I'm not sure that I agree. One of the reasons you think this is a good idea is so that the Internet can be governed by one set of consistent rules and regulations, including this concept of decency. The problem with decency is that it's highly subjective, and highly dependent upon the local community norms. While we in the US have a fairly consistent continuum of decency, even here, the line between "decent" and "indecent" floats from region to region. Radio stations in some areas bleep out words that radio stations in other areas do not. You can't set a national standard without it being inappropriate for some regions, unless your goal is to force those other regions to accept your definition of "decency".
But that's just talking about the US. World-wide, social norms vary in multiple dimensions. Things like nudity in public, language, age of consent, pornography, viewing the faces of women are tolerated in completely different ways across different nations. You cannot hope to apply a common set of rules governing decency without seriously pissing one or more groups off, because decency is strongly defined by local norms and customs. It is not an intrinsic property of all people with one set of rules that's "best" for everyone, as much as some people would like to believe.
The only problem I have with ICANN is that it's too political and its members too selfish. Open everything up, do the right thing that balances technical and non-technical needs, be transparent, document, and absolutely refuse to cater to your benefactors. I personally don't think that ICANN can be effective in its current form.
Being pulled to the middle is not something intrinsic in a two-party system. A three-party system would have exactly the same number of voters to sway, and would be pulled in the same direction to capture the same portion of votes.
Your other concerns could be resolved by abandoning our single-vote, single-candidate system of voting in favor of a more reasonable voting system.
A good rotating leadership should address the primary needs of all of its major constituent groups. Each time a party gets a majority in congress, or a president, that person's term sees policies and laws shift in the direction of that party's vision, at the exclusion of the vision of a large portion of the population. People will eventually think, "This party has gone too far," and will vote in another direction, bring another party to power, where they will see policies and laws shift in another direction.
This doesn't have to be a two-dimensional tug of war, though. There's no reason a three- or four-party system couldn't have the same results.
The trick is *moving* to such a system. Once a two-party system is entrenched, you can't vote for a third party unless you're positive that everyone else wanting that third party will be voting for it. Otherwise, it's a wasted vote, and your second choice has lost a much-needed vote, giving your last choice the win. (Alternative voting systems would deal with this situation as well.)
I do not appreciate the condescending tone. Just because someone disagrees with you does not mean they are not literate or not paying attention to the discussion. Reasonable people can disagree reasonably.
Please show me someone saying the words "XML is hard," or any actual evidence in that direction.
My comment was not intended to be a literal quotation.
From the article:
You see, for a while now, people have tried to extract structured data from the unstructured Web. You hear glimmers of these when people talk about the "semantic Web," a Web in which data is separated from formatting. But for whatever reason, the semantic Web hasn't taken off, and the problem of finding structured data in an unstructured world remains.
He's referring to the "semantic web" generally, which most followers of the semantic web interpret to mean XML and the family of "semantic" markup languages built with XML.
The quoted article says this approach hasn't caught on. I rephrased that, with some creative license, as saying, "XML is hard." I don't think that's an inaccurate characterization. But it's also largely irrelevant to the point I was trying to make.
Namespaces in XML were created solely for the purpose of preventing name conflicts.
I completely, 100% agree. Perhaps I misspoke when I referred to namespaces by themselves. Namespaces themselves do not contain any semantics. The various XML languages do, however. XHTML has its semantics, and My Markup Language contains a completely different set of semantics. How do you embed My Markup Language semantics within XHTML? Namespaces. You don't overload HTML elements and assign new semantics to HTML tags that already have semantics. You don't set yourself up for collisions when two "formats" want to overload the same pattern of elements and attributes. That is what namespaces are here to solve.
The reason you think this problem isn't well solved by the solution is that this just isn't what the tool is meant to solve.
This doesn't change the fact that the problem exists. Both namespaced XML and "microformats" allow the embedding of additional (arbitrary) semantics within another type of document. If microformats inherently don't elect to deal with intermingled data, isn't that just another way of saying it has a deficiency?
IE already supports the part of XHTML that deals with these concerns. It has since IE4. Perhaps you should try it.
IE does not support XHTML unless it's transformed from XML. IE supports HTML tag soup, and it supports raw XML. You can either tell IE that your XHTML is HTML, in which case IE will interpret it as HTML tag soup, or you can tell IE that it's XML, in which case IE will treat it as raw, unformatted XML (no HTML semantics). If you have a piece of XHTML content that you desire to deliver as XML (application/xhtml+xml or some other XML media type), your content will not be interpreted as XHTML in IE, because IE does not support XHTML. You can either deliver it as text/html, in which case XML-aware applications strictly honoring media types will not be aware of your additional XML data, as application/xhtml+xml, in which case your XML-aware applications can extract information from it, but the page is unreadable in IE, or some XSLT modified version of your content that allows IE to transform it to XHTML. This was the problem I was attempting to discuss. With proper XHTML support, documents could be created in true, validating, standards-compliant XML/XHTML, and not only would XML-aware applications be able to extract useful information from it, but it would render properly in popular browsers.
Event? What do events have to do with anything?
I was continuing with the example given in the article.
It's remarkably rude and conceited.
And how do you think your post looked? Please drop the attitude.
but Microformats are about embedding formats, not creating them
It seems to me that creating them is exactly what this is about. Taking a step back, what they're saying is, "XML is hard. But if you make up a pattern of HTML elements and reserve some class names, programs can parse out information in standard ways."
This is the same problem that XML namespaces were intended to solve. OK, so this works for a handful of "formats". Clever (and planned) use of CSS gets the data displayed and compatible user agents can more readily parse information out of it. But this solution doesn't scale! Eventually you're going to get "formats" that start to step on each other's toes. They use the same class names, or the same pattern of elements. Maybe you want an "event" but want to add some supplementary information about that event using another "format". Do they mingle together?
If IE can get off of its ass and properly support XHTML, this problem is already solved. Create your event in XHTML, and supplement it with XML tags or attributes from other XML namespaces to include the machine-readable information. If you're concerned about how to style this XML data, remember that CSS can style (or, by extension, hide) XML just fine.
If the parent document is XHTML, and the browser understands that, CSS can easily be used to style these additional non-XHTML elements any way you like.
I don't know, but the easy way to achieve this would be to resolve "mistyped" addresses to a valid SiteFinder-style IP address that would then accept HTTP connections and deliver the search/error page. This is really geared towards the Intarweb user.
I think your concept of first/last name is a little too Americentric. The point I was trying to make was that not everyone can be uniquely identified with a traditional American-style first and last name. If someone has a very distinctive combination, a reaction to their appearance on a watch list might seem justified. But if someone has a very common combination of first and last name, that starts to become a very significant burden on your society (thousands of false matches looking for one). You're saying the latter is justified, simply because their first and last names match. I'm saying that a person's name does not matter. The policies are targeting a racial and/or religious minority and potentially affecting a large portion of that minority. Whenever you have policies like this that are likely to affect a large number of people innocent of any crime (and some would argue a single person is too much), you really need to reconsider your approach. Denying someone the right to conduct business simply because their first and last names match the first and last name of someone on a watch list is unfair to that person. Denying a large segment of a racial/religious minority as a consequence of that policy is, IMO, crossing an unacceptable line. The person's name doesn't matter. The effect is what matters.
I see no reason why some extra checking is not warranted
But this goes beyond extra checking, doesn't it? This was an outright denial of service, so to speak.
What about the Muslim equivalent to "John Smith"? Many deeply religious societies reuse the same names, to the point where people will change their names to honor someone else.
If this were just a little "extra checking", because someone was trying to send money to someone whose name was on a terrorist watch list, that's one thing. That's the whole point of the watch list, after all (and I'm not going to debate the merits of that). You see something interesting, you investigate. But when a private company starts preventing people from doing things just because it happens to involve a name similar to someone on a terrorist list, I have more of a problem with that. Companies wouldn't unilaterally do this by themselves. The government is "suggesting" it, and at that point, you are depriving someone of their rights without due process. Our justice system isn't intended to work that way.
... using both together from a list of known terrorists to do a little more checking doesn't seem bad to me. In the article it mentioned there are other ways to send money so if it is too much of a hassle then don't use the service.
So you admit that since most of the people committing these crimes are of a certain racial or regilious heritage, that it's OK to subject them (and those with names that suggest they belong to those heritages) to additional inconvenience and scrutiny, and to prevent them from using certain services that other people enjoy without a problem, because... there are alternative services they can use? "Separate but equal." I like it!
I wonder if there is a racial or religious correlation to crime rates in the US. We could use that logic there too! Why hasn't anyone thought of that yet? Brilliant!
I don't think they come with cameras, or a public interface that I can use to collect my own data. That's really what I'm after. I'm aware of the devices that you're talking about and they've been discussed on Slashdot in a few different stories.
Every once in a while I look up the current state of the art and try to figure out how to cost-effectively put a "black box" into my car, not just grabbing video, but hooking into the car's computer to grab things like speed and other details. I'm not sure how useful it would be in court, but the video at least would make some entertaining watching. Submitting a few clips of some of the asshole law-breaking drivers I come across every day to the local law enforcement might feel good too.
I think that approach has some seriously flawed logic.
1. "Targeting" a single auction out of 300 with the intent of winning it (rather than getting the best deal) increases the likelyhood that you will end up bidding more than the item is truly worth. 2. The approach assumes that you would end up with a better deal sniping the auction than if you bid your honest maximum early in the auction's life. 3. It assumes that there will be other snipers working against you (since other snipers would be the only ones hiding their true maximum bid; non-snipers would have already bid their maximum)
The only way you win auctions is if your maximum bid exceeds the maximum bids of the non-snipers, and/or (if you're a sniper) your "sniping bid" can be made before other snipers have a chance to "snipe" their way above your own maximum bid.
A potential sniper should ask himself how he is able to snipe at all. The top bidders in the last couple of minutes of an auction are all snipers. Since these bids all exceed the maximum bids of the non-snipers, this suggests (tiny incremental bids notwithstanding) that non-snipers are either underbidding, or snipers are all overbidding.
This has more to do with the fuzzy nature of peoples' maximum bids. The solution to that is to have larger bid increments, like real auctions do. It's kind of stupid to allow people to bid pennies over someone else's bid, because it just feeds this type of behavior. If the average person knows that their maximum is ~$500, and that they wouldn't go as high as $525, maybe a $10 or a $25 bid increment is appropriate.
I tend to do the fire-and-forget thing. I set my maximum, and if I'm out-bid, then oh well. It's more amusement to me to watch people jump all over the bid in the last 30-60 seconds of the auction. Their time must not be worth very much.
In an auction without snipers, the person with the highest maximum bid wins the auction. The other bidders weren't willing to pay what the winning bidder paid.
In an auction with non-snipers and a single sniper, where the sniper wins the auction, the outcome is the same. The non-sniper bidders lost the auction, but they weren't willing to pay what the sniper paid anyway, so nothing was lost. The sniper would have achieved his goal by placing his "sniping" bid earlier in the process.
In an auction with multiple snipers, the non-snipers still don't care, because the snipers are willing to pay more than they are. The snipers at this point are "battling" with each other, not the non-snipers. If they can hold off, deprive the other snipers of information, and try to "strike" at the last possible moment, they might get a better deal, because none of the snipers are bidding what they are willing to spend, they're bidding the amount that they hope will win them the auction. If we were to clairvoyantly assess each person's true maximum bid, the true maximum would probably be more than what the item sold for, at least sometimes. In other words, if the snipers had simply placed their true maximum bid earlier in the process, the item would have likely sold for more.
With an online auction composed primarily of non-snipers, the non-snipers are going to set a good maximum bid. Those that choose to snipe on top of that are probably spending more money than the item is worth. With an online auction composed of more snipers, these tactics probably "work" in that the sniper can probably get a better deal, because nobody's bidding their real maximum. Pretending your maximum is perhaps lower than it really is could work in your favor, but it's unlikely to affect the decisions of the non-snipers.
Snipers are effective only among themselves. With enough non-snipers bidding in auctions, the possible profit that a sniper can make is reduced. IMO, snipers aren't necessarily in it as a cool, calculated way of maximizing return over a large number of auctions. They're in it because it's a game to them, and it's no fun just to have your maximum bid exceed someone else's maximum bid. But if it turns into a last-second battle of the bids, then you feel like you've "won" something, and eBay bidding becomes exciting. This is stupid, obviously, because these people will usually be overbidding. I suspect this behavior is usually just a form of gambling or shopping addiction and many of these people need to actually seek treatment.
That's not to say that there aren't "legitimate" snipers who do do it for purely economic reasons, but it still doesn't really add up, IMO. I've had lots of auctions sniped from me, and I think the whole thing is pretty amusing. I've seen auctions that I've bid on end up with a winning bid exceeding the price for the same item from Best Buy, all done in the last 30 seconds. Tell me how that makes any economic sense.
I still don't understand why people think they've "stolen" or "won" something cleverly by using "sniping" tactics. They've simply placed a bid that was more than the highest bid I was prepared to make. I've never done it and I've never understood why others do it. If you're really willing to spend more than I am, just set your maximum bid there and let it work out on its own. People don't seem to "get" the whole proxy bidding concept, IMO.
Though small bid increments certainly make it a little more difficult to find your maximum bid. You can say "$50, tops", but then someone wins it with a maximum bid of $50.01, and you feel stupid. That shouldn't happen, but the fix to that is using larger bid increments. You shouldn't have to strategize in a proxy bidding process.
But if you put a sign above your door that says "Open to all visitors!", then it's not. That is effectively what an open WiFi access point is doing, and in most access point configurations, the implications of that setting are clear.
Actually, after looking at the act more closely, it would seem (to my layperson's eyes) to put enforcement squarely within the banking system itself. So if the federal banking system can't get your issue resolved, perhaps you would have to resort to your own suit...
You wouldn't need "10 years of lawyers". This isn't a civil tort. If they're doing something illegal (as in, against the law), you would not prosecute them. Your government (local, state or federal) would. Talk to the branch manager, and if the story really is as has been told, and they aren't willing to release the hold, refer them to your state's attorney general. You can even avoid paying for a stamp if they accept e-mails.
Would you trust even very intelligent humans to be able to follow the three laws as written even if they actually desired to do so?
I think you're assuming that any "proper" artificial intelligence must necessarily resemble human intelligence and be sentient and capable of free will. Long before we reach that point, we will have exceptionally advanced intelligence that is still very much constrained by its programming.
Even if AI must necessarily have free will, that doesn't mean they can't be given laws in the form of instincts. Humans have a tremendous drive for self-preservation, for example. When you're extremely thirsty you have a nearly uncontrollable need to seek water. Your body is hard wired by evolution to do these things, even in the absence of higher brain function. Who's to say machine intelligences can't be hard wired in a similar manner?
They require nearly omniscient comprehension of the effects of ones actions -- how can you know that you have to refuse to drive to the mall and pick up three cans of tomato sauce because if you don't you'll be in a car wreck with a little old lady and break rule 1?
Not really. They just need to be able to have a basic understanding of the situation and the ability to predict the consequences of inaction and of intervention. That requires a lot of information and a lot of processing, but it doesn't really require sentience or a human mind. Understanding consequences is mostly a pattern matching exercise.
You imply that developing a website using web standards takes longer.
No, he's implying that it's more expensive.
High school kids that can vomit out "IEML" web sites that meet business requirements are a dime a dozen. People producing higher quality, standards-compliant web sites that meet and exceed business requirements are significantly more expensive, even considering the likelyhood that they'll be able to do it faster.
A = The Linux Kernel B = The binary display drivers C = The shim that connects A to B
The issue is with the aggregation of the display driver and the shim with the kernel. The shim is the only binary component that needs to be licensed under the GPL, and it is. Source is provided. The binary driver itself (B) is not licensed under the GPL and can be freely aggregated and distributed with the rest of the kernel.
The only point where this gets dicey is when you statically link all three together. Linked together, they form an entirely new thing derived from all three, and could not be legally copied and redistributed unless the whole thing were licensed under the GPL. If you don't statically link everything together, though, there remains no copyright issue, and thus no GPL issue. You can interpret the GPL however you want, but it's copyright law that decides whether or not the GPL even applies, and it's copyright law that establishes what it means to be a derivative work, not the GPL.
That's just it, though, the license doesn't matter squat here. It's copyright law that decides what is a derivative work and what isn't, not the GPL (which is simply a license, after all).
If the core display drivers were written completely independently of the Linux kernel or any other GPL code, then they are completely independent. You can never say that one was derived from the other, or vice-versa.
If you write a "shim" that connects the two, and that shim uses code (source or binary) from both the kernel and the display driver, you could say that the shim is a derivative of both. That will never make the display driver itself a derivative of the kernel. The GPL doesn't rewrite history.
Things only get tricky when you statically link things together. If the kernel is statically linked with the shim, which in turn is statically linked with the display driver itself, you end up with one (new!) functional binary computer program that incorporates both GPLed and proprietary code. Distribution of that is another matter entirely, because that would appear to be a derivative work (a combination--not mere aggregation--of all three). But that's trivial to solve: don't statically link!
None of this would be a violation as long as the script is not run behind the users' backs, and as long as the user is explicitly informed they're installing a proprietary driver, told what the consequences are, and asked to click an "AGREE" button or something similar.
Where do these requirements come from? I can't find them in the GPL. How does giving explicit notice to the user (who isn't doing anything relevant to copyright law) affect what is inherently a copyright issue?
If A and B are developed completely independently, you cannot, in any context whatsoever, legal or otherwise, ever claim that B was derived from A.
If you create C, which incorporates elements from A and B, you can say that C is derived from both A and B. Because it is. If A is licensed under the GPL, you cannot redistribute C unless it, too, is licensed under the GPL. By licensing C under the GPL, the elements from B that made it into C are thus licensed under the GPL. B as a whole is not.
Derivation is not communicative. You cannot say that since C was derived from B, that B is also derived from C. GPL may be viral, but it hasn't mutated to the point where it can infect backwards in time.
The issue of derivation is only controversial when you bring the discussion up in a room of thousands of legal noobs. The issue is actually fairly straightforward from a legal perspective, and I believe it's already been tested.
but they should open their site with a uniform warning page allowing the site to be filtered thereafter, or other such methods.
The PICS standard has been around for years and is designed to do precisely this in a machine-readable way. Put up a PICS label on your web page (or in the HTTP headers), configure your browser to prevent access to PICS-labeled sites that say they contain porn, maybe configure it also to prevent access to sites without PICS ratings, lock the browser configuration with a password, and viola! A kid-safe browsing experience. IE supports all of this today. Mozilla opted not to support it because nobody uses it. Yet in all respects it is the perfect answer to this problem. I don't really understand why no one is jumping on this.
You still have to deal with sex sites that deliberately mis-label, but that really starts to cross a line between a shady porn site and one deliberately trying to target minors by labeling themselves as kid-friendly. Legal remedies for that already exist.
Slashdot Mirror
I'm not sure that I agree. One of the reasons you think this is a good idea is so that the Internet can be governed by one set of consistent rules and regulations, including this concept of decency. The problem with decency is that it's highly subjective, and highly dependent upon the local community norms. While we in the US have a fairly consistent continuum of decency, even here, the line between "decent" and "indecent" floats from region to region. Radio stations in some areas bleep out words that radio stations in other areas do not. You can't set a national standard without it being inappropriate for some regions, unless your goal is to force those other regions to accept your definition of "decency".
But that's just talking about the US. World-wide, social norms vary in multiple dimensions. Things like nudity in public, language, age of consent, pornography, viewing the faces of women are tolerated in completely different ways across different nations. You cannot hope to apply a common set of rules governing decency without seriously pissing one or more groups off, because decency is strongly defined by local norms and customs. It is not an intrinsic property of all people with one set of rules that's "best" for everyone, as much as some people would like to believe.
The only problem I have with ICANN is that it's too political and its members too selfish. Open everything up, do the right thing that balances technical and non-technical needs, be transparent, document, and absolutely refuse to cater to your benefactors. I personally don't think that ICANN can be effective in its current form.
Being pulled to the middle is not something intrinsic in a two-party system. A three-party system would have exactly the same number of voters to sway, and would be pulled in the same direction to capture the same portion of votes.
Your other concerns could be resolved by abandoning our single-vote, single-candidate system of voting in favor of a more reasonable voting system.
A good rotating leadership should address the primary needs of all of its major constituent groups. Each time a party gets a majority in congress, or a president, that person's term sees policies and laws shift in the direction of that party's vision, at the exclusion of the vision of a large portion of the population. People will eventually think, "This party has gone too far," and will vote in another direction, bring another party to power, where they will see policies and laws shift in another direction.
This doesn't have to be a two-dimensional tug of war, though. There's no reason a three- or four-party system couldn't have the same results.
The trick is *moving* to such a system. Once a two-party system is entrenched, you can't vote for a third party unless you're positive that everyone else wanting that third party will be voting for it. Otherwise, it's a wasted vote, and your second choice has lost a much-needed vote, giving your last choice the win. (Alternative voting systems would deal with this situation as well.)
Please show me someone saying the words "XML is hard," or any actual evidence in that direction.
My comment was not intended to be a literal quotation.
From the article:
He's referring to the "semantic web" generally, which most followers of the semantic web interpret to mean XML and the family of "semantic" markup languages built with XML.
The quoted article says this approach hasn't caught on. I rephrased that, with some creative license, as saying, "XML is hard." I don't think that's an inaccurate characterization. But it's also largely irrelevant to the point I was trying to make.
Namespaces in XML were created solely for the purpose of preventing name conflicts.
I completely, 100% agree. Perhaps I misspoke when I referred to namespaces by themselves. Namespaces themselves do not contain any semantics. The various XML languages do, however. XHTML has its semantics, and My Markup Language contains a completely different set of semantics. How do you embed My Markup Language semantics within XHTML? Namespaces. You don't overload HTML elements and assign new semantics to HTML tags that already have semantics. You don't set yourself up for collisions when two "formats" want to overload the same pattern of elements and attributes. That is what namespaces are here to solve.
The reason you think this problem isn't well solved by the solution is that this just isn't what the tool is meant to solve.
This doesn't change the fact that the problem exists. Both namespaced XML and "microformats" allow the embedding of additional (arbitrary) semantics within another type of document. If microformats inherently don't elect to deal with intermingled data, isn't that just another way of saying it has a deficiency?
IE already supports the part of XHTML that deals with these concerns. It has since IE4. Perhaps you should try it.
IE does not support XHTML unless it's transformed from XML. IE supports HTML tag soup, and it supports raw XML. You can either tell IE that your XHTML is HTML, in which case IE will interpret it as HTML tag soup, or you can tell IE that it's XML, in which case IE will treat it as raw, unformatted XML (no HTML semantics). If you have a piece of XHTML content that you desire to deliver as XML (application/xhtml+xml or some other XML media type), your content will not be interpreted as XHTML in IE, because IE does not support XHTML. You can either deliver it as text/html, in which case XML-aware applications strictly honoring media types will not be aware of your additional XML data, as application/xhtml+xml, in which case your XML-aware applications can extract information from it, but the page is unreadable in IE, or some XSLT modified version of your content that allows IE to transform it to XHTML. This was the problem I was attempting to discuss. With proper XHTML support, documents could be created in true, validating, standards-compliant XML/XHTML, and not only would XML-aware applications be able to extract useful information from it, but it would render properly in popular browsers.
Event? What do events have to do with anything?
I was continuing with the example given in the article.
It's remarkably rude and conceited.
And how do you think your post looked? Please drop the attitude.
but Microformats are about embedding formats, not creating them
It seems to me that creating them is exactly what this is about. Taking a step back, what they're saying is, "XML is hard. But if you make up a pattern of HTML elements and reserve some class names, programs can parse out information in standard ways."
This is the same problem that XML namespaces were intended to solve. OK, so this works for a handful of "formats". Clever (and planned) use of CSS gets the data displayed and compatible user agents can more readily parse information out of it. But this solution doesn't scale! Eventually you're going to get "formats" that start to step on each other's toes. They use the same class names, or the same pattern of elements. Maybe you want an "event" but want to add some supplementary information about that event using another "format". Do they mingle together?
If IE can get off of its ass and properly support XHTML, this problem is already solved. Create your event in XHTML, and supplement it with XML tags or attributes from other XML namespaces to include the machine-readable information. If you're concerned about how to style this XML data, remember that CSS can style (or, by extension, hide) XML just fine.
If the parent document is XHTML, and the browser understands that, CSS can easily be used to style these additional non-XHTML elements any way you like.
I don't know, but the easy way to achieve this would be to resolve "mistyped" addresses to a valid SiteFinder-style IP address that would then accept HTTP connections and deliver the search/error page. This is really geared towards the Intarweb user.
I think your concept of first/last name is a little too Americentric. The point I was trying to make was that not everyone can be uniquely identified with a traditional American-style first and last name. If someone has a very distinctive combination, a reaction to their appearance on a watch list might seem justified. But if someone has a very common combination of first and last name, that starts to become a very significant burden on your society (thousands of false matches looking for one). You're saying the latter is justified, simply because their first and last names match. I'm saying that a person's name does not matter. The policies are targeting a racial and/or religious minority and potentially affecting a large portion of that minority. Whenever you have policies like this that are likely to affect a large number of people innocent of any crime (and some would argue a single person is too much), you really need to reconsider your approach. Denying someone the right to conduct business simply because their first and last names match the first and last name of someone on a watch list is unfair to that person. Denying a large segment of a racial/religious minority as a consequence of that policy is, IMO, crossing an unacceptable line. The person's name doesn't matter. The effect is what matters.
I see no reason why some extra checking is not warranted
But this goes beyond extra checking, doesn't it? This was an outright denial of service, so to speak.
What about the Muslim equivalent to "John Smith"? Many deeply religious societies reuse the same names, to the point where people will change their names to honor someone else.
If this were just a little "extra checking", because someone was trying to send money to someone whose name was on a terrorist watch list, that's one thing. That's the whole point of the watch list, after all (and I'm not going to debate the merits of that). You see something interesting, you investigate. But when a private company starts preventing people from doing things just because it happens to involve a name similar to someone on a terrorist list, I have more of a problem with that. Companies wouldn't unilaterally do this by themselves. The government is "suggesting" it, and at that point, you are depriving someone of their rights without due process. Our justice system isn't intended to work that way.
... using both together from a list of known terrorists to do a little more checking doesn't seem bad to me. In the article it mentioned there are other ways to send money so if it is too much of a hassle then don't use the service.
So you admit that since most of the people committing these crimes are of a certain racial or regilious heritage, that it's OK to subject them (and those with names that suggest they belong to those heritages) to additional inconvenience and scrutiny, and to prevent them from using certain services that other people enjoy without a problem, because... there are alternative services they can use? "Separate but equal." I like it!
I wonder if there is a racial or religious correlation to crime rates in the US. We could use that logic there too! Why hasn't anyone thought of that yet? Brilliant!
I don't think they come with cameras, or a public interface that I can use to collect my own data. That's really what I'm after. I'm aware of the devices that you're talking about and they've been discussed on Slashdot in a few different stories.
Every once in a while I look up the current state of the art and try to figure out how to cost-effectively put a "black box" into my car, not just grabbing video, but hooking into the car's computer to grab things like speed and other details. I'm not sure how useful it would be in court, but the video at least would make some entertaining watching. Submitting a few clips of some of the asshole law-breaking drivers I come across every day to the local law enforcement might feel good too.
I think that approach has some seriously flawed logic.
1. "Targeting" a single auction out of 300 with the intent of winning it (rather than getting the best deal) increases the likelyhood that you will end up bidding more than the item is truly worth.
2. The approach assumes that you would end up with a better deal sniping the auction than if you bid your honest maximum early in the auction's life.
3. It assumes that there will be other snipers working against you (since other snipers would be the only ones hiding their true maximum bid; non-snipers would have already bid their maximum)
The only way you win auctions is if your maximum bid exceeds the maximum bids of the non-snipers, and/or (if you're a sniper) your "sniping bid" can be made before other snipers have a chance to "snipe" their way above your own maximum bid.
A potential sniper should ask himself how he is able to snipe at all. The top bidders in the last couple of minutes of an auction are all snipers. Since these bids all exceed the maximum bids of the non-snipers, this suggests (tiny incremental bids notwithstanding) that non-snipers are either underbidding, or snipers are all overbidding.
This has more to do with the fuzzy nature of peoples' maximum bids. The solution to that is to have larger bid increments, like real auctions do. It's kind of stupid to allow people to bid pennies over someone else's bid, because it just feeds this type of behavior. If the average person knows that their maximum is ~$500, and that they wouldn't go as high as $525, maybe a $10 or a $25 bid increment is appropriate.
I tend to do the fire-and-forget thing. I set my maximum, and if I'm out-bid, then oh well. It's more amusement to me to watch people jump all over the bid in the last 30-60 seconds of the auction. Their time must not be worth very much.
In an auction without snipers, the person with the highest maximum bid wins the auction. The other bidders weren't willing to pay what the winning bidder paid.
In an auction with non-snipers and a single sniper, where the sniper wins the auction, the outcome is the same. The non-sniper bidders lost the auction, but they weren't willing to pay what the sniper paid anyway, so nothing was lost. The sniper would have achieved his goal by placing his "sniping" bid earlier in the process.
In an auction with multiple snipers, the non-snipers still don't care, because the snipers are willing to pay more than they are. The snipers at this point are "battling" with each other, not the non-snipers. If they can hold off, deprive the other snipers of information, and try to "strike" at the last possible moment, they might get a better deal, because none of the snipers are bidding what they are willing to spend, they're bidding the amount that they hope will win them the auction. If we were to clairvoyantly assess each person's true maximum bid, the true maximum would probably be more than what the item sold for, at least sometimes. In other words, if the snipers had simply placed their true maximum bid earlier in the process, the item would have likely sold for more.
With an online auction composed primarily of non-snipers, the non-snipers are going to set a good maximum bid. Those that choose to snipe on top of that are probably spending more money than the item is worth. With an online auction composed of more snipers, these tactics probably "work" in that the sniper can probably get a better deal, because nobody's bidding their real maximum. Pretending your maximum is perhaps lower than it really is could work in your favor, but it's unlikely to affect the decisions of the non-snipers.
Snipers are effective only among themselves. With enough non-snipers bidding in auctions, the possible profit that a sniper can make is reduced. IMO, snipers aren't necessarily in it as a cool, calculated way of maximizing return over a large number of auctions. They're in it because it's a game to them, and it's no fun just to have your maximum bid exceed someone else's maximum bid. But if it turns into a last-second battle of the bids, then you feel like you've "won" something, and eBay bidding becomes exciting. This is stupid, obviously, because these people will usually be overbidding. I suspect this behavior is usually just a form of gambling or shopping addiction and many of these people need to actually seek treatment.
That's not to say that there aren't "legitimate" snipers who do do it for purely economic reasons, but it still doesn't really add up, IMO. I've had lots of auctions sniped from me, and I think the whole thing is pretty amusing. I've seen auctions that I've bid on end up with a winning bid exceeding the price for the same item from Best Buy, all done in the last 30 seconds. Tell me how that makes any economic sense.
I still don't understand why people think they've "stolen" or "won" something cleverly by using "sniping" tactics. They've simply placed a bid that was more than the highest bid I was prepared to make. I've never done it and I've never understood why others do it. If you're really willing to spend more than I am, just set your maximum bid there and let it work out on its own. People don't seem to "get" the whole proxy bidding concept, IMO.
Though small bid increments certainly make it a little more difficult to find your maximum bid. You can say "$50, tops", but then someone wins it with a maximum bid of $50.01, and you feel stupid. That shouldn't happen, but the fix to that is using larger bid increments. You shouldn't have to strategize in a proxy bidding process.
But if you put a sign above your door that says "Open to all visitors!", then it's not. That is effectively what an open WiFi access point is doing, and in most access point configurations, the implications of that setting are clear.
Actually, after looking at the act more closely, it would seem (to my layperson's eyes) to put enforcement squarely within the banking system itself. So if the federal banking system can't get your issue resolved, perhaps you would have to resort to your own suit...
You wouldn't need "10 years of lawyers". This isn't a civil tort. If they're doing something illegal (as in, against the law), you would not prosecute them. Your government (local, state or federal) would. Talk to the branch manager, and if the story really is as has been told, and they aren't willing to release the hold, refer them to your state's attorney general. You can even avoid paying for a stamp if they accept e-mails.
Would you trust even very intelligent humans to be able to follow the three laws as written even if they actually desired to do so?
I think you're assuming that any "proper" artificial intelligence must necessarily resemble human intelligence and be sentient and capable of free will. Long before we reach that point, we will have exceptionally advanced intelligence that is still very much constrained by its programming.
Even if AI must necessarily have free will, that doesn't mean they can't be given laws in the form of instincts. Humans have a tremendous drive for self-preservation, for example. When you're extremely thirsty you have a nearly uncontrollable need to seek water. Your body is hard wired by evolution to do these things, even in the absence of higher brain function. Who's to say machine intelligences can't be hard wired in a similar manner?
They require nearly omniscient comprehension of the effects of ones actions -- how can you know that you have to refuse to drive to the mall and pick up three cans of tomato sauce because if you don't you'll be in a car wreck with a little old lady and break rule 1?
Not really. They just need to be able to have a basic understanding of the situation and the ability to predict the consequences of inaction and of intervention. That requires a lot of information and a lot of processing, but it doesn't really require sentience or a human mind. Understanding consequences is mostly a pattern matching exercise.
You imply that developing a website using web standards takes longer.
No, he's implying that it's more expensive.
High school kids that can vomit out "IEML" web sites that meet business requirements are a dime a dozen. People producing higher quality, standards-compliant web sites that meet and exceed business requirements are significantly more expensive, even considering the likelyhood that they'll be able to do it faster.
A = The Linux Kernel
B = The binary display drivers
C = The shim that connects A to B
The issue is with the aggregation of the display driver and the shim with the kernel. The shim is the only binary component that needs to be licensed under the GPL, and it is. Source is provided. The binary driver itself (B) is not licensed under the GPL and can be freely aggregated and distributed with the rest of the kernel.
The only point where this gets dicey is when you statically link all three together. Linked together, they form an entirely new thing derived from all three, and could not be legally copied and redistributed unless the whole thing were licensed under the GPL. If you don't statically link everything together, though, there remains no copyright issue, and thus no GPL issue. You can interpret the GPL however you want, but it's copyright law that decides whether or not the GPL even applies, and it's copyright law that establishes what it means to be a derivative work, not the GPL.
That's just it, though, the license doesn't matter squat here. It's copyright law that decides what is a derivative work and what isn't, not the GPL (which is simply a license, after all).
If the core display drivers were written completely independently of the Linux kernel or any other GPL code, then they are completely independent. You can never say that one was derived from the other, or vice-versa.
If you write a "shim" that connects the two, and that shim uses code (source or binary) from both the kernel and the display driver, you could say that the shim is a derivative of both. That will never make the display driver itself a derivative of the kernel. The GPL doesn't rewrite history.
Things only get tricky when you statically link things together. If the kernel is statically linked with the shim, which in turn is statically linked with the display driver itself, you end up with one (new!) functional binary computer program that incorporates both GPLed and proprietary code. Distribution of that is another matter entirely, because that would appear to be a derivative work (a combination--not mere aggregation--of all three). But that's trivial to solve: don't statically link!
None of this would be a violation as long as the script is not run behind the users' backs, and as long as the user is explicitly informed they're installing a proprietary driver, told what the consequences are, and asked to click an "AGREE" button or something similar.
Where do these requirements come from? I can't find them in the GPL. How does giving explicit notice to the user (who isn't doing anything relevant to copyright law) affect what is inherently a copyright issue?
If A and B are developed completely independently, you cannot, in any context whatsoever, legal or otherwise, ever claim that B was derived from A.
If you create C, which incorporates elements from A and B, you can say that C is derived from both A and B. Because it is. If A is licensed under the GPL, you cannot redistribute C unless it, too, is licensed under the GPL. By licensing C under the GPL, the elements from B that made it into C are thus licensed under the GPL. B as a whole is not.
Derivation is not communicative. You cannot say that since C was derived from B, that B is also derived from C. GPL may be viral, but it hasn't mutated to the point where it can infect backwards in time.
The issue of derivation is only controversial when you bring the discussion up in a room of thousands of legal noobs. The issue is actually fairly straightforward from a legal perspective, and I believe it's already been tested.
but they should open their site with a uniform warning page allowing the site to be filtered thereafter, or other such methods.
The PICS standard has been around for years and is designed to do precisely this in a machine-readable way. Put up a PICS label on your web page (or in the HTTP headers), configure your browser to prevent access to PICS-labeled sites that say they contain porn, maybe configure it also to prevent access to sites without PICS ratings, lock the browser configuration with a password, and viola! A kid-safe browsing experience. IE supports all of this today. Mozilla opted not to support it because nobody uses it. Yet in all respects it is the perfect answer to this problem. I don't really understand why no one is jumping on this.
You still have to deal with sex sites that deliberately mis-label, but that really starts to cross a line between a shady porn site and one deliberately trying to target minors by labeling themselves as kid-friendly. Legal remedies for that already exist.