200dpi screen is actually much better than a fax, and probably better than a 300dpi laser printer.
Anti aliasing is partly to credit (esp with the 600dpi subpixel arrangement), but more importantly is that a 200dpi pixel is 1/200 inch in size. A 300 dpi laser printer has dots significantly larger than 1/300 inch, so the display will be much crisper. I think. I don't know why I know this, but am pretty sure about it.
I'm suprised it's not written in Java. If all you want is a reference implementation, Java seems to be the way to go. These days, with modern JIT technology, it might even be somewhat acceptable from a performance perspective.
I dunno that I was that unjustified in assuming that edge enhancement was a convoultion with a sharpening mask.
Doyou know the mask that was used to enhance the edges? Ifso, does it have an inverse? (img proc was too long ago to recall if all masks had inverses; I suspect not)
Before you start lambasting me for suggesting to run real-time image processing on vanilla hardware, I'd like to point out that MPEG (ok, I'm guessing about this part -- back me up?) compresses frames with DCT (after motion compensation). DCTs are in frequency space, which can be convolved (ie smoothed) in linear time.
DVDs use MPEG internally, so all you need is to insert a filter before DCT decompression. How to do that when decode is in hardware, I leave up to others, but I'm assuming that the chips must have some hooks for soft-upgrades.
I remember that one. $250 billion was it? Aren't those jokers busy investing money again?
I recall Rolling Stone (they weren't completely lame at that point) having an article about all the things you could do for that much money -- like take a cab ride to Jupiter...
Sorry, i don't understand how this is open to fraud. Every elegible person gets one card/code/slip. If they chose to use it themselves, or give/sell it to Adam for his use, that's hardly fraud.... Or is this exactly what you are talking about.
In which case I am perfectly happy with that sort of fraud. I wouldn't sell my vote, but I don't see why I should stop you doing so. I mean, people spend lots of money trying to indirectly influence me, why not make it concrete?
Thank you, I was going to make that point myself. Please mod that man up!
Anyway, Ob. Topic:
So the FDIC presumably (so add salt to taste) has standards that a bank has to be audited to -- somehting like iso 9001 I'd imagine -- concerning data firewalling, air gaps, and the like. I think any 3 out of 10 slashdotters could draft pretty reasonable guidelines.
The problem is x-fold (don't know how many I can come up with) as I see it:
1) The bank has no interest in being any more secure than it needs to be, as long as it is insurable. I'd imagine premiums are flat-rate, and you don't get a cut if you are over-secure. Security costs money.
2) If the FDIC needs to up premiums 'cause too many banks are insecure, it is easier to pass these costs on to customers than to deal with them. All credit cards (and checks for banks) accept that their system is dead-easy to defraud, but that it is cheaper to accept some loss (passed on to customers as fees) than to fix the system. Apparently you need at least $10000 on a check before the signature is verified.
So relying on insurance means that banks can basically turn their insecurity into a systematic cost that everybody else has to pay. It seems to me that the FDIC should reevaluate whether it insures on-line transactions too, at least without very strict security audits.
I was thinking about this. What is needed is a secure yet anonymous system that guarantees that everyone eligible can vote at most once.
I was thinking that perhaps you let the DMV (they have provisions for dealing with heavy loads of people throughout the year) let people register several weeks/months in advance; they verify your eligiblity and print you out a secret (== long, perhaps on a smartcard) code that you can use over any phone to vote.
The main stubling block is anonymity vs losability. You know some people will lose the card, so the system has to know your number in order to revoke the old one when printing you out a new card. So you either get losability, but this necessitates you lose anonymity. I guess most people would prefer anonymity -- even I could hold onto a smart card for a few days, I think.
Doesn't matter. No third party will win; even a coalition of all the third parties likely wouldn't gain double digits. The whole point is to signal your dissatisfaction with "da system".
mind you, I'd like to vote, but I have taxation without representation. Does anybody know if there is a way I can get my social security withholdings back when I leave the country?
Principles alone get you nowhere. They are meant to be moral guides to help you derive a consistent approach to problems.
When principles get in the way of solving problems, your principles are wrong. yes, wrong as in part of the problem, wrong. If they help you achieve solutions, they are correct. Are nader's and brown's principled stand helping or not? (assume that the problem we are trying to tackle is having some -- any -- influence on domestic policy)
Now. If we stop wallowing for a second and think about how to make an incremental step in the right -- erm. correct -- direction we see that the only non-violent way is through votes and assimilation and by being taken seriously. I post that should a more moderate "loon coalition" be formed that dropped the unimplementable parts from their platform and kept the issues we mostly all agree on, this would be a pareto optimal step in the right direction.
But I'm sure you'll judge for yourself, as my cynical viewpoint has little of value to add. I must admit that I'm not quite in the clear as to wanting things to get done makes me an ugly cynic, tho. Would you care to enlighten me?
Many of these questions are answerable with a flat yes or no, followed by an explanation.
Bush's "answers" with the disturbing exception of the war-on-drugs issue, are crafted from generic statements of his values, how he is not part of washington cognicsenti, why the current administration is evil, and how the system needs to balance this against that.
These answers thus force the reader to read a long screed and from that try to judge a) whether the answer was yes or no, and b) whether the you agree with the justification. The whole point of questions is that the voters want concrete datapoints from which to draw conclusions about the candidate's values (this is called "showing" in expositionary writing). These answers are the opposite: instead "telling" us his values and letting us infer the datapoint.
Had each of these answers been prefaced with an unequivocable yes or no, then I would have been much happier.
As for the FUD allegation, yes. I hadn't thought of it in those terms, but I Fear many of the candiates, I have Uncertainty about their vision and values, and I Doubt they will have a positive effect should they win.
nah, it's just that as a rule teachers are not the savvyest bunch of grapes on the vine. They love to teach (not much point being a teacher for the pay these days) so you're not going to get the sort of person who scrutinises a contract for every loophole and repercussion (sp?).
"well meaning" is probably the best generalisation, as it captures the intuitive notion of the strengths and weaknesses of educators as a generalisation.
This is why the companies who prey on this mentality are such slime, IMHO.
The problem with the 3rd parties is that they are crackpots. Sorry. They have great ideas about the small issues like drugs, personal freedoms, overcommercialisation, and things that we like to blather on about.
But they remain crackpots by shooting themselves in the foot by espousing impossible party platforms -- like Nader's 100% taxation above x times minimum wage or Brown's abolish all government idea. I have no idea what the Naturals stand for, but I'm willing to believe you when you claim they're loons.
If these parties would only tone down their crackpot ideas to something that a somewhat mainstream person (ie doesn't stand out on the "T" for whatever reason) could vote for, and maybe get together in some coalition, then maybe they could stand a chance.
For some reason I believe these people are not interested in realpolitik, tho, and would rather wallow in their principles than make a constructive effort to win.
The point has been made on salon (a self admittedly un-republican 'zine) numerous times that Bush's entire campaign has been a bunch of well crafted soundbites that are strung together in a semi-random fashion to answer pretty much any question.
He very rarely answers the question even indirectly, often resorting to wonderful statements like "I have said what I believe in, and if that is what you are asking, then that is what I believe." OR something equally self-referential.
Gore tends to (tho I notice he hasn't bothered to here) say too much, rather than too little, so he tends to drown his answer in a series of snipes and look-at-me's.
But then I get taxation but no representation, so my opinion is moot.
what about you starting in an area that is completely surrounded by mines?
XXXXX
X434X
X303X
X434X
XXXXX
You click on the 0 to start (numbers may be wrong, sorry). the Xs represent mines. To clear the rest of the board, you have to blindly guess on a square outside your cleared area.
That sounds "unsolvable" to me. So you need blind luck.
NB I realise that this is not the what the article is about.
You'll notice that they speak about board "consistency". Basically, give me a board with a bunch of numbers on it and all other cells marked X. I'll tell you whether there exists a layout of mines (in the Xs, of course) so that your numbers are correct. This proof basically states that the best we can do is to enumerate all possible layouts and check them in turn.
say there are n Xs on the board. If you can come up with an algorithm that runs in O(n^k) time (k some finite integer) or prove that it is impossible to do faster than O(k^n) then you win a million dollars.
They're not asking you to play the game. mutter mutter. Reading comprehension. mutter mutter.
Slashdot Mirror
200dpi screen is actually much better than a fax, and probably better than a 300dpi laser printer.
Anti aliasing is partly to credit (esp with the 600dpi subpixel arrangement), but more importantly is that a 200dpi pixel is 1/200 inch in size. A 300 dpi laser printer has dots significantly larger than 1/300 inch, so the display will be much crisper. I think. I don't know why I know this, but am pretty sure about it.
I'm suprised it's not written in Java. If all you want is a reference implementation, Java seems to be the way to go. These days, with modern JIT technology, it might even be somewhat acceptable from a performance perspective.
I dunno that I was that unjustified in assuming that edge enhancement was a convoultion with a sharpening mask.
Doyou know the mask that was used to enhance the edges? Ifso, does it have an inverse? (img proc was too long ago to recall if all masks had inverses; I suspect not)
So run a gaussian smooth on each frame!
Before you start lambasting me for suggesting to run real-time image processing on vanilla hardware, I'd like to point out that MPEG (ok, I'm guessing about this part -- back me up?) compresses frames with DCT (after motion compensation). DCTs are in frequency space, which can be convolved (ie smoothed) in linear time.
DVDs use MPEG internally, so all you need is to insert a filter before DCT decompression. How to do that when decode is in hardware, I leave up to others, but I'm assuming that the chips must have some hooks for soft-upgrades.
I remember that one. $250 billion was it? Aren't those jokers busy investing money again?
I recall Rolling Stone (they weren't completely lame at that point) having an article about all the things you could do for that much money -- like take a cab ride to Jupiter...
Sorry, i don't understand how this is open to fraud. Every elegible person gets one card/code/slip. If they chose to use it themselves, or give/sell it to Adam for his use, that's hardly fraud. ... Or is this exactly what you are talking about.
In which case I am perfectly happy with that sort of fraud. I wouldn't sell my vote, but I don't see why I should stop you doing so. I mean, people spend lots of money trying to indirectly influence me, why not make it concrete?
If that half an arm was useful for clubbing the unwilling mother-to-be over the head, I'd imagine that the 2.5er would win out.
Or perhaps he could use it to fend off other suitors, but you get my drift.
Thank you, I was going to make that point myself. Please mod that man up!
Anyway, Ob. Topic:
So the FDIC presumably (so add salt to taste) has standards that a bank has to be audited to -- somehting like iso 9001 I'd imagine -- concerning data firewalling, air gaps, and the like. I think any 3 out of 10 slashdotters could draft pretty reasonable guidelines.
The problem is x-fold (don't know how many I can come up with) as I see it:
1) The bank has no interest in being any more secure than it needs to be, as long as it is insurable. I'd imagine premiums are flat-rate, and you don't get a cut if you are over-secure. Security costs money.
2) If the FDIC needs to up premiums 'cause too many banks are insecure, it is easier to pass these costs on to customers than to deal with them. All credit cards (and checks for banks) accept that their system is dead-easy to defraud, but that it is cheaper to accept some loss (passed on to customers as fees) than to fix the system. Apparently you need at least $10000 on a check before the signature is verified.
So relying on insurance means that banks can basically turn their insecurity into a systematic cost that everybody else has to pay. It seems to me that the FDIC should reevaluate whether it insures on-line transactions too, at least without very strict security audits.
I was thinking about this. What is needed is a secure yet anonymous system that guarantees that everyone eligible can vote at most once.
I was thinking that perhaps you let the DMV (they have provisions for dealing with heavy loads of people throughout the year) let people register several weeks/months in advance; they verify your eligiblity and print you out a secret (== long, perhaps on a smartcard) code that you can use over any phone to vote.
The main stubling block is anonymity vs losability. You know some people will lose the card, so the system has to know your number in order to revoke the old one when printing you out a new card. So you either get losability, but this necessitates you lose anonymity. I guess most people would prefer anonymity -- even I could hold onto a smart card for a few days, I think.
Doesn't matter. No third party will win; even a coalition of all the third parties likely wouldn't gain double digits. The whole point is to signal your dissatisfaction with "da system".
mind you, I'd like to vote, but I have taxation without representation. Does anybody know if there is a way I can get my social security withholdings back when I leave the country?
I think they are already digitised, by virtue of being sent to earth by radio waves
<\pedantry>
:-P
so many typos, so little time.
s/post/posit
s/that dropped/which dropped/
s/as to wanting/as to how wanting/
Probably shouldn't start a paragraph with "But" either.
Principles alone get you nowhere. They are meant to be moral guides to help you derive a consistent approach to problems.
When principles get in the way of solving problems, your principles are wrong. yes, wrong as in part of the problem, wrong. If they help you achieve solutions, they are correct. Are nader's and brown's principled stand helping or not? (assume that the problem we are trying to tackle is having some -- any -- influence on domestic policy)
Now. If we stop wallowing for a second and think about how to make an incremental step in the right -- erm. correct -- direction we see that the only non-violent way is through votes and assimilation and by being taken seriously. I post that should a more moderate "loon coalition" be formed that dropped the unimplementable parts from their platform and kept the issues we mostly all agree on, this would be a pareto optimal step in the right direction.
But I'm sure you'll judge for yourself, as my cynical viewpoint has little of value to add. I must admit that I'm not quite in the clear as to wanting things to get done makes me an ugly cynic, tho. Would you care to enlighten me?
Many of these questions are answerable with a flat yes or no, followed by an explanation.
Bush's "answers" with the disturbing exception of the war-on-drugs issue, are crafted from generic statements of his values, how he is not part of washington cognicsenti, why the current administration is evil, and how the system needs to balance this against that.
These answers thus force the reader to read a long screed and from that try to judge a) whether the answer was yes or no, and b) whether the you agree with the justification. The whole point of questions is that the voters want concrete datapoints from which to draw conclusions about the candidate's values (this is called "showing" in expositionary writing). These answers are the opposite: instead "telling" us his values and letting us infer the datapoint.
Had each of these answers been prefaced with an unequivocable yes or no, then I would have been much happier.
As for the FUD allegation, yes. I hadn't thought of it in those terms, but I Fear many of the candiates, I have Uncertainty about their vision and values, and I Doubt they will have a positive effect should they win.
nah, it's just that as a rule teachers are not the savvyest bunch of grapes on the vine. They love to teach (not much point being a teacher for the pay these days) so you're not going to get the sort of person who scrutinises a contract for every loophole and repercussion (sp?).
"well meaning" is probably the best generalisation, as it captures the intuitive notion of the strengths and weaknesses of educators as a generalisation.
This is why the companies who prey on this mentality are such slime, IMHO.
The problem with the 3rd parties is that they are crackpots. Sorry. They have great ideas about the small issues like drugs, personal freedoms, overcommercialisation, and things that we like to blather on about.
But they remain crackpots by shooting themselves in the foot by espousing impossible party platforms -- like Nader's 100% taxation above x times minimum wage or Brown's abolish all government idea. I have no idea what the Naturals stand for, but I'm willing to believe you when you claim they're loons.
If these parties would only tone down their crackpot ideas to something that a somewhat mainstream person (ie doesn't stand out on the "T" for whatever reason) could vote for, and maybe get together in some coalition, then maybe they could stand a chance.
For some reason I believe these people are not interested in realpolitik, tho, and would rather wallow in their principles than make a constructive effort to win.
The point has been made on salon (a self admittedly un-republican 'zine) numerous times that Bush's entire campaign has been a bunch of well crafted soundbites that are strung together in a semi-random fashion to answer pretty much any question.
He very rarely answers the question even indirectly, often resorting to wonderful statements like "I have said what I believe in, and if that is what you are asking, then that is what I believe." OR something equally self-referential.
Gore tends to (tho I notice he hasn't bothered to here) say too much, rather than too little, so he tends to drown his answer in a series of snipes and look-at-me's.
But then I get taxation but no representation, so my opinion is moot.
He said nothing of the order, only about the relative frequency (over a largeish number of samples) of the two digits.
I haven't actually verified their frequency, but I'd be all set to believe them pretty equal.
H.L Menken had a great one:
No one ever lost money underestimating the intelligence of the consumer.
There might have been a reference to the nationality of the consumer, but i forget.
"I don't care" in the version I had heard about. And since they asked for it, there's not really a lot the consumer can do about it.
what about you starting in an area that is completely surrounded by mines?
XXXXX
X434X
X303X
X434X
XXXXX
You click on the 0 to start (numbers may be wrong, sorry). the Xs represent mines. To clear the rest of the board, you have to blindly guess on a square outside your cleared area.
That sounds "unsolvable" to me. So you need blind luck.
NB I realise that this is not the what the article is about.
Even if that is the large scale trend, aren't there likely to be eddies (like in a stream) that actually flow backwards, or at least much slower?
That's what I meant about self similarity.
?
That seemed somewhat out of line. I'll assume you have some personal problem that you're just venting here.
No.
You'll notice that they speak about board "consistency". Basically, give me a board with a bunch of numbers on it and all other cells marked X. I'll tell you whether there exists a layout of mines (in the Xs, of course) so that your numbers are correct. This proof basically states that the best we can do is to enumerate all possible layouts and check them in turn.
say there are n Xs on the board. If you can come up with an algorithm that runs in O(n^k) time (k some finite integer) or prove that it is impossible to do faster than O(k^n) then you win a million dollars.
They're not asking you to play the game. mutter mutter. Reading comprehension. mutter mutter.
bravo!
No brownie points to give, sorry!