying by 2, and I couldn't even get gcc to output a multiply instruction for that even when I passed it -O0. After editing the assembly and recompiling, I tried it, and... the multiply was still faster. I figured this was probably because I was setting the destination to one directly afterwards, and register renaming was figuring this out and doing the loop in parallel, and there were probably more multipliers than shifters. So, I took out the redefinition, and the shift was faster afterwards, as I'd expected. It was about 400ms for multiply and 350ms for shifting, for 10 million iterations.
Modern architectures are really complex, but replacing power of 2 multiplies with bit shifts is still usually a win on them. You might have a point for pathological cases where you have the multiplier free but the shifter not free, but it's still a good general rule, and compilers still do it.
> I've seen programmers who grew up with Pascal carefully turning multiplies into sequences of adds and shifts. Great, except that something like the Athlon can do two multiplies in parallel, but only one shift at a time (because most code contains a lot more multiplies than shifts), stalling the pipeline.
That strength reduction of multiplication to bit shifting might not be valuable anymore for modern CPUs surprised me, so I looked up the instruction latency and throughput for the Core 2. To see for yourself, download http://download.intel.com/design/processor/manuals/248966.pdf and look at Appendix C.
The timings show that integer multiply, IMUL, is still quite a bit slower from the bit shift instructions (SHL/SHR). So, turning multiplies into adds and shifts is still a Good Thing.
However, these days, a compiler can do this type of thing for you much of the time, so doing the transformation by hand is usually not necessary. Also, since optimizing the source code in this way does take time and decreases source readability, the only cases where people should really still be doing manual strength reduction are in performance-critical code where the compiler, for whatever reason, is missing the optimization opportunity.
Actually, Linus was, as he sometimes is, completely clueless. He's unaware of the fact that filesystem journaling was *NEVER* intended to give better data integrity guarantees than an ext2-crash-fsck cycle and that the only reason for journaling was to alleviate the delay caused by fscking. All the filesystem can normally promise in the event of a crash is that the metadata will describe a valid filesystem somewhere between the last returned synchronization call and the state at the event of the crash. If you need more than that -- and you really, probably don't -- you have to do special things, such as running an OS that never, ever, ever crashes and putting a special capacitor in the system so the OS can flush everything to disk before the computer loses power in an outage.
As I see it, the only bad thing about a communist economy is that it leads to economic inefficiency due to disincentivizing labor and entrepreneurship. Since my argument supporting a property or land tax was based specifically on avoiding the inefficient use of economic resources, superficial similarities between a property or land tax and communism do not trouble me.
>If someone proved to you that it would make society more efficient to chain up your family and force them to pick cotton, would you happily accept that? May I pick your pocket if an economist tells me I can use your cash more productively than you? That is what you seem to be saying -- no property rights except where society finds them efficient.
The answer to both of those questions (assuming the economist were correct in the second example) is yes, according to me. It's slightly more correct to say "it would be moral in my view to do those things". I am not certain that I am such a good person I would submit to being a slave or to seeing those I love enslaved even if I did believe it to be the moral thing to do.
Now, there are several good economic arguments why slavery is not in fact an efficient use of human labor and why government redistribution of wealth is superior to letting pickpockets run free, so please do not go around saying, "LINUXROCKS123 SUPPORTS SLAVERY" or "THIS GUY WANTS TO LEGALIZE THIEVERY", because that would be misrepresenting my views.
Well, by keeping the land for yourself, you are preventing others from using it. There's a lot of stuff that can be done with land: you can farm it, build a house or business complex on it, drill for oil on it, etc. Moreover, not all land is equal; it differs by location. Some areas (like Japan and California) are very short on land, and it is a very valuable resource.
"Land speculation", where you just hold land as an investment and don't do anything on or with the land, prevents others from doing things on or with the land that are more useful to society than just sitting on it. Property tax decreases the incentive for land speculation or use a field of garden gnomes (unless the pleasure gained from looking at the garden gnomes outweighs its other productive uses). As long as the tax is not raised higher than the productive value of the land in a year, someone will still be willing to buy the land and use it productively (though a property tax, unlike a pure land tax, does decrease the incentive to develop the land).
If you think holding land is a fundamental right, I respectfully disagree. Property and free market rights should only exist to the extent they lead to the efficient use of society's resources.
Theoretically, you should not need motivation to sell your land, because, if you're using your land inefficiently, the free market will take care of providing a buyer who is willing to pay you enough to motivate you to sell your land to him. However, in practice, property tax might be necessary to provide an incentive for someone to sell land that he doesn't really need but just wants to keep and use inefficiently since he happened to have it. Property tax in this case provides an incentive not to do that: cost motivates some people much more than lost profit.
Property tax isn't communism. It's very close to a land tax, which is the economically perfect tax. Calling one thing something else doesn't make it so, it just makes it harder to talk with you.
I could have modded you down, but I decided to argue with you instead. I knocked off my karma bonus in case I fed a troll here.
I think they probably added support in 3 and not 3.0.1, since switching from X11 to Aqua would be kind of a big change for a point release. I don't know, though; I've never used OOo for Mac. I'm glad you have something that works for you with NeoOffice.
Slashdot Mirror
That article was written quite a while ago. Those issues are likely resolved by now.
Oops: http://www.computerandvideogames.com/article.php?id=165699
Did they eventually give you the $200 back? What bank was that?
---linuxrocks123
I don't believe you about the banana peels: http://en.wikipedia.org/wiki/Bananadine
Not to nitpick, but high school was recent enough for me to remember that my life /was/ stressful, and that I /didn't/ have time for other activities -- useful things like self-directed computer science study and piano lessons.
Peter Gutmann also says,
"As a result, they advocate applying the voodoo [a 35-write pass] to PRML and EPRML drives even though it will have no more effect than a simple scrubbing with random data."
---linuxrocks123
No, it's not. Slashdot even did a story on this a while back: data on modern hard disks can't be recovered after being overwritten just once, even with electron microscopes.
Please tell me you're not speaking from experience...
---linuxrocks123
I'm going to jump in here because I've been doing a lot of thinking on this exact issue. I'm atheist, by the way.
As an atheist, I don't think there's any sort of God out there making up morals for us, and it's pretty clear that the laws of physics don't directly imply a morality. So, "morality" as a word and concept can only mean some physically arbitrary principle or set of rules defined by humanity.
The moral relativists catch on to this. Where they go wrong is by saying, "therefore, anything I want to say is equally valid to any of your morals". To put it bluntly, that's semantic nonsense. Morality is a word. Useful words mean things. Morality is a useful concept to many people, and it means something to them. But, since it hasn't been examined mathematically, it doesn't mean anything precise.
So I've been working on making it precise. It is inconsistent when imprecise, so I'm looking for the largest consistent subset of moral concepts that doesn't seem to imply anything totally nonsensical. I look at my approach to the task as similar to debugging someone else's computer program. I didn't make up morality, but I'm making it work right. My product so far is a Utilitarian-based scheme that works pretty well for anything I think I'm likely to come up with, but has problems with regard to the rights of animals, fetuses, children, sentient computers, and those in persistent vegetative states. All of those problems could be defined away trivially in the code, but that would be hack. I don't like hacks, so I'm thinking of what the "right" solution to that problem is.
Once my morality is complete, my claim won't be that it is divinely or universally inspired. It will be that it is self-consistent, intuitively correct, and matches my views of how I and others should behave. I won't care whether you agree with me, except to the extent that I might have to change my own actions toward you if you are prone to immoral behaviors. If you don't agree with me and claim that "morality" is something different to you, I'll just say that we're speaking slightly different languages. Whether or not you agree with me, you could get my attention by pointing out inconsistencies in my work or absurd or unexpected results that follow from it. This is much like the approach taken by mathematics today. In fact, my morality will /be/ a form of math, though a form not likely to be interesting mathematically.
This, I think, is the solution to your argument. It is not to speculate the highly unlikely existence of some supreme being, or to (usually unintentionally) claim that the term "morality" has no meaning. It is to define, logically, what you mean by the term, and then to run with that and see where you can get with it.
I've written more about this on my web log (linked to in the home page), and you can randomly comment there if you want to talk to me about it. I may eventually publish my "morality document", but it's really not ready yet.
Best regards,
---linuxrocks123
If your ISP is, in effect, providing you with a hacked web browser, then, yeah, the people who use it are stuck. But, only those people: those who used their own browsers would object, some quite strenuously, to their ISP doing a man-in-the-middle attack on every SSL website, so that scheme wouldn't be workable in practice. Employers can (sometimes) get away with the antics they pull because they're paying you. I wouldn't visit my bank from work unless I were booted from a CD anyway (and that's against policy at a lot of places, so I'd probably just check from home).
Who is your employer, by the way?
---linuxrocks123
Actually, the only significant problem with his proposal is that a high-traffic website would have significantly higher server processing costs if it had to encrypt everything. There are no known breaks for SSL right now, so it's highly unlikely anyone is reading your https traffic except the website on the other end of the connection. An ISP would certainly be able to purchase a certificate for itself, but that certificate would be useless for MITM because a legitimate certificate authority won't knowingly issue a certificate for a DNS domain not under that ISP's control.
Have a nice day,
---linuxrocks123
Well, that's not what I meant. If doing shift-add sequences were /always/ better, there would be no reason for the imul instruction to exist at all. What I am suggesting is using a shift instruction instead of a multiply in the easy cases, like when you're multiplying by a power of 2.
I did my own test on my laptop just now for multiplying by 2, and I couldn't even get gcc to output a multiply instruction for that even when I passed it -O0. After editing the assembly and recompiling, I tried it, and ... the multiply was still faster. I figured this was probably because I was setting the destination to one directly afterwards, and register renaming was figuring this out and doing the loop in parallel, and there were probably more multipliers than shifters. So, I took out the redefinition, and the shift was faster afterwards, as I'd expected. It was about 400ms for multiply and 350ms for shifting, for 10 million iterations.
Modern architectures are really complex, but replacing power of 2 multiplies with bit shifts is still usually a win on them. You might have a point for pathological cases where you have the multiplier free but the shifter not free, but it's still a good general rule, and compilers still do it.
---linuxrocks123
> I've seen programmers who grew up with Pascal carefully turning multiplies into sequences of adds and shifts. Great, except that something like the Athlon can do two multiplies in parallel, but only one shift at a time (because most code contains a lot more multiplies than shifts), stalling the pipeline.
That strength reduction of multiplication to bit shifting might not be valuable anymore for modern CPUs surprised me, so I looked up the instruction latency and throughput for the Core 2. To see for yourself, download http://download.intel.com/design/processor/manuals/248966.pdf and look at Appendix C.
The timings show that integer multiply, IMUL, is still quite a bit slower from the bit shift instructions (SHL/SHR). So, turning multiplies into adds and shifts is still a Good Thing.
However, these days, a compiler can do this type of thing for you much of the time, so doing the transformation by hand is usually not necessary. Also, since optimizing the source code in this way does take time and decreases source readability, the only cases where people should really still be doing manual strength reduction are in performance-critical code where the compiler, for whatever reason, is missing the optimization opportunity.
---linuxrocks123
You laugh, but...
http://www.physorg.com/news5619.html
---linuxrocks123
Actually, Linus was, as he sometimes is, completely clueless. He's unaware of the fact that filesystem journaling was *NEVER* intended to give better data integrity guarantees than an ext2-crash-fsck cycle and that the only reason for journaling was to alleviate the delay caused by fscking. All the filesystem can normally promise in the event of a crash is that the metadata will describe a valid filesystem somewhere between the last returned synchronization call and the state at the event of the crash. If you need more than that -- and you really, probably don't -- you have to do special things, such as running an OS that never, ever, ever crashes and putting a special capacitor in the system so the OS can flush everything to disk before the computer loses power in an outage.
As I see it, the only bad thing about a communist economy is that it leads to economic inefficiency due to disincentivizing labor and entrepreneurship. Since my argument supporting a property or land tax was based specifically on avoiding the inefficient use of economic resources, superficial similarities between a property or land tax and communism do not trouble me.
---linuxrocks123
>If someone proved to you that it would make society more efficient to chain up your family and force them to pick cotton, would you happily accept that? May I pick your pocket if an economist tells me I can use your cash more productively than you? That is what you seem to be saying -- no property rights except where society finds them efficient.
The answer to both of those questions (assuming the economist were correct in the second example) is yes, according to me. It's slightly more correct to say "it would be moral in my view to do those things". I am not certain that I am such a good person I would submit to being a slave or to seeing those I love enslaved even if I did believe it to be the moral thing to do.
Now, there are several good economic arguments why slavery is not in fact an efficient use of human labor and why government redistribution of wealth is superior to letting pickpockets run free, so please do not go around saying, "LINUXROCKS123 SUPPORTS SLAVERY" or "THIS GUY WANTS TO LEGALIZE THIEVERY", because that would be misrepresenting my views.
---linuxrocks123
Well, by keeping the land for yourself, you are preventing others from using it. There's a lot of stuff that can be done with land: you can farm it, build a house or business complex on it, drill for oil on it, etc. Moreover, not all land is equal; it differs by location. Some areas (like Japan and California) are very short on land, and it is a very valuable resource.
"Land speculation", where you just hold land as an investment and don't do anything on or with the land, prevents others from doing things on or with the land that are more useful to society than just sitting on it. Property tax decreases the incentive for land speculation or use a field of garden gnomes (unless the pleasure gained from looking at the garden gnomes outweighs its other productive uses). As long as the tax is not raised higher than the productive value of the land in a year, someone will still be willing to buy the land and use it productively (though a property tax, unlike a pure land tax, does decrease the incentive to develop the land).
If you think holding land is a fundamental right, I respectfully disagree. Property and free market rights should only exist to the extent they lead to the efficient use of society's resources.
---linuxrocks123
Theoretically, you should not need motivation to sell your land, because, if you're using your land inefficiently, the free market will take care of providing a buyer who is willing to pay you enough to motivate you to sell your land to him. However, in practice, property tax might be necessary to provide an incentive for someone to sell land that he doesn't really need but just wants to keep and use inefficiently since he happened to have it. Property tax in this case provides an incentive not to do that: cost motivates some people much more than lost profit.
---linuxrocks123
Property tax isn't communism. It's very close to a land tax, which is the economically perfect tax. Calling one thing something else doesn't make it so, it just makes it harder to talk with you.
I could have modded you down, but I decided to argue with you instead. I knocked off my karma bonus in case I fed a troll here.
---linuxrocks123
Factually incorrect. Garmin also uses Linux; others probably do as well.
---linuxrocks123
Very strange. I guess OOo has lax standards for what is appropriate for a point release :\
---linuxrocks123
I think they probably added support in 3 and not 3.0.1, since switching from X11 to Aqua would be kind of a big change for a point release. I don't know, though; I've never used OOo for Mac. I'm glad you have something that works for you with NeoOffice.
Best regards,
---linuxrocks123
If that's true, this page is very misleading: http://porting.openoffice.org/mac/download/aqua-Intel.html
---linuxrocks123
What you said used to be true, but OpenOffice.org 3 introduced a native Mac port in October of 2008.
Cheers,
---linuxrocks123