I'd say that's one of the more trivial things in the IP world to spoof. I guess what we call "trivial" is relative.
If spoofing is a concern, then run djbdns instead of BIND. djbdns's cache uses 32-bit identifiers by incorporating the source port into the id.
Aside from the fact that "oh, it works, just replace all instances of the most popular nameserver on the Internet with another" isn't going to be very popular (if we're going to be ripping up major infrastructure, as I said above, I'd rather be doing things right, and fixing more problems that allow through spam than just impersonating servers), a lot of folks are going to have firewalls that can't handle djbdns' technique, and they then need to be told that they need to replace their firewalls, also not very popular. Spam is bad -- replacing the mail, DNS, and firewall daemons throughout the Internet to fix a single issue that does not even come close to stopping spam is unacceptable.
o DomainKeys allows user-level granularity. You can use as many keys as you want to administer.
I am open to the possibility that I am drastically misreading the DomainKeys proposal. I have only seriously taken a look at DomainKeys recently, and while I'm reasonably sure that your statement is not true (at least in the straightforward sense of not requiring a domain-per-user), I am quite open to having my mind changed.
Reading through that document, this is my understanding. DomainKeys-related authentication is entirely done for the benefit of the receiving server, and authorizes the sending server. It seems that one may only set a DomainKeys authentication rule of the following format: "if message is signed by one of the set of keys registered for this domain, then accept the message". There is no way to say "If message is signed by one of the set of keys registed for this user:domain tuple, then accept this message". DomainKeys provides functionality for multiple keys per domain (the design of which I must give the Yahoo folks a hand for -- they worked around a number of DNS-related issues here, including some subtle ones, like the problems of rapidly switching keys due to caching). However, every one of these keys authorizes every user. If a user's account is compromised, he may send spam that appears to the remote system to be valid mail from any user, as is also the case if the source mail server is compromised.
It's not *quite* as silly as it sounds. There's a Spy Hunter 2 release that came out not long ago. It wasn't very much fun.
What makes Hollywood think that a good game means a good movie is beyond me -- perhaps marketing is *so* expensive that it's worth it to double up on effort. I can't figure out why Woo is so obsessed with doing a video game movie. Aside from Resident Evil (which wasn't a *great* movie by any measure), all of the video game movies I can think of have pretty much sucked.
You are assuming an even distribution in alpaca quality. We know that this is not the case, as there are specific alpacas that are banned from being exported.
If smugglers are trying to export the best of the best, it's likely that:
* These alpacas are separated from the rest of the herd to avoid unwanted damage and breeding. This probably makes them easier to identify.
* Even if mixed with lesser speciments of their breed, these alpacas may be visually differentiable to someone familiar with judging alpaca quality. For instance, let us assume that Peru is breeding alpacas to have particularly large, firm rear ends. One would imagine that someone that has worked intimately with alpacas for years would be able to quickly visually skim over alpacas and identify the ones to be stolen. It's even possible that they could take part in an "inside job" -- being hired by an alpaca owner, identifying the best alpacas, and then taking a list or marking these to later be stolen.
* There is presumably some dissuasive factor involved in making the statement of microchip presence at *all*. Heck, the chips don't even really have to be there -- it'd drive a smuggler mad to think that he stole what seems to be a really excellent, high-quality alpaca, but cannot find the supposed embedded chip.
* I'm of the suspicion that many Peruvian alpaca rustlers may have been deprived of a thorough statistics education (thus forcing them to smuggle alpacas instead of becoming credit card market analysts). They may not catch on to how unlikely it is to get the short straw -- Americans certainly don't when it comes to lotteries, for instance.
I haven't really ever watched a Voyager episode that I enjoyed. I quickly gave up on it. I've yet to watch an Enterprise episode. Just kind of lost interest due to Voyager. It was kind of a touchy-feely politically correct show rather than the frequent examination of philosophical problems that came up in the earlier Treks.
I did generally like TNG and DS9, though. Never watched much of The Original Trek.
I don't really understand why people get so rabid about Star Trek in general, though. It's reasonably fun to watch, yes. It elevates the status of science (well, at least pseudoscience, but one can generally put a plausible interpretation on things) and engineering, which is not very common in the media. There was some good acting -- I really do like Patrick Stewart. The makeup is *very* good. It's interesting to see positive predictions about the future -- a *lot* of movies seem to go in for futuristic dystopias. Finally, for such a long-running set of series, things didn't get too formulaic -- there was definitely good writing.
Q is not a real character. He's a personification of a plot device -- deus ex machina. Whenever the writers get stumped about how to connect up a plot, they can always throw in Q and get a usable script.
Writer 1: "Darn, I really hate doing sci-fi movies. I wish we could do some historical fiction for a change."
Writer 2: "Yes, I've always wanted to do something on the Civil War."
Yes, but the legal definition of child pornography is based on whether or not a child was used in the production, not whether it just looks that way. See other replies for links.
Right. I did not use such an example -- I was including a 17-year-old in my scenario.
The police will manage to arrest you if you hire a 17-year old to strip for your pleasure, regardless of whether there's a camera. They'll call it "sexual abuse of children" or somesuch, even if you never touch her.
However, the point that's still disputable is so-called "naturist" photography.
If a photograph is taken of a 17-year-old nudist and then included in a work that is judged to be pornographic (see above-mentioned hypothetical "Hot Babes of Florida), that work is legally child pornography. I think that few folks would consider this art.
I'm not trying to argue that something should be art (and hence non-pornographic) or not. I'm trying to take something that would definitely be classified as pornographic. The scenario is designed as an example where the subject suffers no harm by any universally accepted metric (by Victorian values, being seen in the nude would degrade the person's value as a woman, but this is certainly not a universally-accepted metric, nor does this ethic clearly translate into a real-world benefit).
I do not believe that a court would convict the person taking the pictures of sexual abuse. The person would, however, be charged with the production of child pornography.
As another good example -- while it is unlikely that such a case would go through the courts, if a 17-year-old girlfriend takes a (yes, non-artistic) nude photograph of her 17-year-old boyfriend to keep with her, she is in violation of child pornography laws.
Remember that in the USA, illegal child porn is only pictures whose production actually involved the sexual abuse of children- not just ones that look that way.
Not true. The legal definition of pornography in the United States is based on whether the material is obscene or not. There are not excellent hard rules, but in general, if the main point of the material is to inspire arousal, it's pornography.
It is entirely possible to produce material that is pornographic in nature by the US legal definition and has as a subject children without sexual abuse being present in the least.
Taking pictures, for instance, of a nude seventeen-year-old posing to go in some book called "Hot Babes of Florida" is mostly definitely the production of child pornography, and can result in the invocation of anti-child-pornography laws. Sexual abuse is not a prerequisite.
of course i understand that child porn is wrong and so on..
I'd just like to point out that "child pornography" (that which is illegal in the United States) is not a subset of "content depicting real people having sex with children". Take, say, a nude picture of a seventeen-year-old "intended to incite lust" -- this falls under (the extremely harsh) US child pornography laws.
My take on it is this -- anti-child-pornography laws are primarily a function of a Victorian ethical set. If some kid in a tribe in Africa doesn't give a damn whether some explorer takes a picture of him in the nude (and it's pretty clear to anyone but a religious fundamentalist with a percieved religious mandate against nudity that he isn't being hurt by it), why should the same apply to others?
We have to have a strong social taboo in the US against nudity, and shame those that are seen in the nude. That's not a universal absolute or clearly benificial -- it's a result of us being a heavily Christian nation founded by a bunch of religious extremists.
Anti-child-pornography laws are frequently defended against by people who pull out examples of things like three-year-olds being sexually abused. There are a number of existing laws to deal with this (like, sexual abuse and the like) without ever needing to enter the realm of child pornography. So you don't *need* anti-child-pornography laws to exist to eliminate the social issues that they were introduced to help deal with.
The second main argument that I've seen in favor of anti-child-pornography laws have been those arguing that those exposed to fetish material of one sort or another are more likely to actually engage in such (illegal, in this case) sexual activities. I can't buy into this. There are many widespread fetish communities that enjoy fantasizing about sexual activities that would be illegal to act out in real life -- consider cannibalism/snuff fetishists, or rape fetishists. One does not see mass canniablism in real life -- I can't buy into the second argument without a justification for the existing counterarguments and studies that back such censorship.
There is some argument that it's easier to help avoid sexual interaction with children (which I think can be reasonably argued causes social difficulties) by criminalizing the possession of child pornography. While this has some small degree of reasonableness ("It's hard to catch criminals, but if we crimminalize people that are easier to catch, it's easier for us"), I think that it goes too far. This same rationale can be used to support a number of laws that I find objectionable, including the DMCA ("It's illegal to reverse-engineer copy protection, because it's easier to locate and imprison copy protection researchers than it is to nail those actually committing the crime "). By the same logic, we might as well criminalize being Arabic in the United States (since it's an easier way to track down those who might assist Islamic terrorists). Sure, said Arabs aren't actually causing social problems as a group (as is the case, in my opinion, with fetishists), but they might be involved with assisting Islamic fundamentalist terrorists.
My operating systems professor held up our "evacuation plan" for the case of a terrorist attack. The university administration felt the need to produce one after 9/11.
It was really pretty funny.
First of all, you have to understand that CMU borders Squirrel Hill, which is a sizeable and extremely heavily Jewish community.
So a lot of "community centers" like churches volunteer for such things in the case of emergencies -- to be gathering points to identify people that need medical care and do head counts and all that.
So there's a long list of these, including a nearby Jewish temple listed as gathering points.
Naturally, the temple (and *only* the temple) has a asterisk next to it and a note at the bottom saying "in the event that this location is unavailable, children in the day care center will be taken to an undisclosed location", yadda yadda yadda.
Some suit clearly thought "Well, when those Arabs come over here, you can be *darn* sure that in addition to nailing a research institution, they're also going to be sure to waste any temples in the area." Sigh.
(That said, the fact that kids in day care would be taken to an "undisclosed location" would seem to do more to panic parents than anything else, but what do I know.:-) )
Research institutions are pretty boring targets. Many universities do work that end up in military stuff eventually, but there is a significant lag between a university doing work and practical stuff showing up in US military hardware.
Because of some quirks (like encryption and nuclear weapon simulation, some of the early computer uses) fell under the purview of the military, a good amount of computer science funding comes from DARPA (Defense Advanced Research Projects Agency) rather than the NSF (National Science Foundation) as one would expect. Technically, CMU doesn't do classified research on campus, but it does do a *lot* of military-funded work. The SEI doesn't like to talk about exactly what it does, even though technically most of their research is supposed to be non-classified. Also, sometimes researchers get sold things as being somewhat different from their actual use. I've twice now spoken to people that got grants and worked on "non-combat" systems that were rather misleadingly labeled. One person was working what was billed to him as a "search and rescue" vehicle that could autnomously track people, map areas, and the like. He was rather appalled when he got the final vehicle chassis and there was a rather large weapon mount and fire control system on the controls system -- hardly the innocent "search and rescue" application that he had been told about.
CMU claims that it generally doesn't work directly on "combat systems". I get the vague impression that what this tends to come down to is that DARPA and friends have CMU (and some similar institutions) do the hard work (map-building, pathfinding, missile guidance, and the like), and then hire defense contractors to do the actual integration of such systems. The academics can, as long as they choose to do a bit of eye-averting, maintain a clean conscience and truthfully claim in PR releases "we don't make weapons here".
Unfortunately, as long as so much CS funding comes from DARPA, there isn't a whole lot that can be done about the situation -- if people want to be able to do research, they need to get funding from somewhere, and that is very frequently DARPA. The only fix would be to move more government budget from the DoD to the NSF, which doesn't seem very likely to happen.
It's a lot easier for Bush to demand billions for "homeland security" (of which much eventually winds up in the pockets of research institutions and defense contractors) from scared people than it is for someone to make a convincing request for "money for research in the sciences for the betterment of mankind" when so many people are getting old and are watching the Social Security funding that they were counting on rapidly slip away.
One of the key principles is that posting anonymously with and without logging in should be completely indistinguishable.
To an *external user*. As I've pointed out, Slashdot necessarily keeps IP logs for at least 36 hours or so (and may keep them permanently), so currently it is quite possible to distinguish between a random AC and a user marked as an AC.
If your suggestion were to happen, member details would have to be tied to anonymous postings, which is a Bad Thing.
Here is a system that I claim fulfills the features I claimed and does not require member details to be tied to anonymous postings (well, any more than is currently the case).
* When a post is made, the system checks to see if the user is logged in, if "Anonymous" posting is selected by the user, and if the user is a subscriber. If all three of these are the case, the system marks the log entry as being "short-lived". The system periodically purges all logged IP associations with entries that are "short-lived" and more than 36 hours ago.
This does not require tying member details to anonymous postings (in either an internal or external manner). It is externally completely indistinguishable from regular AC posting (and as I pointed out before, differentiated internally only by whether the IP is permanently logged or not). It is certain that Slashdot currently at least temporarily logs IPs associated with pots. The only issue that I can think of is that it is clear which posts belong to subscribers and which to non-subscribers (but only internally, not externally). There is no more data than that single bit, though.
Ease up on the FUD, AAC isn't any more closed than the MP3 you're so eager to praise. They're both open MPEG standards.
They are both MPEG standards -- but implementing AAC encoders or decoders requires purchasing licenses.
Unfortunately, unlike W3C-approved standards, MPEG standards do *not* need to be either patent-unencumbered or have a blanket license granted for implementation purposes upon standardization.
Fraunhoffer has claimed that it has patent rights over MP3. You can look at an analysis here. Basically, they have patents that cover encoders (but have been ignoring free encoders thus far), and while they claim to have patents that cover decoders, some folks have taken issue with this point, and concluded that decoders can be freely produced.
To be honest, the only times I've read anything directly written by Tanenbaum (and been aware of the fact) was when reading the minor flames between him and Linus and this recent webpage. (My OS class used the "dinosaur book".)
From this tiny dataset, it doesn't really seem that AST is particularly arrogant. I've written much nastier flames than AST did on that infamous occasion. He backs up his claims -- "microkernels are better -- I can say this because I know better than you due to 20 years studying them" would be arrogant. "Microkernels take a relatively minor 20% performance hit or so and provide easier debugging" is a different story.
Re:His comment on Slashdot:
on
More From Tanenbaum
·
· Score: 2, Interesting
Despite popular opinion, I really don't have a PDP-11 in my basement.
Damn! Me and my friends have all been sure that you've had one in there for *years*.
well tough. It's not like Bill Gates, Steve Jobs, Gary Killdall, et al isn't a little bit of a "prick" too.
Jobs is an arrogant guy that didn't do the amazing engineering work that he gets credit for ("The Mac kicks ass so Steve Jobs kicks ass" is very faulty logic.) Maybe he's a good organizer, but he's also the source of a lot of the decisions that I feel hurt the Mac, including deliberately limiting expandability and the single button mouse. AFAICT, Jobs' main notable talent is his ability for marketing himself and associating himself with every good thing that's been achieved by all the engineers working on the Mac.
No it isn't, and I resent that! Slander is spoken. In print it's "Libel".
Sincerely,
Kenneth Brown President, Alexis de Tocqueville Institution
I am fairly certain that this person is not the real Kenneth Brown.
Based on his misleading attacks on Linus, I would say that the real Kenneth Brown would have written something like the following:
No it isn't. Three legal experts that I've consulted have agreed that the person is making entirely unsubstantiated claims. Open source advocates have been known to make wildly inaccurate legal claims to suppress people that disagree with them. This economist that I consulted confirmed that suppressing data leads to less efficient markets, which leads to loss of money. Personally, I wouldn't use any OS that causes you to lose money.
A future version could offer the option of running the filesystems in userspace if you want.
There are two separate systems for doing this in Linux today: LUFS and FUSE.
I don't believe anyone's bothered to port the existing kernel-based filesystems to either because they, y'know, work and are faster, but it would be possible to do so.
Re:We should set up better Open Source Marketing
on
More From Tanenbaum
·
· Score: 1
And Tanembaum provides lots of nice quotes for the profesional marketers from Red Hat/IBM/Novell/whatever, computer columnists, etc.
Yes -- despite the "Li'l ol' me? I'm just a naive ivory-tower academic!" spiel, Tanenbaum seemed to be throwing some awfully juicy tidbits out there.
Seriously, if someone wrote a book and said "0x0d0a said that Bill Atkinson and Steve Wozniak stole all their work from Bill Gates!", I'd be more than a little angry.
Re:I have the PDF of the first 92 pages of the boo
on
More From Tanenbaum
·
· Score: 1
The difference is that even Stallman, the most extreme of the extreme, is hardly anti-business, though he does have issues with IP, which affect some businesses.
Microsoft is very frequently "anti-puppy". Perhaps one day, SuSE, Red Hat, and IBM will be pulling orchestrated FUD campaigns, trying to lock people in to their products, and foist off technically inferior products that enhance their business positioning. Until then, though, Microsoft gets the "anti-puppy" award.
Re:I'm with Tannebaum about microkernels
on
More From Tanenbaum
·
· Score: 3, Insightful
The only thing they accomplish is code memory savings.
Ironically enough, this is not really the case for Windows due to a pretty poor (in retrospect) design decision.
Windows supports pageable kernel memory, unlike Linux. This leads to vastly more complexity than Linux (since you *really* don't want kernel code like your paging code, say, to accidently cause a page fault). You also don't want anything that might be involved with paging (want to support a remote pagefile? Hope network support doesn't page! A pagefile on a USB keydrive? You'll need USB code and thus probably power-saving code and so forth to never cause pagefaults). Any *data* touched by any code that can be invoked by a page fault cannot ever be paged out, or else the risk of a hang again appears. Windows maintains different lists of "pageable" and "unpageable" memory.
Linux took a much simpler approach -- kernel memory isn't pageable, but kernel modules can be unloaded -- that increases the simplicity and reliability of kernel code.
The basic penalty for using a microkernel is one extra copy and context switch for every file system operation. If your system is doing anything besides I/O, you'll probably never notice. If you're running a web server that serves mostly plain pages (little Perl, Java, PHP, etc.), you'd probably notice the overhead.
The bus multiplier on processors and the cycles-for-a-main-memory-access have steadily increased over the past decade or so. This has steadily increased the cost of a page table cache flush, and thus steadily increased the cost of a context switch.
Re:The "Linux is obsolete" flamewar
on
More From Tanenbaum
·
· Score: 3, Interesting
Read more closely in Linus' old writings -- Linus originally intended Linux to be a stopgap solution, and expected that HURD would end up taking over the position of flagbearer.
As it happened, HURD ended up sucking, and so Linux remained the default.
I think the thing that set Linus off was more the fact that Linux was being insulted (probably Prof. Tanenbaum was feeling a bit cranky that day or something, and Linus was in a fighting mood...)
It's funny how emails waaay back then, from when Linus was still a pretty small fry guy, can come back to haunt the people involved.
It's something to think about before posting to a mailing list: If I get really famous ten years from now, is this going to cause me or someone I respect hurt?
(2) "Linus didn't write Linux"... sounds like a dorky meme. Besides looking stupid on a shallow marketing level (why do you think it's called Linux?) and being factually stupid (he sure did write it), it's one of those big yawner don't-care issues. Joe CIO isn't gonna go "oooh! better not deploy Linux after all! Linus didn't actually write it!"
Yes, but I think that the "Linus Didn't Write Linux" bit is more present on Slashdot and friends because it looks so ridiculous to us. The actual exerpts, if you read them, are very misleading, and strongly imply that Linus misappropriated IP in writing Linux. This is particularly entertaining due to the fact that all the parties involved, *including* Tanenbaum (whom you can be quite certain would be crying foul if someone was stealing his work) have pointed out that Linus certainly *is* responsible.
My theory is that Microsoft uses AdTI to float many different trial balloons.
My theory is that AdTI is a trial balloon for Microsoft, the same as SCO was. They need to find third parties that can bash Linux without appearing to be associated with Microsoft.
Yahoo's DomainKeys proposal, unlike Caller ID and SPF, eliminates many of the drawbacks that Caller ID/SPF have. If I *had* to see one of the three used, I would call DomainKeys the technically superior of the three, and push for it. Caller ID and SPF both use host-address-based authentication (Are the lessons of yesteryear so quickly forgotten?). Because of caching, they have a delay period while mail hosts can be moved around. DomainKeys uses key-based authentication -- a much better system. The only disadvantage of DomainKeys versus Caller ID or SPF is its greater CPU load. DomainKeys does not break forwarding, unlike Caller ID and SPF. DomainKeys was actually designed with consideration for needing to move mail servers to new IPs (and the fact that hosts cache) and will not cause mail to bounce as "unauthorized" for some period of time after you change your mail server IP.
DomainKeys still imposes some of the drawbacks of Caller ID/SPF. It is generally not possible to run your own mail server. It's technically possible to have multiple servers, but this requires copying the domain private key to the other servers. It's possible to have multiple servers, in Caller ID/SPF by adding those servers to the DNS record as authorized servers. However, with DomainKeys/Caller ID/SPF, one must be on good terms with one's DNS admin to do so, as all require the addition of DNS elements. If I work for Chrysler, I can easily send work-related email from home in a system that checks signatures on a per-user basis. If I want to do so with Caller ID/SPF/DomainKeys, I have to convince the admins to add DNS entries (which is going to require months of meetings in any large company).
DomainKeys *still* passes the security buck off when it comes to DNS (with a vague "oh, people should start using DNSSEC soon anyway"). There are known and clear holes in DNS that make it currently unsuitable for authentication procedures (it's spoofable, its caching can cause successful spoofs to be much more damaging, etc). DNS was *never* designed to be a secure authentication system, and DomainKeys/Caller ID/SPF try to retrofit it into doing so.
DomainKeys/Caller ID/SPF do not handle the throwaway domain problem.
DomainKeys/Caller ID/SPF all follow the "the entire system works only as long as there are no security breeches anywhere within the system -- in such an event, everyone may be negatively impacted" model. All mail servers are considered trusted sources of mail for their domain. If I can beat the local security within *any* domain from the MUA to the local mailserver, I can send spam to everyone. There are a lot of ways to do this -- compromise a host on the network, an account or a mailserver. This security model does *not* work on large-scale systems like the Internet -- if it did, we wouldn't *need* any anti-spam systems because there would be no open relays in existence (the last time we tried using such an approach) and hence no spam.
Caller ID/SPF/DomainKeys all use domain-level granularity rather than user-level granularity. If I compromise a user's computer (let us suppose it is at ford.com), the only option the rest of the world has is to ban all of ford.com, since there is no guaranteed way for them to ensure that a user at ford.com can't just spoof other users at ford.com to the mailserver.
The problem is that Caller ID, SPF, and DomainKeys all fall far, far short of what is necessary (and the "well, it's better than before" argument does not hold much water when severe limitations are being imposed). Every one of the three is aimed at solving a single problem that does *not* encompass the spam problem -- ensuring that mail from a domain passed through that domain's mail servers. There are times when legitimate mail does not do this (and removing this allowance would break functionality that cannot otherwise be provided). Furthermore (and probably of more concern to the masses) there are many ways to still send spam with such a constraint. None of the three, for instance, attempts to deal with throwaway domains (though the SPF people do some vague handwaving about "trust networks" -- ignoring the fact that if trust networks are put in place, there are much better solutions to spam than SPF).
Slashdot Mirror
A) It doesn't make much sense to have cps without data on the CPU of the system.
B) Is this *true* rendering, or using glyph caching?
o It's far from trivial to spoof DNS queries.
I'd say that's one of the more trivial things in the IP world to spoof. I guess what we call "trivial" is relative.
If spoofing is a concern, then run djbdns instead of BIND. djbdns's cache uses 32-bit identifiers by incorporating the source port into the id.
Aside from the fact that "oh, it works, just replace all instances of the most popular nameserver on the Internet with another" isn't going to be very popular (if we're going to be ripping up major infrastructure, as I said above, I'd rather be doing things right, and fixing more problems that allow through spam than just impersonating servers), a lot of folks are going to have firewalls that can't handle djbdns' technique, and they then need to be told that they need to replace their firewalls, also not very popular. Spam is bad -- replacing the mail, DNS, and firewall daemons throughout the Internet to fix a single issue that does not even come close to stopping spam is unacceptable.
o DomainKeys allows user-level granularity. You can use as many keys as you want to administer.
I am open to the possibility that I am drastically misreading the DomainKeys proposal. I have only seriously taken a look at DomainKeys recently, and while I'm reasonably sure that your statement is not true (at least in the straightforward sense of not requiring a domain-per-user), I am quite open to having my mind changed.
Reading through that document, this is my understanding. DomainKeys-related authentication is entirely done for the benefit of the receiving server, and authorizes the sending server. It seems that one may only set a DomainKeys authentication rule of the following format: "if message is signed by one of the set of keys registered for this domain, then accept the message". There is no way to say "If message is signed by one of the set of keys registed for this user:domain tuple, then accept this message". DomainKeys provides functionality for multiple keys per domain (the design of which I must give the Yahoo folks a hand for -- they worked around a number of DNS-related issues here, including some subtle ones, like the problems of rapidly switching keys due to caching). However, every one of these keys authorizes every user. If a user's account is compromised, he may send spam that appears to the remote system to be valid mail from any user, as is also the case if the source mail server is compromised.
It's not *quite* as silly as it sounds. There's a Spy Hunter 2 release that came out not long ago. It wasn't very much fun.
What makes Hollywood think that a good game means a good movie is beyond me -- perhaps marketing is *so* expensive that it's worth it to double up on effort. I can't figure out why Woo is so obsessed with doing a video game movie. Aside from Resident Evil (which wasn't a *great* movie by any measure), all of the video game movies I can think of have pretty much sucked.
You are assuming an even distribution in alpaca quality. We know that this is not the case, as there are specific alpacas that are banned from being exported.
If smugglers are trying to export the best of the best, it's likely that:
* These alpacas are separated from the rest of the herd to avoid unwanted damage and breeding. This probably makes them easier to identify.
* Even if mixed with lesser speciments of their breed, these alpacas may be visually differentiable to someone familiar with judging alpaca quality. For instance, let us assume that Peru is breeding alpacas to have particularly large, firm rear ends. One would imagine that someone that has worked intimately with alpacas for years would be able to quickly visually skim over alpacas and identify the ones to be stolen. It's even possible that they could take part in an "inside job" -- being hired by an alpaca owner, identifying the best alpacas, and then taking a list or marking these to later be stolen.
* There is presumably some dissuasive factor involved in making the statement of microchip presence at *all*. Heck, the chips don't even really have to be there -- it'd drive a smuggler mad to think that he stole what seems to be a really excellent, high-quality alpaca, but cannot find the supposed embedded chip.
* I'm of the suspicion that many Peruvian alpaca rustlers may have been deprived of a thorough statistics education (thus forcing them to smuggle alpacas instead of becoming credit card market analysts). They may not catch on to how unlikely it is to get the short straw -- Americans certainly don't when it comes to lotteries, for instance.
"Subject line" is pushing it -- I don't believe that this is a major problem.
:-) ).
Buffer overflows based on parsing mail is not an uncommon problem, though.
Let's take a look (I'm not going to bother with more than one per client).
Here are bugs for mutt, pine, evolution, kmail, elm (elm is apparently vulnerable to an overflow in the Subject line
I assume that the last is what the grandparent was referring to.
I haven't really ever watched a Voyager episode that I enjoyed. I quickly gave up on it. I've yet to watch an Enterprise episode. Just kind of lost interest due to Voyager. It was kind of a touchy-feely politically correct show rather than the frequent examination of philosophical problems that came up in the earlier Treks.
I did generally like TNG and DS9, though. Never watched much of The Original Trek.
I don't really understand why people get so rabid about Star Trek in general, though. It's reasonably fun to watch, yes. It elevates the status of science (well, at least pseudoscience, but one can generally put a plausible interpretation on things) and engineering, which is not very common in the media. There was some good acting -- I really do like Patrick Stewart. The makeup is *very* good. It's interesting to see positive predictions about the future -- a *lot* of movies seem to go in for futuristic dystopias. Finally, for such a long-running set of series, things didn't get too formulaic -- there was definitely good writing.
Q is not a real character. He's a personification of a plot device -- deus ex machina. Whenever the writers get stumped about how to connect up a plot, they can always throw in Q and get a usable script.
Writer 1: "Darn, I really hate doing sci-fi movies. I wish we could do some historical fiction for a change."
Writer 2: "Yes, I've always wanted to do something on the Civil War."
In Unison: "Q!"
Yes, but the legal definition of child pornography is based on whether or not a child was used in the production, not whether it just looks that way. See other replies for links.
Right. I did not use such an example -- I was including a 17-year-old in my scenario.
The police will manage to arrest you if you hire a 17-year old to strip for your pleasure, regardless of whether there's a camera. They'll call it "sexual abuse of children" or somesuch, even if you never touch her.
However, the point that's still disputable is so-called "naturist" photography.
If a photograph is taken of a 17-year-old nudist and then included in a work that is judged to be pornographic (see above-mentioned hypothetical "Hot Babes of Florida), that work is legally child pornography. I think that few folks would consider this art.
I'm not trying to argue that something should be art (and hence non-pornographic) or not. I'm trying to take something that would definitely be classified as pornographic. The scenario is designed as an example where the subject suffers no harm by any universally accepted metric (by Victorian values, being seen in the nude would degrade the person's value as a woman, but this is certainly not a universally-accepted metric, nor does this ethic clearly translate into a real-world benefit).
I do not believe that a court would convict the person taking the pictures of sexual abuse. The person would, however, be charged with the production of child pornography.
As another good example -- while it is unlikely that such a case would go through the courts, if a 17-year-old girlfriend takes a (yes, non-artistic) nude photograph of her 17-year-old boyfriend to keep with her, she is in violation of child pornography laws.
Remember that in the USA, illegal child porn is only pictures whose production actually involved the sexual abuse of children- not just ones that look that way.
Not true. The legal definition of pornography in the United States is based on whether the material is obscene or not. There are not excellent hard rules, but in general, if the main point of the material is to inspire arousal, it's pornography.
It is entirely possible to produce material that is pornographic in nature by the US legal definition and has as a subject children without sexual abuse being present in the least.
Taking pictures, for instance, of a nude seventeen-year-old posing to go in some book called "Hot Babes of Florida" is mostly definitely the production of child pornography, and can result in the invocation of anti-child-pornography laws. Sexual abuse is not a prerequisite.
of course i understand that child porn is wrong and so on..
I'd just like to point out that "child pornography" (that which is illegal in the United States) is not a subset of "content depicting real people having sex with children". Take, say, a nude picture of a seventeen-year-old "intended to incite lust" -- this falls under (the extremely harsh) US child pornography laws.
My take on it is this -- anti-child-pornography laws are primarily a function of a Victorian ethical set. If some kid in a tribe in Africa doesn't give a damn whether some explorer takes a picture of him in the nude (and it's pretty clear to anyone but a religious fundamentalist with a percieved religious mandate against nudity that he isn't being hurt by it), why should the same apply to others?
We have to have a strong social taboo in the US against nudity, and shame those that are seen in the nude. That's not a universal absolute or clearly benificial -- it's a result of us being a heavily Christian nation founded by a bunch of religious extremists.
Anti-child-pornography laws are frequently defended against by people who pull out examples of things like three-year-olds being sexually abused. There are a number of existing laws to deal with this (like, sexual abuse and the like) without ever needing to enter the realm of child pornography. So you don't *need* anti-child-pornography laws to exist to eliminate the social issues that they were introduced to help deal with.
The second main argument that I've seen in favor of anti-child-pornography laws have been those arguing that those exposed to fetish material of one sort or another are more likely to actually engage in such (illegal, in this case) sexual activities. I can't buy into this. There are many widespread fetish communities that enjoy fantasizing about sexual activities that would be illegal to act out in real life -- consider cannibalism/snuff fetishists, or rape fetishists. One does not see mass canniablism in real life -- I can't buy into the second argument without a justification for the existing counterarguments and studies that back such censorship.
There is some argument that it's easier to help avoid sexual interaction with children (which I think can be reasonably argued causes social difficulties) by criminalizing the possession of child pornography. While this has some small degree of reasonableness ("It's hard to catch criminals, but if we crimminalize people that are easier to catch, it's easier for us"), I think that it goes too far. This same rationale can be used to support a number of laws that I find objectionable, including the DMCA ("It's illegal to reverse-engineer copy protection, because it's easier to locate and imprison copy protection researchers than it is to nail those actually committing the crime "). By the same logic, we might as well criminalize being Arabic in the United States (since it's an easier way to track down those who might assist Islamic terrorists). Sure, said Arabs aren't actually causing social problems as a group (as is the case, in my opinion, with fetishists), but they might be involved with assisting Islamic fundamentalist terrorists.
My operating systems professor held up our "evacuation plan" for the case of a terrorist attack. The university administration felt the need to produce one after 9/11.
:-) )
It was really pretty funny.
First of all, you have to understand that CMU borders Squirrel Hill, which is a sizeable and extremely heavily Jewish community.
So a lot of "community centers" like churches volunteer for such things in the case of emergencies -- to be gathering points to identify people that need medical care and do head counts and all that.
So there's a long list of these, including a nearby Jewish temple listed as gathering points.
Naturally, the temple (and *only* the temple) has a asterisk next to it and a note at the bottom saying "in the event that this location is unavailable, children in the day care center will be taken to an undisclosed location", yadda yadda yadda.
Some suit clearly thought "Well, when those Arabs come over here, you can be *darn* sure that in addition to nailing a research institution, they're also going to be sure to waste any temples in the area." Sigh.
(That said, the fact that kids in day care would be taken to an "undisclosed location" would seem to do more to panic parents than anything else, but what do I know.
Research institutions are pretty boring targets. Many universities do work that end up in military stuff eventually, but there is a significant lag between a university doing work and practical stuff showing up in US military hardware.
Because of some quirks (like encryption and nuclear weapon simulation, some of the early computer uses) fell under the purview of the military, a good amount of computer science funding comes from DARPA (Defense Advanced Research Projects Agency) rather than the NSF (National Science Foundation) as one would expect. Technically, CMU doesn't do classified research on campus, but it does do a *lot* of military-funded work. The SEI doesn't like to talk about exactly what it does, even though technically most of their research is supposed to be non-classified. Also, sometimes researchers get sold things as being somewhat different from their actual use. I've twice now spoken to people that got grants and worked on "non-combat" systems that were rather misleadingly labeled. One person was working what was billed to him as a "search and rescue" vehicle that could autnomously track people, map areas, and the like. He was rather appalled when he got the final vehicle chassis and there was a rather large weapon mount and fire control system on the controls system -- hardly the innocent "search and rescue" application that he had been told about.
CMU claims that it generally doesn't work directly on "combat systems". I get the vague impression that what this tends to come down to is that DARPA and friends have CMU (and some similar institutions) do the hard work (map-building, pathfinding, missile guidance, and the like), and then hire defense contractors to do the actual integration of such systems. The academics can, as long as they choose to do a bit of eye-averting, maintain a clean conscience and truthfully claim in PR releases "we don't make weapons here".
Unfortunately, as long as so much CS funding comes from DARPA, there isn't a whole lot that can be done about the situation -- if people want to be able to do research, they need to get funding from somewhere, and that is very frequently DARPA. The only fix would be to move more government budget from the DoD to the NSF, which doesn't seem very likely to happen.
It's a lot easier for Bush to demand billions for "homeland security" (of which much eventually winds up in the pockets of research institutions and defense contractors) from scared people than it is for someone to make a convincing request for "money for research in the sciences for the betterment of mankind" when so many people are getting old and are watching the Social Security funding that they were counting on rapidly slip away.
One of the key principles is that posting anonymously with and without logging in should be completely indistinguishable.
To an *external user*. As I've pointed out, Slashdot necessarily keeps IP logs for at least 36 hours or so (and may keep them permanently), so currently it is quite possible to distinguish between a random AC and a user marked as an AC.
If your suggestion were to happen, member details would have to be tied to anonymous postings, which is a Bad Thing.
Here is a system that I claim fulfills the features I claimed and does not require member details to be tied to anonymous postings (well, any more than is currently the case).
* When a post is made, the system checks to see if the user is logged in, if "Anonymous" posting is selected by the user, and if the user is a subscriber. If all three of these are the case, the system marks the log entry as being "short-lived". The system periodically purges all logged IP associations with entries that are "short-lived" and more than 36 hours ago.
This does not require tying member details to anonymous postings (in either an internal or external manner). It is externally completely indistinguishable from regular AC posting (and as I pointed out before, differentiated internally only by whether the IP is permanently logged or not). It is certain that Slashdot currently at least temporarily logs IPs associated with pots. The only issue that I can think of is that it is clear which posts belong to subscribers and which to non-subscribers (but only internally, not externally). There is no more data than that single bit, though.
Ease up on the FUD, AAC isn't any more closed than the MP3 you're so eager to praise. They're both open MPEG standards.
They are both MPEG standards -- but implementing AAC encoders or decoders requires purchasing licenses.
Unfortunately, unlike W3C-approved standards, MPEG standards do *not* need to be either patent-unencumbered or have a blanket license granted for implementation purposes upon standardization.
Fraunhoffer has claimed that it has patent rights over MP3. You can look at an analysis here. Basically, they have patents that cover encoders (but have been ignoring free encoders thus far), and while they claim to have patents that cover decoders, some folks have taken issue with this point, and concluded that decoders can be freely produced.
Ogg Vorbis is not encumbered by patents.
To be honest, the only times I've read anything directly written by Tanenbaum (and been aware of the fact) was when reading the minor flames between him and Linus and this recent webpage. (My OS class used the "dinosaur book".)
From this tiny dataset, it doesn't really seem that AST is particularly arrogant. I've written much nastier flames than AST did on that infamous occasion. He backs up his claims -- "microkernels are better -- I can say this because I know better than you due to 20 years studying them" would be arrogant. "Microkernels take a relatively minor 20% performance hit or so and provide easier debugging" is a different story.
Despite popular opinion, I really don't have a PDP-11 in my basement.
Damn! Me and my friends have all been sure that you've had one in there for *years*.
well tough. It's not like Bill Gates, Steve Jobs, Gary Killdall, et al isn't a little bit of a "prick" too.
Jobs is an arrogant guy that didn't do the amazing engineering work that he gets credit for ("The Mac kicks ass so Steve Jobs kicks ass" is very faulty logic.) Maybe he's a good organizer, but he's also the source of a lot of the decisions that I feel hurt the Mac, including deliberately limiting expandability and the single button mouse. AFAICT, Jobs' main notable talent is his ability for marketing himself and associating himself with every good thing that's been achieved by all the engineers working on the Mac.
No it isn't, and I resent that! Slander is spoken. In print it's "Libel".
Sincerely,
Kenneth Brown
President, Alexis de Tocqueville Institution
I am fairly certain that this person is not the real Kenneth Brown.
Based on his misleading attacks on Linus, I would say that the real Kenneth Brown would have written something like the following:
No it isn't. Three legal experts that I've consulted have agreed that the person is making entirely unsubstantiated claims. Open source advocates have been known to make wildly inaccurate legal claims to suppress people that disagree with them. This economist that I consulted confirmed that suppressing data leads to less efficient markets, which leads to loss of money. Personally, I wouldn't use any OS that causes you to lose money.
A future version could offer the option of running the filesystems in userspace if you want.
There are two separate systems for doing this in Linux today: LUFS and FUSE.
I don't believe anyone's bothered to port the existing kernel-based filesystems to either because they, y'know, work and are faster, but it would be possible to do so.
Mac OS 9?
It's not a microkernel.
And Tanembaum provides lots of nice quotes for the profesional marketers from Red Hat/IBM/Novell/whatever, computer columnists, etc.
Yes -- despite the "Li'l ol' me? I'm just a naive ivory-tower academic!" spiel, Tanenbaum seemed to be throwing some awfully juicy tidbits out there.
Seriously, if someone wrote a book and said "0x0d0a said that Bill Atkinson and Steve Wozniak stole all their work from Bill Gates!", I'd be more than a little angry.
The difference is that even Stallman, the most extreme of the extreme, is hardly anti-business, though he does have issues with IP, which affect some businesses.
Microsoft is very frequently "anti-puppy". Perhaps one day, SuSE, Red Hat, and IBM will be pulling orchestrated FUD campaigns, trying to lock people in to their products, and foist off technically inferior products that enhance their business positioning. Until then, though, Microsoft gets the "anti-puppy" award.
The only thing they accomplish is code memory savings.
Ironically enough, this is not really the case for Windows due to a pretty poor (in retrospect) design decision.
Windows supports pageable kernel memory, unlike Linux. This leads to vastly more complexity than Linux (since you *really* don't want kernel code like your paging code, say, to accidently cause a page fault). You also don't want anything that might be involved with paging (want to support a remote pagefile? Hope network support doesn't page! A pagefile on a USB keydrive? You'll need USB code and thus probably power-saving code and so forth to never cause pagefaults). Any *data* touched by any code that can be invoked by a page fault cannot ever be paged out, or else the risk of a hang again appears. Windows maintains different lists of "pageable" and "unpageable" memory.
Linux took a much simpler approach -- kernel memory isn't pageable, but kernel modules can be unloaded -- that increases the simplicity and reliability of kernel code.
The basic penalty for using a microkernel is one extra copy and context switch for every file system operation. If your system is doing anything besides I/O, you'll probably never notice. If you're running a web server that serves mostly plain pages (little Perl, Java, PHP, etc.), you'd probably notice the overhead.
The bus multiplier on processors and the cycles-for-a-main-memory-access have steadily increased over the past decade or so. This has steadily increased the cost of a page table cache flush, and thus steadily increased the cost of a context switch.
Read more closely in Linus' old writings -- Linus originally intended Linux to be a stopgap solution, and expected that HURD would end up taking over the position of flagbearer.
As it happened, HURD ended up sucking, and so Linux remained the default.
I think the thing that set Linus off was more the fact that Linux was being insulted (probably Prof. Tanenbaum was feeling a bit cranky that day or something, and Linus was in a fighting mood...)
It's funny how emails waaay back then, from when Linus was still a pretty small fry guy, can come back to haunt the people involved.
It's something to think about before posting to a mailing list: If I get really famous ten years from now, is this going to cause me or someone I respect hurt?
(2) "Linus didn't write Linux" ... sounds like a dorky meme. Besides looking stupid on a shallow marketing level (why do you think it's called Linux?) and being factually stupid (he sure did write it), it's one of those big yawner don't-care issues. Joe CIO isn't gonna go "oooh! better not deploy Linux after all! Linus didn't actually write it!"
Yes, but I think that the "Linus Didn't Write Linux" bit is more present on Slashdot and friends because it looks so ridiculous to us. The actual exerpts, if you read them, are very misleading, and strongly imply that Linus misappropriated IP in writing Linux. This is particularly entertaining due to the fact that all the parties involved, *including* Tanenbaum (whom you can be quite certain would be crying foul if someone was stealing his work) have pointed out that Linus certainly *is* responsible.
My theory is that Microsoft uses AdTI to float many different trial balloons.
My theory is that AdTI is a trial balloon for Microsoft, the same as SCO was. They need to find third parties that can bash Linux without appearing to be associated with Microsoft.
Yahoo's DomainKeys proposal, unlike Caller ID and SPF, eliminates many of the drawbacks that Caller ID/SPF have. If I *had* to see one of the three used, I would call DomainKeys the technically superior of the three, and push for it. Caller ID and SPF both use host-address-based authentication (Are the lessons of yesteryear so quickly forgotten?). Because of caching, they have a delay period while mail hosts can be moved around. DomainKeys uses key-based authentication -- a much better system. The only disadvantage of DomainKeys versus Caller ID or SPF is its greater CPU load. DomainKeys does not break forwarding, unlike Caller ID and SPF. DomainKeys was actually designed with consideration for needing to move mail servers to new IPs (and the fact that hosts cache) and will not cause mail to bounce as "unauthorized" for some period of time after you change your mail server IP.
DomainKeys still imposes some of the drawbacks of Caller ID/SPF. It is generally not possible to run your own mail server. It's technically possible to have multiple servers, but this requires copying the domain private key to the other servers. It's possible to have multiple servers, in Caller ID/SPF by adding those servers to the DNS record as authorized servers. However, with DomainKeys/Caller ID/SPF, one must be on good terms with one's DNS admin to do so, as all require the addition of DNS elements. If I work for Chrysler, I can easily send work-related email from home in a system that checks signatures on a per-user basis. If I want to do so with Caller ID/SPF/DomainKeys, I have to convince the admins to add DNS entries (which is going to require months of meetings in any large company).
DomainKeys *still* passes the security buck off when it comes to DNS (with a vague "oh, people should start using DNSSEC soon anyway"). There are known and clear holes in DNS that make it currently unsuitable for authentication procedures (it's spoofable, its caching can cause successful spoofs to be much more damaging, etc). DNS was *never* designed to be a secure authentication system, and DomainKeys/Caller ID/SPF try to retrofit it into doing so.
DomainKeys/Caller ID/SPF do not handle the throwaway domain problem.
DomainKeys/Caller ID/SPF all follow the "the entire system works only as long as there are no security breeches anywhere within the system -- in such an event, everyone may be negatively impacted" model. All mail servers are considered trusted sources of mail for their domain. If I can beat the local security within *any* domain from the MUA to the local mailserver, I can send spam to everyone. There are a lot of ways to do this -- compromise a host on the network, an account or a mailserver. This security model does *not* work on large-scale systems like the Internet -- if it did, we wouldn't *need* any anti-spam systems because there would be no open relays in existence (the last time we tried using such an approach) and hence no spam.
Caller ID/SPF/DomainKeys all use domain-level granularity rather than user-level granularity. If I compromise a user's computer (let us suppose it is at ford.com), the only option the rest of the world has is to ban all of ford.com, since there is no guaranteed way for them to ensure that a user at ford.com can't just spoof other users at ford.com to the mailserver.
The problem is that Caller ID, SPF, and DomainKeys all fall far, far short of what is necessary (and the "well, it's better than before" argument does not hold much water when severe limitations are being imposed). Every one of the three is aimed at solving a single problem that does *not* encompass the spam problem -- ensuring that mail from a domain passed through that domain's mail servers. There are times when legitimate mail does not do this (and removing this allowance would break functionality that cannot otherwise be provided). Furthermore (and probably of more concern to the masses) there are many ways to still send spam with such a constraint. None of the three, for instance, attempts to deal with throwaway domains (though the SPF people do some vague handwaving about "trust networks" -- ignoring the fact that if trust networks are put in place, there are much better solutions to spam than SPF).