Which law? If so, the law is seriously out of touch with reality--good luck convincing students and professors that they can't use their university email address for personal mail. And I don't see how the public would be particularly benefited by enforcing such a strict separation. (OK, I'm being a bit unfair--handling personal email at the university isn't quite the same as setting up a personal server there. Hard to know exactly where to draw the line, though.)
I would appreciate if you didn't use state owned, state maintained servers, bandwidth, and infrastructure for personal use.
Eh, I don't know. I also tend to err on the side of caution when it comes to using work resources for personal stuff. But when it comes to setting policy I'd rather be lenient--there are also costs to enforcing a rigorous separation between work and personal uses, and I wouldn't expect the small bandwidth/power/whatever savings you get from doing that to be particularly worth it. And at a place like a university, along with the predictable abuses, you also get people that do some pretty cool stuff just for fun, and I'd rather not make those people jump through a lot of red tape.
1) Don't run your own Email Server. It's a pain in the rear, and it'll get blacklisted for being on a consumer subnet anyway.
I've been sending and recieving my own email directly from a debian/unstable box on a home DSL line (speakeasy) for about 4 years. Unless I've forgotten something, I believe the setup was just a matter of apt-get installing exim and answering the questions in the obvious way. This is the only email address I use for personal and work use, I use it pretty heavily, and I've only seen the blacklist problem on one email. Though it may have happened other times without my noticing, of course. And maybe I just lucked out and got network neighbors that don't get us on too many blacklists.
My most serious problem currently is that my current hardware is kind of poor at recovering from power outages. So if I'm travelling a while and something bad happens at exactly the wrong time I may be without email till I get back home.
So it's not without tradeoffs. But I wouldn't describe it as a "pain in the rear" either. In terms of ongoing administration, there's not much beyond the regular apt-get update && apt-get upgrade (or equivalent for your distro).
Whether or not you want a mathematician's career, whatever that is, math is one of those things, like good writing skills, that just tends to be useful background for a lot of things.
Often, people claim that (as pointed out in 'The Magic Cauldron' by ESR) over 95% of software is not for sale (so called 'custom' jobs), but it is ridiculous to expect programmers to bank on the availability of such jobs, especially because they don't get much attention.
You'd rather bank on the other 5%?
I don't believe ESR's numbers either--he's just not reliable. But there *are* plenty of people making a living writing free software. Most of the Linux kernel community does these days, for example.
Yes, I am, but I have no other choice. The time for you Nader-ites to make a difference is before the election. Convince the Dems (your closest allies) to pick the most liberal candidate they can.
Also note that a national election is a process of compromising with 100 million of your fellow citizens, and the fact is that by the time we get down to the final 2 choices, they're both going to be pretty middle-of-the-road. I'll be an activist on the issues, but when it comes time to vote I'd rather be a part of the real process, and that means accepting compromises.
So, for example, 1024 is 1 * 1000 + 0 * 100 + 2 * 10 + 4.
Doesn't everything work out nicely when you use 0 as the placeholder?
So, what we should have is "1-10" being counted as "0-9" and an additional "null" character, so that "10" (11 in "normal" parlance) might be written as "1_" or "1*".
I use ing direct (www.ingdirect.com), another bank with a big online presence, and I've been happy with them. In parcitular, I've never had any trouble with their site under any browser.
And this is because the credit card companies make money coming and going. They make vast sums from merchants, by charging a few percent of the purchase price when you buy something.
Sure. But that doesn't prevent them from wanting to make even more money if they could.
So if they could save X in fraud by spending Y<X on security infrastructure, you'd think they would. Beats me.
Merchants should never be given master account authentication credentials. They should be given signed message digests, or one-time authentication codes, or something along these lines, which are tied to the transaction date and amount. Even if credit card companies just put secureID readouts on their cards it would MASSIVELY cut fraud.
I would have thought so too, but the credit card companies (ast least in the US) seem to have been happy enough just paying the price of fraud for a long time now....
Also it seems fair to assume that most consumer PC's can be hacked. If that's the case, then you need more than just a little token with a private key on it--the token needs its own display so it can, for example, display the transactions it's authorizing instead of relying on the PC's compromised display.
Does anyone here really think the banks are going to pay this money out from their bottom line? They'll recover it from those customers who do protect their identity through increased fees and interest.
The whole "identity theft" terminology is screwed up; it's not your "identity" you're protecting--you're still you after someone else manages to clear out your checking account. What the "identity thief" has done is to fool the bank's authentication system into thinking their transactions were authorized by you. You do have some control over whether this happens, by your choice of password, choice of when to type it in, etc. But the decisions with the greatest affect on the security of that authentication system are completely in the bank's hands: e.g. the decision to authenticate you by asking you to enter a password into a form on a web page.
The decision to make banks responsible for losses isn't because of a preference for consumers over banks--as you point out, expenses may be passed on to customers either way--it's because the best way to make the banking system more secure is to make sure that the entities with the most power to fix the system are the ones that see the incentives to fix it.
This is the same reason we limit consumer's liability for credit card losses--it's the credit card company that's in the best position to detect and prevent fraud, and if we pass on the cost to them then we enable them to weigh the costs of fraud against the costs of improved security infrastructure, something that's impossible for an individual consumer to do.
By staying so true to the events of the book it loses a little something. The book can narrate a lot of fealing, thoughts and emotions. A film needs to show events to explain that kind of thing.
This is a point that was really brought home to me by watching the Lord of the Rings movies. Tolkein explains directly that Farimir is tempted to take the ring back to Gondor, whereas in the movie Farimir is actually shown taking Frodo partway back, getting into a battle, then releasing him. His indecision is transcribed into action.
Those movies were interesting to watch because, of all the movie adaptations I've seen, they were the ones where I was most familiar with the original text. So it was fun watching them make these kinds of adaptation decisions, even when I wasn't sure I loved the result.
"Each game has it's own limitations, and each fight has it's own strategy. Many British generals decried the foul tactics of the American Revolutionary army as they used guerilla tactics. Come out and fight us face to face! Line up in bright colors! Don't attack during tea time!"
Fortunately, we're playing a game, not shooting at each other on actual battlefields. So fun counts, and players have some responsibility to make sure other players are having fun.
Most games are meant to be fun if everyone just forgets all that and plays to win. But they don't always work that way all the time.
Insisting that device manufacturers need to have on-staff kernel hackers in order to keep ahead of a frequently-changing kernel makes it that much harder for manufacturers to support Linux as a viable alternative.
They have on-staff kernel hackers right now. (Unless those ATI and NVidia binary drivers I keep hearing about are written by the Linux Kernel Fairy.)
Merging their code with the mainline linux kernel should allow them to share their maintenance burden, not increase it....
The point is you should *always* be able to do this wihtin the same major kernel version.
Note that there isn't a stable binary interface even within one kernel version right now. Config options (like CONFIG_SMP) can change binary interfaces.
So a stable binary interface would require extremely fundamental changes.
For some drivers we already do have an API. For example, sound cards can use ALSA instead of coding directly to the kernel.
Most of the interfaces the ALSA driver uses are the same as any other driver. For that matter, do you have information to support the claim that in-kernel alsa-specific interfaces have been stable? I very much doubt it....
In my world you don't get a lot of credit for thinking up an idea that can be explained in two sentences.
So E=MC2 is probably small potatoes to you, isn't it?
If all Einstein produced was those 5 symbols, his work would have been (quite correctly) ignored. There's a larger theory behind that equation, which he worked out in detail, and explained carefully in papers. He wasn't just a guy saying "hey, maybe matter and energy are related somehow". He worked out the details and explained them to us. (And for all its elegance you're not going to explain special relativity in two sentences. The same can't be said for hyperlinks or micropayments.)
Also, Linux isn't a pure monolithic kernel. Almost everything is loaded as a module these days, and modules can communicate through message passing, and can even be implemented is userspace in some cases (e.g. filesystem drivers with FUSE). Even the direct function calling that modules do can be seen as just an optimization of what MINIX does with message passing.
A microkernel runs in a separate address space--that's why it is can't be brought down by driver bugs. Modules don't have that property--they're just bits of code that are loaded dynamically but have rights to all the same stuff as any other bit of kernel code. So modules aren't microkernel-like, but FUSE filesystems (which run in user space--they aren't kernel modules) arguably are.
Remember, this is a guy who thought up hypertext and micro payments at a time when people were literally telling him he was insane.
In my world you don't get a lot of credit for thinking up an idea that can be explained in two sentences. The people I admire are the ones that actually roll up their sleeves and figure out the details.
Tons of people have "thought up" hypertext and micro payments. They're simple ideas. The reason you didn't hear people talk about them much before is that the other bits of infrastructure that needed to be in place to make them work really well weren't there yet, so it didn't make sense to talk about them yet. (And maybe that's still true for micropayments....)
Does this mean that 28% of my er.. body belongs to someone else?
Presumably the patents are phrased in such a way that they effectively give the holders monopolies on the commercial exploitation of those genes without claiming the right to demand permission for anyone's cells to continue transcribing their own DNA....
In some ways this makes more sense to me than software patents: we granted these temporary monopolies in order to encourage the (large) investment required to produce the gene data. By comparison it requires little or no effort to come up with something like one-click shopping, so it shouldn't require special incentives to encourage that kind of work.
"I, a typical citizen, have a great idea. I patent it. Unfortunately, I have $20k in credit card debt and $450 in my checking account. So, while I have a great idea, I lack the means to produce that idea. What to do?"
Where's the problem? The patent system doesn't exist to help people with their credit card debt.
If the idea is so great, then someone else will figure it out soon enough. Probably someone with the means to develop it. The world's better off if they just do it, instead of first making them waste time searching for anyone that might have had the same idea before and getting their permission.
"If you go with the model of "you can only patent what you can build" -- you will squeeze out all the "little guys" with limited resources."
The patent system isn't intended as some sort of charity for tinkerers with "ideas".
It's intended to encourage investment in new ideas, by providing a limited temporary monopoly to support such investment. If we're awarding patents for things that don't require a lot of time and resources to come up with, then the there's something wrong with the system.
Forget the whole myth of the lone inventor walking through the park one day and having that one brilliant world-changing idea. The fact is, *lots* of people have those brilliant world-changing ideas, all the time. Some of those people make it into the history books, but not because of one moment of brilliance, but because they happened to be in the right time and place to develop the idea, or because they did a lot of work getting the details right.
Big ideas are a dime a dozen. The hard work is always in the details.
"Russell showed that the domain of logic gets tangled if you use it to think about itself. Well (with hindsight) that is no surprise at all."
The vague idea of Gödel's proof, if I remember right, was to show that any consistent theory was incomplete--that is to say, either your axioms disagree with each other somehow, or there are theorems out there which can't be proved or disproved. His proof involved a construction that was, as you say, rather self-referential and "tangled". This does not at all imply that it is *only* such self-referential statements are problematic. There may be lots of other innocent-seeming theorems which are fundamentally unproveable.
And as a practical matter it's easy to find theorems in mathematics that are extremely difficult (possibly impossible) to prove. (Though *proving* they're impossible to prove seems to always be virtually impossible.)
In short, math is hard. There are some important problems that it does solve completely, but pick a practical problem at random, something that you need fixed today, and it may not be there for you.
--Bruce Fields (an algebraist, definitely not a logician...)
Slashdot Mirror
Which law? If so, the law is seriously out of touch with reality--good luck convincing students and professors that they can't use their university email address for personal mail. And I don't see how the public would be particularly benefited by enforcing such a strict separation. (OK, I'm being a bit unfair--handling personal email at the university isn't quite the same as setting up a personal server there. Hard to know exactly where to draw the line, though.)
Eh, I don't know. I also tend to err on the side of caution when it comes to using work resources for personal stuff. But when it comes to setting policy I'd rather be lenient--there are also costs to enforcing a rigorous separation between work and personal uses, and I wouldn't expect the small bandwidth/power/whatever savings you get from doing that to be particularly worth it. And at a place like a university, along with the predictable abuses, you also get people that do some pretty cool stuff just for fun, and I'd rather not make those people jump through a lot of red tape.
I've been sending and recieving my own email directly from a debian/unstable box on a home DSL line (speakeasy) for about 4 years. Unless I've forgotten something, I believe the setup was just a matter of apt-get installing exim and answering the questions in the obvious way. This is the only email address I use for personal and work use, I use it pretty heavily, and I've only seen the blacklist problem on one email. Though it may have happened other times without my noticing, of course. And maybe I just lucked out and got network neighbors that don't get us on too many blacklists.
My most serious problem currently is that my current hardware is kind of poor at recovering from power outages. So if I'm travelling a while and something bad happens at exactly the wrong time I may be without email till I get back home.
So it's not without tradeoffs. But I wouldn't describe it as a "pain in the rear" either. In terms of ongoing administration, there's not much beyond the regular apt-get update && apt-get upgrade (or equivalent for your distro).
Whether or not you want a mathematician's career, whatever that is, math is one of those things, like good writing skills, that just tends to be useful background for a lot of things.
You'd rather bank on the other 5%?
I don't believe ESR's numbers either--he's just not reliable. But there *are* plenty of people making a living writing free software. Most of the Linux kernel community does these days, for example.
Also note that a national election is a process of compromising with 100 million of your fellow citizens, and the fact is that by the time we get down to the final 2 choices, they're both going to be pretty middle-of-the-road. I'll be an activist on the issues, but when it comes time to vote I'd rather be a part of the real process, and that means accepting compromises.
No, they're the same thing. Given the decimal representation of an integer, d_nd_{n-1}...d_1d_0, its value is
d_n * 10^n + d_{n-1} * 10^{n-1} + ... + d_1 * 10^1 + d_0 * 10^0
So, for example, 1024 is 1 * 1000 + 0 * 100 + 2 * 10 + 4.
Doesn't everything work out nicely when you use 0 as the placeholder?
Do you also pine for the days of roman numerals?
I use ing direct (www.ingdirect.com), another bank with a big online presence, and I've been happy with them. In parcitular, I've never had any trouble with their site under any browser.
Sure. But that doesn't prevent them from wanting to make even more money if they could.
So if they could save X in fraud by spending Y<X on security infrastructure, you'd think they would. Beats me.
I would have thought so too, but the credit card companies (ast least in the US) seem to have been happy enough just paying the price of fraud for a long time now....
Also it seems fair to assume that most consumer PC's can be hacked. If that's the case, then you need more than just a little token with a private key on it--the token needs its own display so it can, for example, display the transactions it's authorizing instead of relying on the PC's compromised display.
The whole "identity theft" terminology is screwed up; it's not your "identity" you're protecting--you're still you after someone else manages to clear out your checking account. What the "identity thief" has done is to fool the bank's authentication system into thinking their transactions were authorized by you. You do have some control over whether this happens, by your choice of password, choice of when to type it in, etc. But the decisions with the greatest affect on the security of that authentication system are completely in the bank's hands: e.g. the decision to authenticate you by asking you to enter a password into a form on a web page.
The decision to make banks responsible for losses isn't because of a preference for consumers over banks--as you point out, expenses may be passed on to customers either way--it's because the best way to make the banking system more secure is to make sure that the entities with the most power to fix the system are the ones that see the incentives to fix it.
This is the same reason we limit consumer's liability for credit card losses--it's the credit card company that's in the best position to detect and prevent fraud, and if we pass on the cost to them then we enable them to weigh the costs of fraud against the costs of improved security infrastructure, something that's impossible for an individual consumer to do.
How is "it is less confusing this way" a bogus usability argument?
This is a point that was really brought home to me by watching the Lord of the Rings movies. Tolkein explains directly that Farimir is tempted to take the ring back to Gondor, whereas in the movie Farimir is actually shown taking Frodo partway back, getting into a battle, then releasing him. His indecision is transcribed into action.
Those movies were interesting to watch because, of all the movie adaptations I've seen, they were the ones where I was most familiar with the original text. So it was fun watching them make these kinds of adaptation decisions, even when I wasn't sure I loved the result.
Why "smaller projects" in particular? It's designed to scale well with tree size, and people are using it on the Linux kernel source.
"Each game has it's own limitations, and each fight has it's own strategy. Many British generals decried the foul tactics of the American Revolutionary army as they used guerilla tactics. Come out and fight us face to face! Line up in bright colors! Don't attack during tea time!" Fortunately, we're playing a game, not shooting at each other on actual battlefields. So fun counts, and players have some responsibility to make sure other players are having fun. Most games are meant to be fun if everyone just forgets all that and plays to win. But they don't always work that way all the time.
They have on-staff kernel hackers right now. (Unless those ATI and NVidia binary drivers I keep hearing about are written by the Linux Kernel Fairy.)
Merging their code with the mainline linux kernel should allow them to share their maintenance burden, not increase it....
Note that there isn't a stable binary interface even within one kernel version right now. Config options (like CONFIG_SMP) can change binary interfaces.
So a stable binary interface would require extremely fundamental changes.
Most of the interfaces the ALSA driver uses are the same as any other driver. For that matter, do you have information to support the claim that in-kernel alsa-specific interfaces have been stable? I very much doubt it....
--Bruce Fields
If all Einstein produced was those 5 symbols, his work would have been (quite correctly) ignored. There's a larger theory behind that equation, which he worked out in detail, and explained carefully in papers. He wasn't just a guy saying "hey, maybe matter and energy are related somehow". He worked out the details and explained them to us. (And for all its elegance you're not going to explain special relativity in two sentences. The same can't be said for hyperlinks or micropayments.)
A microkernel runs in a separate address space--that's why it is can't be brought down by driver bugs. Modules don't have that property--they're just bits of code that are loaded dynamically but have rights to all the same stuff as any other bit of kernel code. So modules aren't microkernel-like, but FUSE filesystems (which run in user space--they aren't kernel modules) arguably are.
In my world you don't get a lot of credit for thinking up an idea that can be explained in two sentences. The people I admire are the ones that actually roll up their sleeves and figure out the details.
Tons of people have "thought up" hypertext and micro payments. They're simple ideas. The reason you didn't hear people talk about them much before is that the other bits of infrastructure that needed to be in place to make them work really well weren't there yet, so it didn't make sense to talk about them yet. (And maybe that's still true for micropayments....)
Presumably the patents are phrased in such a way that they effectively give the holders monopolies on the commercial exploitation of those genes without claiming the right to demand permission for anyone's cells to continue transcribing their own DNA....
In some ways this makes more sense to me than software patents: we granted these temporary monopolies in order to encourage the (large) investment required to produce the gene data. By comparison it requires little or no effort to come up with something like one-click shopping, so it shouldn't require special incentives to encourage that kind of work.
--b.
Where's the problem? The patent system doesn't exist to help people with their credit card debt.
If the idea is so great, then someone else will figure it out soon enough. Probably someone with the means to develop it. The world's better off if they just do it, instead of first making them waste time searching for anyone that might have had the same idea before and getting their permission.
The patent system isn't intended as some sort of charity for tinkerers with "ideas".
It's intended to encourage investment in new ideas, by providing a limited temporary monopoly to support such investment. If we're awarding patents for things that don't require a lot of time and resources to come up with, then the there's something wrong with the system.
Forget the whole myth of the lone inventor walking through the park one day and having that one brilliant world-changing idea. The fact is, *lots* of people have those brilliant world-changing ideas, all the time. Some of those people make it into the history books, but not because of one moment of brilliance, but because they happened to be in the right time and place to develop the idea, or because they did a lot of work getting the details right.
Big ideas are a dime a dozen. The hard work is always in the details.
The vague idea of Gödel's proof, if I remember right, was to show that any consistent theory was incomplete--that is to say, either your axioms disagree with each other somehow, or there are theorems out there which can't be proved or disproved. His proof involved a construction that was, as you say, rather self-referential and "tangled". This does not at all imply that it is *only* such self-referential statements are problematic. There may be lots of other innocent-seeming theorems which are fundamentally unproveable.
And as a practical matter it's easy to find theorems in mathematics that are extremely difficult (possibly impossible) to prove. (Though *proving* they're impossible to prove seems to always be virtually impossible.)
In short, math is hard. There are some important problems that it does solve completely, but pick a practical problem at random, something that you need fixed today, and it may not be there for you.
--Bruce Fields (an algebraist, definitely not a logician...)