The problem with tiered web browsing is that unlike electricity, gas and water, the internet and technologies surrounding it have ALWAYS been developed with the ideals of unlimited bandwidth capabilities.
That's nonsense. Any networking technology that's not point-to-point involves many nodes sharing limited bandwidth. One of the goals of packet switching is precisely to allow nodes that use some shared bandwidth intermittently to get full bandwidth during their use.
Suppose you have 10 nodes sharing a 100Mbps network, and each of these nodes only talks about 5% of the time. What way would you prefer the bandwidth to be shared?
Give each node its own dedicated 10Mbps channel. No node can then ever transmit or receive at a rate of more than 10Mbps.
Use packet switching so that each node takes turns using the whole 100Mbps. Effective bandwidth depends on how much those 5%
utilizations overlap, but basically, unless all 10 nodes always try to use the wire at the same time, everybody gets more than 10Mbps average.
From your description, you're talking about compact cameras, which are simply not designed to give the photographer any control over the process. If you want to have manual focus, you have to go with one of the following alternatives:
Advanced compact cameras like the Canon G11, Canon S90, Panasonic LX3, the higher-end Ricoh models, or the Sigma DP1/DP2. Even with these, you have to double check before buying, because controlling focus manually may be very clumsy.
Any interchangeable lens camera. Lenses always have focus rings. The alternatives here are the traditional DSLRs and the new mirrorless cameras like Micro Four Thirds and Samsung NX.
If you don't take some control over the focus system of an SLR, the camera will very often focus on the completely wrong thing. On a handheld shot at medium distance in low light, that can quite easily ruin a shot.
Need to meter the scene and decide where to place the values, no, just trial and error it.
Sure. How many of your guests know how to do that? Are they going to be happy to spend the whole ceremony taking photos at your behest instead of, um, participating?
Photography is all composition today and your mastery of a graphics editor, whoop-dee-do.
So your forced one of your guests to spend the whole ceremony taking photos. How good are they with Lightroom?
How can someone justify a desire to get paid ridiculously when all they did was buy a new camera? Sure "it's an investment" but it's not what I would consider to be a creative effort as their skill hasn't been the influence so much as the equipment.
You say it's only the equipment. Well, how many family members do you have who own a pair of Canon 5D bodies, a 24-70mm f/2.8 and a 70-200mm f/2.8 lens, external flashes and flip flash brackets?
Suppose you do know somebody in your family that has this gear. Do they know how to use them effectively? How often do they shoot ceremonies like yours? How good are they going through the results on Lightroom?
And don't forget: are they willing to spend all of the ceremony taking photos when you supposedly invited them as guests? Will they want to socialize at all? Will they be reluctant to be pulled away from socializing because you want them to take some posed shots of the people after the ceremony?
When you pay for a photographer to cover your ceremony, you're not just paying for equipment. You're paying for their time more than anything, followed by skill (but mostly technical skill in operating their camera), and only then for gear. You might want to argue that they should be charging per hour and not per copies of photos; well, that's not unreasonable. But again, you're arguing that you're paying primarily for their equipment, and I'm telling you that you can rent the same gear, and good luck getting Uncle Bob to get it right for you.
Japanese people aren't especially known for their legginess. Or height, for that matter.
However, a lot of their reputation for being short seems to come from before their post-WWII economic boom. Today's younger Japanese are significantly taller than their ancestors. Wikipedia's data is hard to compare (numbers for different countries sample different sub-populations), but roughly, it seems that between 1850 and 2000 (roughly), young Japanese men went from 5' 1" to 5' 7.5". In the same time period, American men seem to have gone from 5' 7" to 5' 10.5". So, again stressing that the data isn't necessarily the best to compare, Americans used to be 6 inches taller than the Japanese, but now are only 3 inches taller.
Another interesting data point (from the same article): young South Koreans are about 4.7 inches taller on average than comparable North Koreans.
Given that there are plenty of bacteria that can do this (including those that find oxygen toxic) it's not surprising that multicellular creatures have evolved to take advantage of low oxygen environments.
Do you mean that some multicellular aerobic organisms evolved into multicellular anaerobes, or that some monocellular anaerobes evolved into multicellular anaerobes? The article seems to suggest the latter, with the no-mitochondria claim.
So, who actually dominates, intelligent people, or libertarians?
The thing you're missing is that intelligent people are all too often vulnerable to forms of stupidity that average people aren't, and libertarianism is one--it's a One Grand Theory That Explains Everything in a Nutshell(TM), and nerds tend to be suckers for that sort of stuff.
Various sorts of technical challenges involved making a radio contact with somebody very far away. Sometimes natural challenges (distance, propagation), sometimes self-imposed (deliberate use of a low-power transmitter, bouncing radio signals off the moon or meteor trails).
If you're into DIY electronics, ham radio is heaven. You can build, design and/or use your own equipment. Lately, this extends also to software, too--if you're interested in DSP, software radios can be pretty neat; if you're interested in networking technology, likewise packet radio can be fun.
It's occasionally useful when regular communications channels go down.
You do need to make assumptions for the sake of an argument. Assuming that Toyota ownership of these models skew a little older is valid assumption, in case you want to contest this 20% then please say so. Making the assumption into 30% (30% of Toyota owners are older than 65) won't change the calculation very much and the statistically significant skew towards older drivers is still valid.
I certainly don't disagree, but I think the biggest problem in this thread is that we're not pointing how thoroughly wrong about statistics the people who are crying "small sample size" are.
People who complain about the small sample sizes of studies whose conclusions they don't like are operating under the false assumption that the thing that primarily determines the accuracy of an unbiased sample is its relative size, i.e., the proportion of the population that was sampled. But in fact, the accuracy of the sample is primarily determined by the absolute size, and the frequency of the outcome we're trying to estimate.
As a simple thought experiment, imagine a big, opaque black bag with a million marbles, half of them red, the other half blue, randomly scattered. If you put your hand into the bag and draw a marble without looking, you have a 500,000/1,000,000 chance of drawing red. If you drew red on the first one and now draw a second one, you have a 499,999/1,000,000 chance of drawing red; then in your third draw, the chance of drawing red again is 499,998/1,000,000, and so on. So each draw is has a roughly 50/50 chance of getting red, except that drawing red in one turn makes red slightly less likely in the next one.
So, what's your chance of drawing 27 reds on a row? It's easy to see that it's something in the order of 1/2^27, i.e. 1/134,217,728, except that that's actually an overestimate, given that earlier reds make later ones less likely. The formula that calculates the probability of drawing only red marbles from the 50%/50% red/blue bag is looks like an exponential function over the number of marbles you draw, with a relatively minor correction for sample size.
This is a crude version of statistical error estimates, but it applies pretty directly to the Prius driver age situation, because getting all 27 drivers to fall in a narrow age range is very much like drawing 27 red marbles in a row. Even if 66% of Prius drivers were in that age group, the chance of drawing 27 of them in a random draw is something like 1 in 56,800 (1/(2/3)^27).
So, repeat the mantra: Sampling error depends on absolute sample size, not relative sample size.
If Ford was worth $15 a share at the beginning of 2005, and it's worth $14 a share now, was it really only worth $1.40 on November 21, 2008? Or was that the "wisdom of crowds" working for you?
Your point isn't necessarily wrong, but your example isn't great. The government stepped in during late 2008 and early 2010 to support the car industry. Government guarantees, even when implict, surely will boost stock prices, and it's difficult to separate what part of the price owes to opinion and which to government support.
Don't forget, they've shown that a monkey, randomly choosing stocks can beat the market on a regular basis.
No, the monkey gets the same gross return as the market, on average. The monkey is a metaphor for a market index, and the market index can't do better than the market, by definition.
The original point of the monkey methaphot is that professional investment managers' average net returns fall short of the market's gross return, i.e., that active stock investing leads to lower net returns than buying a representative sample of the market and minimizing investing costs. Most investors would do better off with a portfolio of index funds that match the market's gross return at lower-than average costs. As long as indexing is a minority strategy, its lower costs ensure that it will produce above average net returns despite only getting average gross returns.
Cultures/religions are against homosexuality for cultural evolution reasons, more or less the same reason why Catholics are against birth control: it is to the advantage of the culture/religious group to have more children (and therefore out-breed competitors), so they hold beliefs that increase the number of children their followers will have.
Islam doesn't forbid contraception, and it is growing faster than Christianity.
Cultures/religions are against homosexuality for cultural evolution reasons, more or less the same reason why Catholics are against birth control: it is to the advantage of the culture/religious group to have more children (and therefore out-breed competitors), so they hold beliefs that increase the number of children their followers will have.
That's a justification you just pulled out of where the sun don't shine. Don't talk as if it counts as knowledge.
The most famous known case of a society that tolerated or encouraged homosexual behavior is pederasty in ancient Greece. Older men took younger men as lovers, in addition to marrying women. This pattern, BTW, recurs all over the world. The marginalization or acceptance of same-sex relationships is orthogonal to reproduction.
As to the second point, you're presumably not telling us that the Catholic Church adopted their birth-control policies as a calculated move to boost their numbers, but rather, that they have such large numbers because they adopted that policy in the distant past. However, knowledge of effective birth control didn't really exist in Europe until the late 20th century, and Protestant churches used to have similar prohibitions until then. So, your explanation requires us to uncover some sort of pre-20th century church that both allowed its members to practice birth control, and even more, whose members actually knew how to achieve it effectively.
Not to mention that the Catholic Church is historically notorious for having a serious anti-sex attitude and all but telling people not to have sex.
OK, so enlighten us with your brilliance! Share with us the ultimate answer of what should be done to differentiate a null (logically, "I don't know") with a blank string (logically, "We know there's nothing there") and what should be done differently
The way languages like F# or Haskell treat "nulls" would be a straightforward definite improvement over the way SQL does (or for that matter, the common C/C++/Java/C# paradigm). A type whose value is either a Foo or nothing is just a tagged union type. So allowing columns to take tagged unions as their type would solve that right away--and also allow to impose further logical distinctions as needed by the application.
The whole three-valued-logic "null is not a value" paradigm of SQL is a disaster, that one's for sure. There are all sorts of query optimizations that are impossible to do on the face of it.
Why should I give two shits about what database system someone else uses?
Because they're gonna tell your non-technical boss to make you use it, and he's gonna listen when they start telling him that Google, Twitter and Facebook do.
No, it says "in such manner", which means they can go door-to-door, send a letter, use email, ask you to visit your polling place, etc. It doesn't redefine the word "enumerate", which means "count".
The word "enumerate" doesn't mean count. If a test question says "enumerate the causes of the American Revolution," the answer "seven" would be incorrect off the bat, whereas it wouldn't be so if the question was to count the causes.
There's something very important that you're also missing: even if the purpose of the Census was simply to produce an accurate count of people, you cannot produce accurate counts unless you ask for considerably more information than just how many people live in the household. The power of Congress to dictate the manner in which the Census is conducted extends also to adopting methods to make it accurate. Extra questions allow for more statistical checks on the accuracy of the raw survey results, which are inevitably going to have all sorts of errors. The classic one is people getting counted more than once; names and dates of birth allow to reduce that error substantially.
The census data is absolutely useless to medical researchers. "Black" doesn't describe anything about an individuals genetic code other than melanin content. The genetic variation among "black" people is as great or even greater than the genetic variation between any given black person and white people.
Then why do researchers routinely find a link between race and various health conditions?
Race may be a social construct with no genetic basis, but social constructs are real, and often have medical consequences.
"Asian" is used by the census generally to describe anybody from about Pakistan eastward, lumping Indians with Koreans, Japanese, and Chinese, all of which are very distinct from each other.
...and, surprise surprise, this is why question #9 in the 2010 Census form doesn't have a checkbox for "Asian" (contrary to what you imply); it has separate checkboxes for Asian Indian, Chinese, Filipino, Japanese, Korean, Vietnamese, and Other Asian--with a write-in box for people to write in their ethnicity in the latter, complete with a list of examples (Hmong, Laotian, Thai, Pakistani, Cambodian).
So you're very much wrong about the Census lumping "asians" together.
And what constitutes "black" and "white" today, anyway? Is Tiger Woods black, asian, or what? Are his kids black or white? Do you want to bring back the old "one drop" test, so if any of your ancestors are black, you are deemed black?
It has been absolutely clear for decades now that the answers to the census questions about race and ethnicity are about the respondent's self-identification and their identification of the other people in their household, not some objective adjudication of their "real" race. The upsides and downsides of this, its (un)reliability and its problems are very much a matter of public discussion.
Again, race is today understood as a social construct, and race is very much a matter of personal identity. What race people identify themselves as is a significant fact about the population. Nobody who's not constructing a strawman is pretending that this data is to be treated as any more than what it is.
Demographers are among those who continue to insist that we define our society by skin color, so I don't feel much need to help them out.
I'm sorry, but the American society is very much defined by skin color. Native-born and raised black people, whose ancestors have been in this country no shorter than the majority, overwhelmingly live in different neighborhoods, speak a distinctive dialect, have different artistic expressions, have different naming customs for their children, have worse health, are discriminated against in housing, health and employment, and a host of other differences that would not be possible if the nation's culture did not see them as a different race in the first place.
I also put American for race.
Thank you for completing and submitting your Census form. Your answer will be routinely adjudicated as "Non-Hispanic White," perhaps after some minor statistical controls to estimate the very small number of non-whites who fill out "American."
No matter how you try to obfuscate the destination - a base-10 "number", octal, binary, who effing cares how - it still goes out on the wire as an IP packet with a destination address field, either sourced from your desktop or your proxy. Packets don't lie.
Not all IP address filtering is done by IP firewalls. These days there are many applications, most notably web browsers, that consult online databases of known or suspected malicious hosts in order to protect users from malicious hosts. I know for a fact that Firefox and Safari do this--if you try to go to a known suspected malware site, the browser pops up a warning page instead of the page you asked for. Google also do it for their search results--suspected malware site results don't link to the site in question, they link to a warning page. Many websites also have anti-XSS submission filters that perform textual matching against known "bad" addresses, to protect their users from attacks.
Apparently, many such programs are not parsing the textual IP addresses into a canonical form, and are therefore vulnerable to this sort of obfuscation. So the typical result here is that a comment submission system will fail to block a comment that has some XSS in it, and the users' browsers, running on a network whose firewally doesn't filter the IP address in question, will then fetch a malicious script from a known malware site.
Because the computer doesn't really understand IP addresses. To the computer, it sees one 32-bit number. People aren't really good at memorizing numbers, so dotted-octet IP notation was designed. It isn't that we have the ability to enter a number (in any of the many ways any given number can be written) and have it interpreted as an IP. It's that we have the ability to enter an IP and have it converted into a number.
But you see, the point is that there should be three levels of abstraction here, not two:
The low-level computer representation of a 32-bit number. This is where the checks whether two IP addresses are the same should happen.
The canonical textual representation of IP addresses, as used in textual URLs. This is where GP and I think the protocols allow for too much flexibility. Our argument the claim is that as far as the protocol is concerned, each IP address should ideally have just one unique textual representation, and applications should just reject URLs that don't fit the canonical representation. This simplifies applications that have to deal with that protocol.
Tools and libraries that take some input to generate textual representations of URLs. Such tools may take non-canonical representation of IP addresses in hex, base 10, or whatever they want, as long as they can spit out the correct representation in #2 for the benefit of other applications that implement the protocol. I.e., if your application wants to be able to deal with IP addresses represented in texually as hex, octal and decimal numerals, fine, but it shouldn't be allowed to assume that other applications will accept such encodings. It should convert its internal representations into the normal form before it sends stuff over the wire.
Equally important is the question: why do we have to be able specify IP adresses in more than one number format? I don't see any benefit in that.
Indeed. That points to another rule, I think, from the protocol/language design side: the syntax should be as closely isomorphic to the semantics as possible. 10 different ways to say the same thing means 10 different ways things can go wrong.
You can't just do things like this based on the syntax of the input, but rather on the semantics. In this case, to properly block the URLs, you need to parse them and transform them into an abstract representation of what they mean, e.g. a struct that encodes the protocol, host, port, document and query strings, and then examine the parse result to check if it matches the rule.
The IT industry just systematically fails this over and over, because of people's bad habit of doing shit with regular expressions instead of parsing and semantic analysis. See, for example, the gazillion ways that people get around cross-site scripting filters; or if you want to see it from the other angle (generation instead of parsing), see SQL injection.
Slashdot Mirror
That's nonsense. Any networking technology that's not point-to-point involves many nodes sharing limited bandwidth. One of the goals of packet switching is precisely to allow nodes that use some shared bandwidth intermittently to get full bandwidth during their use.
Suppose you have 10 nodes sharing a 100Mbps network, and each of these nodes only talks about 5% of the time. What way would you prefer the bandwidth to be shared?
From your description, you're talking about compact cameras, which are simply not designed to give the photographer any control over the process. If you want to have manual focus, you have to go with one of the following alternatives:
If you don't take some control over the focus system of an SLR, the camera will very often focus on the completely wrong thing. On a handheld shot at medium distance in low light, that can quite easily ruin a shot.
Sure. How many of your guests know how to do that? Are they going to be happy to spend the whole ceremony taking photos at your behest instead of, um, participating?
So your forced one of your guests to spend the whole ceremony taking photos. How good are they with Lightroom?
You say it's only the equipment. Well, how many family members do you have who own a pair of Canon 5D bodies, a 24-70mm f/2.8 and a 70-200mm f/2.8 lens, external flashes and flip flash brackets?
Suppose you do know somebody in your family that has this gear. Do they know how to use them effectively? How often do they shoot ceremonies like yours? How good are they going through the results on Lightroom?
And don't forget: are they willing to spend all of the ceremony taking photos when you supposedly invited them as guests? Will they want to socialize at all? Will they be reluctant to be pulled away from socializing because you want them to take some posed shots of the people after the ceremony?
When you pay for a photographer to cover your ceremony, you're not just paying for equipment. You're paying for their time more than anything, followed by skill (but mostly technical skill in operating their camera), and only then for gear. You might want to argue that they should be charging per hour and not per copies of photos; well, that's not unreasonable. But again, you're arguing that you're paying primarily for their equipment, and I'm telling you that you can rent the same gear, and good luck getting Uncle Bob to get it right for you.
However, a lot of their reputation for being short seems to come from before their post-WWII economic boom. Today's younger Japanese are significantly taller than their ancestors. Wikipedia's data is hard to compare (numbers for different countries sample different sub-populations), but roughly, it seems that between 1850 and 2000 (roughly), young Japanese men went from 5' 1" to 5' 7.5". In the same time period, American men seem to have gone from 5' 7" to 5' 10.5". So, again stressing that the data isn't necessarily the best to compare, Americans used to be 6 inches taller than the Japanese, but now are only 3 inches taller.
Another interesting data point (from the same article): young South Koreans are about 4.7 inches taller on average than comparable North Koreans.
Do you mean that some multicellular aerobic organisms evolved into multicellular anaerobes, or that some monocellular anaerobes evolved into multicellular anaerobes? The article seems to suggest the latter, with the no-mitochondria claim.
I've been out of this for a number of years now, but my understanding is that PSK31 gives Morse a run for its money, and is therefore popular for QRP.
The thing you're missing is that intelligent people are all too often vulnerable to forms of stupidity that average people aren't, and libertarianism is one--it's a One Grand Theory That Explains Everything in a Nutshell(TM), and nerds tend to be suckers for that sort of stuff.
Beware the key term there: "up to."
I certainly don't disagree, but I think the biggest problem in this thread is that we're not pointing how thoroughly wrong about statistics the people who are crying "small sample size" are.
People who complain about the small sample sizes of studies whose conclusions they don't like are operating under the false assumption that the thing that primarily determines the accuracy of an unbiased sample is its relative size, i.e., the proportion of the population that was sampled. But in fact, the accuracy of the sample is primarily determined by the absolute size, and the frequency of the outcome we're trying to estimate.
As a simple thought experiment, imagine a big, opaque black bag with a million marbles, half of them red, the other half blue, randomly scattered. If you put your hand into the bag and draw a marble without looking, you have a 500,000/1,000,000 chance of drawing red. If you drew red on the first one and now draw a second one, you have a 499,999/1,000,000 chance of drawing red; then in your third draw, the chance of drawing red again is 499,998/1,000,000, and so on. So each draw is has a roughly 50/50 chance of getting red, except that drawing red in one turn makes red slightly less likely in the next one.
So, what's your chance of drawing 27 reds on a row? It's easy to see that it's something in the order of 1/2^27, i.e. 1/134,217,728, except that that's actually an overestimate, given that earlier reds make later ones less likely. The formula that calculates the probability of drawing only red marbles from the 50%/50% red/blue bag is looks like an exponential function over the number of marbles you draw, with a relatively minor correction for sample size.
This is a crude version of statistical error estimates, but it applies pretty directly to the Prius driver age situation, because getting all 27 drivers to fall in a narrow age range is very much like drawing 27 red marbles in a row. Even if 66% of Prius drivers were in that age group, the chance of drawing 27 of them in a random draw is something like 1 in 56,800 (1/(2/3)^27).
So, repeat the mantra: Sampling error depends on absolute sample size, not relative sample size.
Your point isn't necessarily wrong, but your example isn't great. The government stepped in during late 2008 and early 2010 to support the car industry. Government guarantees, even when implict, surely will boost stock prices, and it's difficult to separate what part of the price owes to opinion and which to government support.
Islam doesn't forbid contraception, and it is growing faster than Christianity.
That's a justification you just pulled out of where the sun don't shine. Don't talk as if it counts as knowledge.
The most famous known case of a society that tolerated or encouraged homosexual behavior is pederasty in ancient Greece. Older men took younger men as lovers, in addition to marrying women. This pattern, BTW, recurs all over the world. The marginalization or acceptance of same-sex relationships is orthogonal to reproduction.
As to the second point, you're presumably not telling us that the Catholic Church adopted their birth-control policies as a calculated move to boost their numbers, but rather, that they have such large numbers because they adopted that policy in the distant past. However, knowledge of effective birth control didn't really exist in Europe until the late 20th century, and Protestant churches used to have similar prohibitions until then. So, your explanation requires us to uncover some sort of pre-20th century church that both allowed its members to practice birth control, and even more, whose members actually knew how to achieve it effectively.
Not to mention that the Catholic Church is historically notorious for having a serious anti-sex attitude and all but telling people not to have sex.
Can you explain your reasoning here? I really don't see how it would.
The way languages like F# or Haskell treat "nulls" would be a straightforward definite improvement over the way SQL does (or for that matter, the common C/C++/Java/C# paradigm). A type whose value is either a Foo or nothing is just a tagged union type. So allowing columns to take tagged unions as their type would solve that right away--and also allow to impose further logical distinctions as needed by the application.
The whole three-valued-logic "null is not a value" paradigm of SQL is a disaster, that one's for sure. There are all sorts of query optimizations that are impossible to do on the face of it.
Because they're gonna tell your non-technical boss to make you use it, and he's gonna listen when they start telling him that Google, Twitter and Facebook do.
You were one of the unlucky 10% to receive the 2000 Census Long For Questionnaire (PDF link) ; most people got the Short Form Questionnaire (PDF link). The 2010 Census doesn't have a long form.
The word "enumerate" doesn't mean count. If a test question says "enumerate the causes of the American Revolution," the answer "seven" would be incorrect off the bat, whereas it wouldn't be so if the question was to count the causes.
There's something very important that you're also missing: even if the purpose of the Census was simply to produce an accurate count of people, you cannot produce accurate counts unless you ask for considerably more information than just how many people live in the household. The power of Congress to dictate the manner in which the Census is conducted extends also to adopting methods to make it accurate. Extra questions allow for more statistical checks on the accuracy of the raw survey results, which are inevitably going to have all sorts of errors. The classic one is people getting counted more than once; names and dates of birth allow to reduce that error substantially.
Then why do researchers routinely find a link between race and various health conditions?
Race may be a social construct with no genetic basis, but social constructs are real, and often have medical consequences.
...and, surprise surprise, this is why question #9 in the 2010 Census form doesn't have a checkbox for "Asian" (contrary to what you imply); it has separate checkboxes for Asian Indian, Chinese, Filipino, Japanese, Korean, Vietnamese, and Other Asian--with a write-in box for people to write in their ethnicity in the latter, complete with a list of examples (Hmong, Laotian, Thai, Pakistani, Cambodian).
So you're very much wrong about the Census lumping "asians" together.
It has been absolutely clear for decades now that the answers to the census questions about race and ethnicity are about the respondent's self-identification and their identification of the other people in their household, not some objective adjudication of their "real" race. The upsides and downsides of this, its (un)reliability and its problems are very much a matter of public discussion.
Again, race is today understood as a social construct, and race is very much a matter of personal identity. What race people identify themselves as is a significant fact about the population. Nobody who's not constructing a strawman is pretending that this data is to be treated as any more than what it is.
I'm sorry, but the American society is very much defined by skin color. Native-born and raised black people, whose ancestors have been in this country no shorter than the majority, overwhelmingly live in different neighborhoods, speak a distinctive dialect, have different artistic expressions, have different naming customs for their children, have worse health, are discriminated against in housing, health and employment, and a host of other differences that would not be possible if the nation's culture did not see them as a different race in the first place.
Thank you for completing and submitting your Census form. Your answer will be routinely adjudicated as "Non-Hispanic White," perhaps after some minor statistical controls to estimate the very small number of non-whites who fill out "American."
Pepper spray is made by extracting the capsaicin from hot peppers. Hotter peppers means more efficient manufacturing process.
Not all IP address filtering is done by IP firewalls. These days there are many applications, most notably web browsers, that consult online databases of known or suspected malicious hosts in order to protect users from malicious hosts. I know for a fact that Firefox and Safari do this--if you try to go to a known suspected malware site, the browser pops up a warning page instead of the page you asked for. Google also do it for their search results--suspected malware site results don't link to the site in question, they link to a warning page. Many websites also have anti-XSS submission filters that perform textual matching against known "bad" addresses, to protect their users from attacks.
Apparently, many such programs are not parsing the textual IP addresses into a canonical form, and are therefore vulnerable to this sort of obfuscation. So the typical result here is that a comment submission system will fail to block a comment that has some XSS in it, and the users' browsers, running on a network whose firewally doesn't filter the IP address in question, will then fetch a malicious script from a known malware site.
But you see, the point is that there should be three levels of abstraction here, not two:
Indeed. That points to another rule, I think, from the protocol/language design side: the syntax should be as closely isomorphic to the semantics as possible. 10 different ways to say the same thing means 10 different ways things can go wrong.
You can't just do things like this based on the syntax of the input, but rather on the semantics. In this case, to properly block the URLs, you need to parse them and transform them into an abstract representation of what they mean, e.g. a struct that encodes the protocol, host, port, document and query strings, and then examine the parse result to check if it matches the rule.
The IT industry just systematically fails this over and over, because of people's bad habit of doing shit with regular expressions instead of parsing and semantic analysis. See, for example, the gazillion ways that people get around cross-site scripting filters; or if you want to see it from the other angle (generation instead of parsing), see SQL injection.