Copyright and trademark works with relatively little objection, because there are very many ways to write a novel, compose a piece of music, and (to a lesser extent) design a trademark, so even if one of the ways is monopolized by copyright/trademark, the scope for innovation does not decrease significantly.
And the problem many have with most software patents (and many in other fields too) is that they cover one of very few good ways of doing something, or the idea of doing some obvious thing itself, so others have no way around it no matter how good they are. Some people might find it fair, but many do not, and find it counterproductive.
For patents/trademarks on smells and tastes, the same thing applies --- I won't object much if and only if the smell/taste must be extremely special (but do we have such a good nose/tongue?), in other words, it must contain a large amount of information.
I could not RTFA but if it just mean what you have said, who cares? I ignore the trademark part for now; as for patents, if no one can hold a patent on software they created, this is not very different from the case in many countries where software patents are not allowed at all. I have never heard of FSF holding any patents, and if Redhat has some, they have stated that they will be using them defensively only, so they probably won't mind losing them very much if the main purpose of them has disappeared in the first place.
Actually even if copyrights vanished overnight, the free software community will be one of the least affected. Of course GPL would no longer mean anything, and much troubles may ensue, but life will probably go on in the community. Now that we are talking about patents, a court decision such as the one you mentioned will hardly do any harm to our community.
AFAIK, if encrypted protocols like SSL are used, each of the two parties just generates a new public/private keypair randomly, and they exchange the public key, using certificates to verify the authenticity of the other party's public key. After that, all traffic from A to B is encrypted by A using B's public key, so only B can decrypt it using his private key (actually a symmetric key is generated and exchanged in this way and symmetric ciphers are used after that to improve performance, but this does not matter much to security).
I'm not a security expert, so some of the above may be wrong, but it is obvious that capturing everything on the wire is not sufficient for successful wiretapping, since the private keys required for deciphering all these information is never sent on the wire (actually they are often never written to disk either), and inferring the public key from the private key is considered practically impossible.
Although some outside people has done the patent investigation, the developers themselves decide not to, so they really don't know what patents are they really stepping on.
I think they are wise not to spend time investigating patents. Assuming that all the patents concerned are valid (actually many of them can probably be struck down, but at considerable cost of time and money), the situation may change from unknowningly infringement of 340 patents (57 undiscovered by that MySQL employee) to knowingly infringement of the uncorrectable 97 and unknowningly infringement of the undiscovered 57. Not a better situation IMHO, and a lot of time would have been spent in avoiding patents.
Last November I replaced my then 6-year-old P2-based machine to a Pentium 4 with integrated graphics (i845G). Rather cheap, the open source driver works good enough, glxgears get 400fps on default resolution and 56fps on 1024x768, glBoom (a Doom source port) and glQuake2 runs smoothly without slowdowns, and many small OpenGL games work well too, but I haven't tried anything newer (from my classmates' experience Quake3 and CS should be playable, but newer stuff probably won't work well). In short, it performs just a little better (up to 2x in glBoom) than the 4MB Riva128 on my old machine (cost me $250 in early 1998), but then the Riva128 didn't have hardware-accelerated 3D under linux, even with their closed-source drivers.
Of course, if you want to use AMD, or want to play newer eye-candy-filled games like Doom3, my advice cannot help you.
Even if he dies the copyright still lasts for 5 to 10 years... the grandparent's way has nothing to do with the author's life. Just make sure his descendents have the ability to publish the work and profit from it.
Well, what you say is true, as far as I know. Many active members (some from overseas) of BBSs and other message boards often "smuggle" political articles from these blocked sites, and everyone can then have a discussion about them, at least until they get removed at the request of authorities (about 20% of the time). Generally the participants can agree on the truthfulness of the article in question after some time, and indeed a fairly large percentage of the articles are deemed untrue with evidence to the contrary cited; for other "probably true" articles, we can often have a healthy discussion from them --- at least we get to know how many foreigners think about Tibet or Taiwan, for instance. So you see the blocking isn't really effective except for the most casual surfers who probably won't be interested them anyway, and the leakage of controversial information apparently does more good than harm, for grumpy people would have complaints even if they had seen none of these, and with real sources of information blocked their imagination is often worse than reality.
I don't think Freenet and Entropy are trash... Maybe most of the content there are (I have used them), but the technology and the concept are interesting, if theoretically immature.
As for non-political sites... Yes, it is not the government's intention to block these sites, but it causes inconvenience and blocked access to valuable information, just the same. For political stuff, they at least have a large enough interested audience that much of them are "smuggled" into BBS boards or message boards anyway, even the false and slanderous ones, while if you are interested in some not-so-popular software or some LMM novel reviews located in blocked sites/IP addresses, you are shut out from the information except with a unblocked foreign proxy. How ironic.
Seems that BBC and CNN have been unblocked some time ago. Irritated by so many blocked sites recently (mostly software sites and personal web sites), I just didn't bother rechecking the facts. Sorry.
As a Chinese in China, this is hardly a surprise. Considering that Google news is accessible in China, while most foreign news sites such as CNN and BBC are blocked, I'd be very surprised if Google news are allowed to serve anything censored by the authority to those in China.
Note that I don't think this is right, and the current internet censorship really sucks, neither does it work --- new sites containing western political views spring up every day and they can't censor them one by one. The recent efforts against porn sites are even more laughable, considering that it is still hard to find a news site in China that does not contain sexual content deemed inappropriate for children by most parents. Hopefully some time in the future they will admit that such efforts are useless and use the money on places that really need them (such as some poor rural areas).
On some numerical tests icc seems to give twice faster code than gcc, and I think it is probably due to icc's automatic vectorization (the difference in almabench is said to be mainly due to the superior software SSE-based sin() and cos() functions used by ICC, correct me if this is wrong).
I think the author had better mark the tests whose inner loops have been vectorized, since while some algorithms are easily vectorizable, some are not, so the performance of both are interesting. After all, we care most about the features (such as automatic vectorization) of the compiler, while benchmarks only very roughly reflect the existence of such features, the applicability of them, and their effects.
I guess no library that need any security now has much code containing arbitrarily-sized arrays like "char x[256]" now (unless specified in some standard). Most arrays have a well-defined size that possibly depends on the input, allocating an extra 1024 bytes will not save you if the array should have a size of DSIZE+SSIZE instead of 2*DSIZE, where DSIZE and SSIZE depends on the input, or if the 32-bit array index from the input data is not properly bounds-checked. What's more, the extra size would confuse maintainers as of how much space is really needed.
Even if you use Java or other "safe" languages, such a vulnerability will still result in an array-out-of-bounds exception (hopefully properly handled) in the best case, and it won't be long before you get an OutOfMemoryException or an infinite loop or even worse.
I guess taint checking would be useful here, whatever the language is.
Demand publishers make a new movie/novel/whatever easily available to people in a sufficiently uncrippled format at a reasonable price, and any copying during the time by the people to whom the thing is unavailable will not be counted as infringement, though the publishers are allowed to demand unauthorized copying to stop, and (possibly) ask people who has obtained these copies to either destroy them or buy one at that reasonable price (no fines), when they make the thing available.
Here an "sufficiently uncrippled format" should be a format that allows users to enjoy the work in perpetuity, with no further obligation to the publisher, possibly by using backups and/or software (not applicable if such things are precluded by DRM, patents or whatever). For example, software in ordinary CD-ROMs without timebombs in them is included, so are paper books (you can scan them) and non-crippled music CDs (you can rip them and backup them forever, and you will always be able to play the PCM data). DVDs should also be included, especially when related patents expire and DeCSS is legalized, so that you can rip the bits and play it on the computer anytime in future, when hardware DVD players and DVD-ROMs may be no longer available. In contrast, any time-limited or player-limited versions, such as those using that SVP technology mentioned here, will not count (unless it can be legally hacked), and the publisher had better make it available in some other less-crippled format at the same time. This rule can be loosened for new kinds of copyrightable works for which no such perfect backup mechanisms are available yet, but these should be special cases.
As for a "reasonable" price, I think up to twice the normal price would be acceptable at first, for example up to $40 for a DVD. If the publisher want higher prices, they should make every buyer sign an agreement with them promising that they will not copy the thing they have bought, i.e., it should no longer be of the copyright law's concern.
And if movie publishers want to stop people cameraing their movies and making bootleg copies, they'd better either release the thing in DVD at the same time, or sign an agreement with everyone watching it (no children allowed).
In short, I want to respect your copyright, but if you make your thing public (i.e., not a trade secret or privacy-related stuff), and you don't want to accept my money, you still have no right to prevent me from enjoying it.
Of course, it is not too expensive for my own use, but there are times when I want to exchange data with my classmates (the 'net isn't always usable; we might even meet somewhere without computers around), when it would be convenient to just give him a floppy disk, while a USB drive would be much less convenient, since I probably won't have many spare $30 drives around if the classmate is unable to promptly return the drive to me.
RMS has always maintained that there should be no restrictions placed on using the software, and actually if your license contains such a clause you will have to make the user click through it, since starting a service using your software is allowed by default in copyright law.
As other posts have pointed out, what GLib does is mostly the same; although GLib seems to be used mostly in client-side apps, I have written quite a few server programs quite comfortably with it.
Looking at the API it seems that GLib is a bit easier to use, while APR has some features that may lead to better performance, in particular memory pools (GLib does not have it, so all memory allocation are done by malloc(), saving a parameter but also losing a little flexibility).
Both are quite complete; but from a quick glance I can't find any data structure in APR that implements a sorted set/map, such as balanced binary trees or red-black trees or heaps, which seems to be quite handy when implementing timeouts (you always need to get the timeout that will expire first in order to provide the "timeout" parameter for "select"). Does anyone know how to solve this problem?
I guess there are quite a number of software that use MD5 or SHA-1 hashes as if a collision is impossible, and indeed even when there is code to handle hash collision, it had been hard to test that code when no real collisions were known. Now, if such software is fed with two pieces of data with identical MD5 hashes, it might well break in horrible ways --- at least a DoS attack.
Actually a QC can do simultaneous computations by using superpositioned states, although getting the desired result out can be challenging.
If there is only one correct key, you can use the Grover algorithm to reduce the complexity to O(sqrt(N)). It first make a superposition of all possible N=2^n keys, then for each iteration the encryption algorithm is applied on it (simulataneously to all keys), yielding a "it is the right key or not" answer for each key, still superposed. Then a conditional 180deg-phase-shift operator and a "diffuse" operator is applied on the resulting superposed state, the result being that the eigenstate containing the correct key now has a larger weight than others in the superposed state, so that it has a higher probability to be observed when measured. By repeating the encryption-conditionPhaseShift-diffuse process O(sqrt(N)) times, the eigenstate containing the correct key will have a sufficiently large weight in the superposed state, and we can measure it and get the correct key with a high probability. So you see the encryption operation is always done simultaneously to every key, but you need to repeat the operation O(sqrt(N)) times to make the desired result strong enough to be measured.
Of course, the Grover algorithm is O(sqrt(N)), as opposed to O(N) of a purely brute-force attack, so just double the key length and you will be safe.
To be simple, we suppose the encryption is done by multiplication by 3 (X1=3*P1), then it can be done by adding P1 to itself left-shifted one bit, in which the shifting can be omitted (it doesn't do anything real to the bits) and replaced by some crazy bit-twiddling. After it is done you will see not only X1 but also a copy of P1, and possibly some intermidiate results.
Now it is difficult to get rid of the P1 (i.e. make the bits zero) using only reversible operations. If you don't do that, when mixing X2 and X1 you will also need a correct P2=X2/3 to mix with the P1, before the reverse operation will work. So you see that it isn't easy to reverse things in QC, whether it is multiplying by 3 or the DES transform.
I'm Chinese and in China and can RTFA, and my friends here mostly agree that the google filter works on search results, not keywords (if you search for FaLunGong, you will certainly have some FaLunGong-related sites in the result, so the filter works in this simple case). If the search results contain unapproved sites, it will be filtered however "innocuous" the keyword may be. That's the reason why some keywords seem to be filtered for no reason --- heck, even searching for DengXiaoPing (in Chinese) fails sometimes. Of course, since this is result-dependent, whether or not a keyword seems to be blocks changes randomly with time without Chinese authorities changing anything, and it is IMHO pointless to compile a list of banned keywords.
BTW, the government is cracking down on porn sites by asking ordinary people to help, and this action is quite popular among most people. Personally I'm indifferent to this, since although I don't visit porn sites, such cracking-downs may well extend to sites disagreeing with the Party, some of the information in which can be quite instructive to a Chinese student (even for someone who mostly supports the Party!). The ironic thing is that although hardcore porn sites are not easy to find in China, and porn is indeed illegal, it is still hard to find a news site that porn-paranoid parents can allow their children to roam freely... you need to be really careful not to stumble onto some pornographic material (or things that are obviously very unsuitable for children) on the biggest news sites.
IIRC, with a sufficiently big parabolic antenna we will have an arbitrarily large antenna gain (since the beam is concentrated in one direction only, not spreading everywhere), which means that the signal strength will drop off at the rate of 1/r^2 only for very large r, and it will be much stronger than the signal by an omnidirectional antenna at the same distance. Of course, IIRC to reach 10 times the distance you will need an antenna gain 20dB higher, or the diameter of the antenna must also be 10 times larger. Probably such antennas are just too big to build, and it becomes really hard to point the beam accurately.
Slashdot Mirror
And the problem many have with most software patents (and many in other fields too) is that they cover one of very few good ways of doing something, or the idea of doing some obvious thing itself, so others have no way around it no matter how good they are. Some people might find it fair, but many do not, and find it counterproductive.
For patents/trademarks on smells and tastes, the same thing applies --- I won't object much if and only if the smell/taste must be extremely special (but do we have such a good nose/tongue?), in other words, it must contain a large amount of information.
Actually even if copyrights vanished overnight, the free software community will be one of the least affected. Of course GPL would no longer mean anything, and much troubles may ensue, but life will probably go on in the community. Now that we are talking about patents, a court decision such as the one you mentioned will hardly do any harm to our community.
I'm not a security expert, so some of the above may be wrong, but it is obvious that capturing everything on the wire is not sufficient for successful wiretapping, since the private keys required for deciphering all these information is never sent on the wire (actually they are often never written to disk either), and inferring the public key from the private key is considered practically impossible.
You seemed to be talking about copyrights, while he was talking about patents. They are different.
I think they are wise not to spend time investigating patents. Assuming that all the patents concerned are valid (actually many of them can probably be struck down, but at considerable cost of time and money), the situation may change from unknowningly infringement of 340 patents (57 undiscovered by that MySQL employee) to knowingly infringement of the uncorrectable 97 and unknowningly infringement of the undiscovered 57. Not a better situation IMHO, and a lot of time would have been spent in avoiding patents.
Of course, if you want to use AMD, or want to play newer eye-candy-filled games like Doom3, my advice cannot help you.
Even if he dies the copyright still lasts for 5 to 10 years... the grandparent's way has nothing to do with the author's life. Just make sure his descendents have the ability to publish the work and profit from it.
I don't think Freenet and Entropy are trash... Maybe most of the content there are (I have used them), but the technology and the concept are interesting, if theoretically immature.
As for non-political sites... Yes, it is not the government's intention to block these sites, but it causes inconvenience and blocked access to valuable information, just the same. For political stuff, they at least have a large enough interested audience that much of them are "smuggled" into BBS boards or message boards anyway, even the false and slanderous ones, while if you are interested in some not-so-popular software or some LMM novel reviews located in blocked sites/IP addresses, you are shut out from the information except with a unblocked foreign proxy. How ironic.
Seems that BBC and CNN have been unblocked some time ago. Irritated by so many blocked sites recently (mostly software sites and personal web sites), I just didn't bother rechecking the facts. Sorry.
Note that I don't think this is right, and the current internet censorship really sucks, neither does it work --- new sites containing western political views spring up every day and they can't censor them one by one. The recent efforts against porn sites are even more laughable, considering that it is still hard to find a news site in China that does not contain sexual content deemed inappropriate for children by most parents. Hopefully some time in the future they will admit that such efforts are useless and use the money on places that really need them (such as some poor rural areas).
I think the author had better mark the tests whose inner loops have been vectorized, since while some algorithms are easily vectorizable, some are not, so the performance of both are interesting. After all, we care most about the features (such as automatic vectorization) of the compiler, while benchmarks only very roughly reflect the existence of such features, the applicability of them, and their effects.
Here an "sufficiently uncrippled format" should be a format that allows users to enjoy the work in perpetuity, with no further obligation to the publisher, possibly by using backups and/or software (not applicable if such things are precluded by DRM, patents or whatever). For example, software in ordinary CD-ROMs without timebombs in them is included, so are paper books (you can scan them) and non-crippled music CDs (you can rip them and backup them forever, and you will always be able to play the PCM data). DVDs should also be included, especially when related patents expire and DeCSS is legalized, so that you can rip the bits and play it on the computer anytime in future, when hardware DVD players and DVD-ROMs may be no longer available. In contrast, any time-limited or player-limited versions, such as those using that SVP technology mentioned here, will not count (unless it can be legally hacked), and the publisher had better make it available in some other less-crippled format at the same time. This rule can be loosened for new kinds of copyrightable works for which no such perfect backup mechanisms are available yet, but these should be special cases.
As for a "reasonable" price, I think up to twice the normal price would be acceptable at first, for example up to $40 for a DVD. If the publisher want higher prices, they should make every buyer sign an agreement with them promising that they will not copy the thing they have bought, i.e., it should no longer be of the copyright law's concern.
And if movie publishers want to stop people cameraing their movies and making bootleg copies, they'd better either release the thing in DVD at the same time, or sign an agreement with everyone watching it (no children allowed).
In short, I want to respect your copyright, but if you make your thing public (i.e., not a trade secret or privacy-related stuff), and you don't want to accept my money, you still have no right to prevent me from enjoying it.
Of course, it is not too expensive for my own use, but there are times when I want to exchange data with my classmates (the 'net isn't always usable; we might even meet somewhere without computers around), when it would be convenient to just give him a floppy disk, while a USB drive would be much less convenient, since I probably won't have many spare $30 drives around if the classmate is unable to promptly return the drive to me.
RMS has always maintained that there should be no restrictions placed on using the software, and actually if your license contains such a clause you will have to make the user click through it, since starting a service using your software is allowed by default in copyright law.
Looking at the API it seems that GLib is a bit easier to use, while APR has some features that may lead to better performance, in particular memory pools (GLib does not have it, so all memory allocation are done by malloc(), saving a parameter but also losing a little flexibility).
Both are quite complete; but from a quick glance I can't find any data structure in APR that implements a sorted set/map, such as balanced binary trees or red-black trees or heaps, which seems to be quite handy when implementing timeouts (you always need to get the timeout that will expire first in order to provide the "timeout" parameter for "select"). Does anyone know how to solve this problem?
For best effects, have no licenses slapped onto it, then you may claim that they infringed YOUR copyright!
IANAL, but I believe there may be restrictions on using titles of well-known movies in your product.
N=2^n where n is the key length, so doubling the key length is enough.
Disclaimer: this is a repost.
If there is only one correct key, you can use the Grover algorithm to reduce the complexity to O(sqrt(N)). It first make a superposition of all possible N=2^n keys, then for each iteration the encryption algorithm is applied on it (simulataneously to all keys), yielding a "it is the right key or not" answer for each key, still superposed. Then a conditional 180deg-phase-shift operator and a "diffuse" operator is applied on the resulting superposed state, the result being that the eigenstate containing the correct key now has a larger weight than others in the superposed state, so that it has a higher probability to be observed when measured. By repeating the encryption-conditionPhaseShift-diffuse process O(sqrt(N)) times, the eigenstate containing the correct key will have a sufficiently large weight in the superposed state, and we can measure it and get the correct key with a high probability. So you see the encryption operation is always done simultaneously to every key, but you need to repeat the operation O(sqrt(N)) times to make the desired result strong enough to be measured.
Of course, the Grover algorithm is O(sqrt(N)), as opposed to O(N) of a purely brute-force attack, so just double the key length and you will be safe.
For references, google for "quprog.pdf".
To be simple, we suppose the encryption is done by multiplication by 3 (X1=3*P1), then it can be done by adding P1 to itself left-shifted one bit, in which the shifting can be omitted (it doesn't do anything real to the bits) and replaced by some crazy bit-twiddling. After it is done you will see not only X1 but also a copy of P1, and possibly some intermidiate results.
Now it is difficult to get rid of the P1 (i.e. make the bits zero) using only reversible operations. If you don't do that, when mixing X2 and X1 you will also need a correct P2=X2/3 to mix with the P1, before the reverse operation will work. So you see that it isn't easy to reverse things in QC, whether it is multiplying by 3 or the DES transform.
BTW, the government is cracking down on porn sites by asking ordinary people to help, and this action is quite popular among most people. Personally I'm indifferent to this, since although I don't visit porn sites, such cracking-downs may well extend to sites disagreeing with the Party, some of the information in which can be quite instructive to a Chinese student (even for someone who mostly supports the Party!). The ironic thing is that although hardcore porn sites are not easy to find in China, and porn is indeed illegal, it is still hard to find a news site that porn-paranoid parents can allow their children to roam freely... you need to be really careful not to stumble onto some pornographic material (or things that are obviously very unsuitable for children) on the biggest news sites.
IIRC, with a sufficiently big parabolic antenna we will have an arbitrarily large antenna gain (since the beam is concentrated in one direction only, not spreading everywhere), which means that the signal strength will drop off at the rate of 1/r^2 only for very large r, and it will be much stronger than the signal by an omnidirectional antenna at the same distance. Of course, IIRC to reach 10 times the distance you will need an antenna gain 20dB higher, or the diameter of the antenna must also be 10 times larger. Probably such antennas are just too big to build, and it becomes really hard to point the beam accurately.
It does contain something like MS Access, it is just not a separate application.