I think I've read of people getting Lyx to work with other revision control software (cvs, svn), but I've never actually gotten around to exploring it, so I couldn't say.
I believe the submitter is well aware he could use latex with svn, but his whole point was that he was looking for a way to combine version control with some editing software/process more user friendly than latex.
Though I can use latex, even I prefer using Lyx anyway, as I find I can write equations more quickly with fewer errors than directly in latex. (But then, this might be different if I were willing to add the additional task of mastering some vi or emacs derivative like texmacs or auctex.)
I use LyX to write my LaTeX docs, and it has some support for using version control (using some version control software called RCS). I haven't tried it yet, but I've been tempted.
Thus far, I've been in the position where I just write most of my contribution in Lyx, then export it to plain Latex and sent it to collaborators. From there we just do the collaboration in plain Latex. The problem for me hasn't been the lack of version control but rather the ability/willingness of collaborators to all use LyX. Now, one can import LaTeX into Lyx, but if you do a closed loop (write -> export -> import again) you'll find things are not quite as nice in the end, so this hasn't seemed to be an optimal solution.
As for people saying that technical writers ought to be able to use technical software: A) in many cases it's a question of willingness to commit the time, not ability and B) just because you're technically knowledgeable in, say, cosmological physics, doesn't mean you're adept with computers....trust me on this one.
I think your comment about objective tests is on the money. I do have to say, though, that my Firefox (running for about a two days with, say, 50 tabs open) is currently showing resident memory usage of just under 1 GB (and about 1.5 GB of virtual memory). Now what I don't know about these numbers is how much of that is just caching of pages, etc., that couldn't be harmlessly swapped out if I started running out of memory. I just don't know much about how memory is used in Firefox or in the OS (Linux, in this case).
Is there a viable source for electronic copies of books (i.e., one that has a significant library beyond the contents of project Gutenberg? If I thought I could get non-DRMed electronic versions of most books I'm interested in, I'd buy an ebook reader fast, but for right now it looks like that's not an option. Suffice it to say that I'm not eager to pay $10 a piece for ebooks I can only use with hardware from one vendor.
Well, I was trying to address the issue of POWs, which was not AFAIK the claimed status of those in internment camps. Suffice it to say that that was an absurd and unjust exercise in racism. Not to say that there weren't any Japanese-American sympathizers, but imprisoning every Japanese immigrant was just insane. Again, it failed to properly balance the rights of the various parties.
The government has *always* had a right to intercept foreign shipments and communications.
Well, except for since 1978 when the FISA statute has limited that power in certain ways, which is what has been at issue in the case of Bush administration warrantless wiretaps.
Intelligence collection on non-US Persons outside of the US has never required a warrant, throughout the entire history of the United States.
The difference occurred when traffic of non-US Persons outside of the US started traveling through the US. ...
If you believe that a warrant should be required for intelligence collection on persons outside of the US with no legal standing of any kind with the US (i.e., citizen, vistor, legal resident, etc.), then you are completely out of step with all law, intelligence policy, and scholarship on the issue.
If that were the extent of the issue then there would not be nearly as much disagreement. The issue most often brought up by proponent of recent changes to FISA is that of purely international communications that are simply routed through the US. This is a problem with FISA as it stood before, and anyone who was comfortable with the original FISA would agree that the law should be changed so that the routing is irrelevant. This is precisely why proponents bright this up; it's the strongest argument for a change.
However, the proposed (and the eventually enacted) changes to FISA are much broader, loosening the restrictions on other sorts of eavesdropping. This makes sense, because AFAIK the administration has not stated that the surveillance is purely of this sort. Many of their statements suggest it also includes communications in which one endpoint is within the US. The claim is that the target is not a "US person", but I think in a terrorism investigation this is sort of a tautology, because a terrorist is defined under FISA to be an "agent of a foreign power" and, thus, not a US person (even if they would otherwise be). It also seems that whatever things they wanted to do they the couldn't be entirely sure that no communications of US persons would get caught up in the net.
Since it's secret, we can't be precisely sure what nature is of what they want to do, but there doesn't seem to be good reason to think it's only limited to eavesdropping on international-to-international communications routed through the US. I believe that people arguing for the changes have simply focused on this because it's the strongest argument and somewhat disingenuously ignored that the sought authority is considerably broader.
The job of our foreign intelligence services is to collect information on the activities and plans of US adversaries. This activity has never required a warrant...
Well, except between 1978 and 2007 when the versions of FISA then in effect put limitations on them if interception occurred inside the US. The limitations were certainly not as strict as for domestic law enforcement (and practically non-existent where no "US persons" were involved), but they were there nonetheless and, in all likelihood, illegally violated by Bush administration.
"The most disturbing case, however, is that of Jose Padilla [wikipedia.org], who was never held in Guantanamo, to our knowledge, but is an American citizen arrested in the United States and declared an enemy combatant."
You left out the part where he eventually got his day in court and was convicted by a jury.
He was held incommunicado without charges for more than three years and put into the civilian justice system at the 11th hour to try to prevent the Supreme Court from ruling on the legality of his detention. He was subsequently convicted, but not of anything directly related to the original claims by the government (that he was involved in a plot to make a dirty bomb). So there doesn't seem to be much reason to think that he ever would have seen his day in court were it not for the legal challenges to his detention.
When has the Constitution ever been held to apply to prisoners of war? Do you really mean to tell me that all those German and Japanese POWs we captured could have petitioned for habeas corpus? ...
You are confusing domestic law enforcement with alien combatants captured on foreign battlefields while engaged in combat with US forces.
I'm certainly not scholar of constitutional or military law, but this is my view:
As is, I believe, discussed in the Federalist papers, the framers viewed the bill of rights as a set of limitations of the government that reflect some of the "unalienable rights" of every human being (to borrow the terminology of the Declaration of Independence). Today we would call these the fundamental human rights. Since all people have these rights, according to the framework of natural law on which the framers predicated the constitution, the government has no authority to violate those rights. However, the limitations on what the government may do can be somewhat different when applying to a visitor than when applying to a citizen. Concerns of national security are higher in the case of a visitor, and if he does not like his treatment, he can usually simply leave.
Even enemy soldiers in a war retain their inalienable rights, but as always those rights are balanced against the rights of others and necessity. So POWs retain their right to liberty, but they may be held (in a humane fashion) temporarily until the end of hostilities at which point they should be released. This balances their right to liberty against the safety of citizens in a reasonable way.
In WWII there was a relatively clearly defined theater of war in which we could identify enemy soldiers. It was a conflict with a clearly defined set of groups who would eventually win or lose at which point a treaty would be signed ending hostilities. There was, by and large, no need for habeas corpus challenges by POWs, and given the monumental and existential nature of the conflict it would have been totally impractical to hear such challenges. This was a reasonable balancing of the fundamental rights of all parties and necessity.
The problem with so-called enemy combatants is that a) in many cases it's unclear whether they really were enemy combatants (some were even turned over by third parties, not captured by US forces) and b) there is no clearly defined condition that would end the war in question (they are not being held as POWs in a war against any defined group of powers with any clearly defined geographic or temporal boundaries). If people who may or may not be enemy soldiers are imprisoned indefinitely this clearly is a violation of their fundamental right to liberty, not any sort of reasonable compromise. This is precisely why US courts have found that they must have something "substantively similar" to habeas corpus. If the government can show cause that these people be imprisoned, then balance will be restored. It may be inconvenient to hear these challenges, but we certainly are capable of dealing with it, and we must if we mean to maintain our the commitment to the inalienable rights the framers were trying to protect.
I relate to being upset with what Obama did on the retroactive FISA issue, but I guess I don't understand why this was some major shock. I personally wanted to see a candidate who was reasonable on other policy positions but treated civil liberties as a central issue. I never found one, certainly not Barack Obama. Anyway, I've been over this before, so I won't repeat myself, but suffice it to say I think people blow that incident out of proportion. Regarding your comment on Hillary, I'm quite certain she had not even bothered to show up for votes on the issue on previous occasions and had not taken an anti-immunity provision, so I wouldn't construe that vote as an indicator of any deep conviction either. I think each was doing what they held to be politically pragmatic at the moment, and I think that each made the right choice by that criterion.
If it makes the laptop smaller and lighter, some poeple (myself included) happily will give up a replaceable battery. ...
Just because you don't like a fixed battery doesn't mean "people want it" or it was a bad decision.
If it were some PC manufacturer we were talking about, I think you'd be completely correct. They've decided on something that they feel is most marketable, and if it doesn't suit you then you ought to buy from somewhere else. Unfortunately, since Macs use a proprietary OS that is (supposed to be) limited to Mac hardware, there's a significant lock-in factor. If people want to buy a computer from a different manufacturer, they will have to start using a new OS, which means learning that OS, buying and learning new software, etc. So, while I'm sure you're right that surely Apple has done the research and figured out this won't impact sales negatively overall, if you happen to be one of the Mac users for which this runs against your preferences or usage patterns you're simply SOL. You either have to like what Apple tells you to like, or you have to incur big pain the the butt (and very likely significant costs) of switching OSs.
I think that Apple makes great hardware and software in many instances, but that is one of the main reasons I probably won't ever buy any of it.
My parents' Dish Network receiver allows them to setup a "favorites" list so that some channels are never shown in the channel guide. That's effectively as good.
The other thing that may be worth pointing out is that without some sort of repercussions for misleading people, that's probably always where the market will drive things. Clearly if all your competitors say they offer "unlimited" connections you're unlikely to be able to compete while honestly explaining the caps on your service. Well, maybe amongst geeks who really care, but not amongst the general population.
The other sticky issue is that if you do explain what you're actually offering, many people may not really understand it. To add to my other story, I was recently at the mall and there was a booth advertising a new Sprint WiMAX service called XOHM. I tried to ask the guy what the latency of the connection was, but he didn't understand the difference between that and bandwidth. And he was selling the service. So, I'm not confident that the average customer will understand the difference between a data cap and a data rate cap.:-/
But when traffic is being routed through my home system - when I am the "upstream provider" - there are questions I need answered:
1 What is my legal exposure?
I am betting I do not have protection as a common carrier.
You are betting correctly, because no ISP has common carrier status. This is a common myth that is unfortunately widely propagated on Slashdot. Due to the safe harbor provision of of the DMCA and some other laws, ISPs do enjoy some similar protections (see, for example here and here for more details).
As for the remainder of the questions, they are valid ones and I'm not certain of the answers. Personally, I'd be willing to chance any potential exposure (which I believe would be limited but I am not certain of that), but YMMV.
Unfortunately, pretty much all ISPs (in my experience) offer "unlimited" service, but for residential broadband connections it's always a lie. AFAIK, for the prices they're charging, it would always have to be a lie for them to turn a profit. So, on one level it's probably unreasonable to expect they'll let you max out your bandwidth 24/7. The second point is that if you're getting a connection with any ISP you should always read the Terms of Service (TOS) (at least if you're going to complain later that they violated their deal). If you look at Speakeasy's TOS you'll see that they have weasel words in there prohibiting use of "excessive bandwidth". As far as what you may have been promised beforehand, I don't know how much of that reflects systematic behavior of Speakeasy and how much was just that the individual you talked to was an idiot/lying.
All that being said, Speakeasy, like pretty much all other ISPs has often advertised "unlimited" connections of a certain bandwidth and it's dishonest if that's not actually what they provide (even if a caveat is buried on page 43 or whatever of the TOS). It also doesn't make any sense to provide people with limited Internet service without telling them what sorts of caps they're actually agreeing to. I would not claim that Speakeasy is perfect, but from my own experience and those of people I've talked to they are the best of a set of imperfect alternatives.
Interestingly, we've recently had Verizon giving us the hard sell on getting FIOS. The TOS for their residential connection prohibits running a server of any sort and "excessive" use of bandwidth (not that I have any idea what constitutes "excessive" on a 15/2 Mbps connection). When I mentioned this to the guy trying to sell me the service he said he was "surprised" to hear that's in the TOS and that he had been selling the service to people who wanted to run servers (he seemed to be trying to imply it wasn't true, but you can read for yourself). I guess that outlines the degree of duplicity of ISPs.
I think that's true in many cases. Like I said in reply to the parent, that's why I think the Netshare program my ISP (Speakeasy) runs is cool. The only problems are that a) Speakeasy isn't cheap and b) you then have to get your neighbors interested/convince them it's not some kind of scam.
Interestingly, if your neighbors had their broadband through Speakeasy you could be doing this on the up and up through their Netshare program, because it some cases this probably is a practical solution.
Unfortunately, innovative offerings come through competition, and with high speed broadband moving increasingly toward cable and phone companies' offerings (i.e., cable and fiber) you'll be dealing with local monopolies who have no real incentive to offer you those sorts of options.
I'm not sure the number of possible outcomes has anything to do with it. These quantum algorithms can in principle give the same wrong answer over and over again (though this is extremely improbable), so it's not like you exhaust the set of wrong answers or something. In any case, it is certainly true that you can put a probability distribution on the time it will take the algorithm to get the answer, and you ask how long it will take to get a solution on average or that it will take less than T operations 99.999% of the time. Clearly, if you're comparing a polynomial to an exponential time growth, at any level of confidence the quantum polynomial algorithm will win asymptotically.
But that's all beside the point, because my question is really more about the classical problem. By my (admittedly not thorough) reading, it looks like all they're requiring of the quantum algorithm is to find the solution vector up to an relative error epsilon in the 2-norm. (i.e., it appears epsilon is a distance in the vector space, not a probability) To wit:
we can approximately construct |x> = A^1 |b>... ...then our residual state (up to relative error O(t_1^2 kappa^2 )) is A^-1 |b> = |x> . Taking t1 much less than 1/kappa gives a good success probability.
What I'm wondering is whether similarly relaxing the requirements on the classical computation (so that you only require such an approximate solution in the same sense, not an exact solution) will reduce the classical computational complexity of the problem significantly.
It's true that something like Shor's algorithm is probabilistic, and so, yes, you're always sort of comparing apples to oranges. So in that situation you can define a probability distribution on the amount of time it will take to get the correct answer. To compare apples to apples you simply compare some number like the mean time for solution, or the 95th percentile longest time for solution (there's also the question of best- and worst-case scenarios) for the two algorithms. If the classical algorithm is deterministic, it's probability distribution is just simpler. I assume that in the case of factoring there's no faster non-deterministic classical algorithm so that it's a moot point.
In this case it looks like they're talking about solving the system to within accuracy epsilon and comparing that to solving exactly, which seems a bit different than the question of solving exactly but sometimes being wrong. Mostly, in this case you're "wrong" most of the time but only a little wrong, and you don't go back and correct it. So my question is whether in this case there is a faster classical algorithm if we let it also get the answer only to within epsilon; this is the classical algorithm we would wish to compare to.
So, I think you're right that you're always sort of comparing apples to oranges, but it seems like here we're doing so to an unnecessary degree. In all likelihood, relaxing the condition on the classical solutions doesn't admit any known faster solution, and the authors just assumed the reader knows this. But I don't know it, and I am hoping someone on/. does.
This looks like a really interesting result, but having look a little at the paper it's not 100% clear to me what the quantum algorithm is being compared to. First a caveat, I study quantum physics but not CS, so my knowledge of algorithms and complexity theory is quite limited. Anyway, in solving problems on a good old classical computer, you can seek to solve a linear system exactly (up to the bounds of finite arithmetic) or you can seek to solve it approximately to within some error.
So the thing I'm wondering is what classical algorithm are they comparing their result to, and is this comparing apples to apples? They say
In this case, when A is sparse and well-conditioned, with
largest dimension n, the best classical algorithms can find x and estimate x* M x
in O(n) time. Here, we exhibit a quantum algorithm for solving linear sets of
equations that runs in O(log n) time, an exponential improvement over the best
classical algorithm.... In particular, we can approximately
construct |x> = A^-1 |b> and make a measurement M whose expectation value x* M x
corresponds to the feature of x that we wish to evaluate.
So it looks like the authors' algorithm gives an approximate solution to the linear system and they're comparing it to a classical algorithm that gives an exact solution to the problem. This seems like comparing apples to oranges. Perhaps someone who knows more about the features of the various classical algorithms can comment on whether it looks like the correct comparison to make and why.
I bring this up because I recall that given a matrix representing the density matrix of a bipartite quantum system, determining whether it represents an entangled or separable quantum state is in general NP-Hard, but IIRC there exist semi-definite programming techniques to get the answer probabilistically in polynomial time. The point is that in that case there's a big gain for accepting an answer that will be wrong every once in a while. I was just curious if settling for an approximate answer in solving linear systems changes the complexity of that problem drastically as well.
Slashdot Mirror
I think I've read of people getting Lyx to work with other revision control software (cvs, svn), but I've never actually gotten around to exploring it, so I couldn't say.
I believe the submitter is well aware he could use latex with svn, but his whole point was that he was looking for a way to combine version control with some editing software/process more user friendly than latex.
Though I can use latex, even I prefer using Lyx anyway, as I find I can write equations more quickly with fewer errors than directly in latex. (But then, this might be different if I were willing to add the additional task of mastering some vi or emacs derivative like texmacs or auctex.)
I use LyX to write my LaTeX docs, and it has some support for using version control (using some version control software called RCS). I haven't tried it yet, but I've been tempted.
Thus far, I've been in the position where I just write most of my contribution in Lyx, then export it to plain Latex and sent it to collaborators. From there we just do the collaboration in plain Latex. The problem for me hasn't been the lack of version control but rather the ability/willingness of collaborators to all use LyX. Now, one can import LaTeX into Lyx, but if you do a closed loop (write -> export -> import again) you'll find things are not quite as nice in the end, so this hasn't seemed to be an optimal solution.
As for people saying that technical writers ought to be able to use technical software: A) in many cases it's a question of willingness to commit the time, not ability and B) just because you're technically knowledgeable in, say, cosmological physics, doesn't mean you're adept with computers. ...trust me on this one.
I think your comment about objective tests is on the money. I do have to say, though, that my Firefox (running for about a two days with, say, 50 tabs open) is currently showing resident memory usage of just under 1 GB (and about 1.5 GB of virtual memory). Now what I don't know about these numbers is how much of that is just caching of pages, etc., that couldn't be harmlessly swapped out if I started running out of memory. I just don't know much about how memory is used in Firefox or in the OS (Linux, in this case).
Is there a viable source for electronic copies of books (i.e., one that has a significant library beyond the contents of project Gutenberg? If I thought I could get non-DRMed electronic versions of most books I'm interested in, I'd buy an ebook reader fast, but for right now it looks like that's not an option. Suffice it to say that I'm not eager to pay $10 a piece for ebooks I can only use with hardware from one vendor.
What you're saying sounds quite plausible, but it's be useful if you actually pointed out some references.
Well, I was trying to address the issue of POWs, which was not AFAIK the claimed status of those in internment camps. Suffice it to say that that was an absurd and unjust exercise in racism. Not to say that there weren't any Japanese-American sympathizers, but imprisoning every Japanese immigrant was just insane. Again, it failed to properly balance the rights of the various parties.
Well, except for since 1978 when the FISA statute has limited that power in certain ways, which is what has been at issue in the case of Bush administration warrantless wiretaps.
If that were the extent of the issue then there would not be nearly as much disagreement. The issue most often brought up by proponent of recent changes to FISA is that of purely international communications that are simply routed through the US. This is a problem with FISA as it stood before, and anyone who was comfortable with the original FISA would agree that the law should be changed so that the routing is irrelevant. This is precisely why proponents bright this up; it's the strongest argument for a change.
However, the proposed (and the eventually enacted) changes to FISA are much broader, loosening the restrictions on other sorts of eavesdropping. This makes sense, because AFAIK the administration has not stated that the surveillance is purely of this sort. Many of their statements suggest it also includes communications in which one endpoint is within the US. The claim is that the target is not a "US person", but I think in a terrorism investigation this is sort of a tautology, because a terrorist is defined under FISA to be an "agent of a foreign power" and, thus, not a US person (even if they would otherwise be). It also seems that whatever things they wanted to do they the couldn't be entirely sure that no communications of US persons would get caught up in the net.
Since it's secret, we can't be precisely sure what nature is of what they want to do, but there doesn't seem to be good reason to think it's only limited to eavesdropping on international-to-international communications routed through the US. I believe that people arguing for the changes have simply focused on this because it's the strongest argument and somewhat disingenuously ignored that the sought authority is considerably broader.
Well, except between 1978 and 2007 when the versions of FISA then in effect put limitations on them if interception occurred inside the US. The limitations were certainly not as strict as for domestic law enforcement (and practically non-existent where no "US persons" were involved), but they were there nonetheless and, in all likelihood, illegally violated by Bush administration.
He was held incommunicado without charges for more than three years and put into the civilian justice system at the 11th hour to try to prevent the Supreme Court from ruling on the legality of his detention. He was subsequently convicted, but not of anything directly related to the original claims by the government (that he was involved in a plot to make a dirty bomb). So there doesn't seem to be much reason to think that he ever would have seen his day in court were it not for the legal challenges to his detention.
I'm certainly not scholar of constitutional or military law, but this is my view:
As is, I believe, discussed in the Federalist papers, the framers viewed the bill of rights as a set of limitations of the government that reflect some of the "unalienable rights" of every human being (to borrow the terminology of the Declaration of Independence). Today we would call these the fundamental human rights. Since all people have these rights, according to the framework of natural law on which the framers predicated the constitution, the government has no authority to violate those rights. However, the limitations on what the government may do can be somewhat different when applying to a visitor than when applying to a citizen. Concerns of national security are higher in the case of a visitor, and if he does not like his treatment, he can usually simply leave.
Even enemy soldiers in a war retain their inalienable rights, but as always those rights are balanced against the rights of others and necessity. So POWs retain their right to liberty, but they may be held (in a humane fashion) temporarily until the end of hostilities at which point they should be released. This balances their right to liberty against the safety of citizens in a reasonable way.
In WWII there was a relatively clearly defined theater of war in which we could identify enemy soldiers. It was a conflict with a clearly defined set of groups who would eventually win or lose at which point a treaty would be signed ending hostilities. There was, by and large, no need for habeas corpus challenges by POWs, and given the monumental and existential nature of the conflict it would have been totally impractical to hear such challenges. This was a reasonable balancing of the fundamental rights of all parties and necessity.
The problem with so-called enemy combatants is that a) in many cases it's unclear whether they really were enemy combatants (some were even turned over by third parties, not captured by US forces) and b) there is no clearly defined condition that would end the war in question (they are not being held as POWs in a war against any defined group of powers with any clearly defined geographic or temporal boundaries). If people who may or may not be enemy soldiers are imprisoned indefinitely this clearly is a violation of their fundamental right to liberty, not any sort of reasonable compromise. This is precisely why US courts have found that they must have something "substantively similar" to habeas corpus. If the government can show cause that these people be imprisoned, then balance will be restored. It may be inconvenient to hear these challenges, but we certainly are capable of dealing with it, and we must if we mean to maintain our the commitment to the inalienable rights the framers were trying to protect.
This can't properly refer to Coulter; One of the defining characteristics of mammals is that they are warm blooded.
I relate to being upset with what Obama did on the retroactive FISA issue, but I guess I don't understand why this was some major shock. I personally wanted to see a candidate who was reasonable on other policy positions but treated civil liberties as a central issue. I never found one, certainly not Barack Obama. Anyway, I've been over this before, so I won't repeat myself, but suffice it to say I think people blow that incident out of proportion. Regarding your comment on Hillary, I'm quite certain she had not even bothered to show up for votes on the issue on previous occasions and had not taken an anti-immunity provision, so I wouldn't construe that vote as an indicator of any deep conviction either. I think each was doing what they held to be politically pragmatic at the moment, and I think that each made the right choice by that criterion.
If it were some PC manufacturer we were talking about, I think you'd be completely correct. They've decided on something that they feel is most marketable, and if it doesn't suit you then you ought to buy from somewhere else. Unfortunately, since Macs use a proprietary OS that is (supposed to be) limited to Mac hardware, there's a significant lock-in factor. If people want to buy a computer from a different manufacturer, they will have to start using a new OS, which means learning that OS, buying and learning new software, etc. So, while I'm sure you're right that surely Apple has done the research and figured out this won't impact sales negatively overall, if you happen to be one of the Mac users for which this runs against your preferences or usage patterns you're simply SOL. You either have to like what Apple tells you to like, or you have to incur big pain the the butt (and very likely significant costs) of switching OSs.
I think that Apple makes great hardware and software in many instances, but that is one of the main reasons I probably won't ever buy any of it.
My parents' Dish Network receiver allows them to setup a "favorites" list so that some channels are never shown in the channel guide. That's effectively as good.
Yes, especially considering the poor grades the DOD and DHS were given on computer security, this doesn't seem too serious.
The other thing that may be worth pointing out is that without some sort of repercussions for misleading people, that's probably always where the market will drive things. Clearly if all your competitors say they offer "unlimited" connections you're unlikely to be able to compete while honestly explaining the caps on your service. Well, maybe amongst geeks who really care, but not amongst the general population.
The other sticky issue is that if you do explain what you're actually offering, many people may not really understand it. To add to my other story, I was recently at the mall and there was a booth advertising a new Sprint WiMAX service called XOHM. I tried to ask the guy what the latency of the connection was, but he didn't understand the difference between that and bandwidth. And he was selling the service. So, I'm not confident that the average customer will understand the difference between a data cap and a data rate cap. :-/
You are betting correctly, because no ISP has common carrier status. This is a common myth that is unfortunately widely propagated on Slashdot. Due to the safe harbor provision of of the DMCA and some other laws, ISPs do enjoy some similar protections (see, for example here and here for more details).
As for the remainder of the questions, they are valid ones and I'm not certain of the answers. Personally, I'd be willing to chance any potential exposure (which I believe would be limited but I am not certain of that), but YMMV.
Unfortunately, pretty much all ISPs (in my experience) offer "unlimited" service, but for residential broadband connections it's always a lie. AFAIK, for the prices they're charging, it would always have to be a lie for them to turn a profit. So, on one level it's probably unreasonable to expect they'll let you max out your bandwidth 24/7. The second point is that if you're getting a connection with any ISP you should always read the Terms of Service (TOS) (at least if you're going to complain later that they violated their deal). If you look at Speakeasy's TOS you'll see that they have weasel words in there prohibiting use of "excessive bandwidth". As far as what you may have been promised beforehand, I don't know how much of that reflects systematic behavior of Speakeasy and how much was just that the individual you talked to was an idiot/lying.
All that being said, Speakeasy, like pretty much all other ISPs has often advertised "unlimited" connections of a certain bandwidth and it's dishonest if that's not actually what they provide (even if a caveat is buried on page 43 or whatever of the TOS). It also doesn't make any sense to provide people with limited Internet service without telling them what sorts of caps they're actually agreeing to. I would not claim that Speakeasy is perfect, but from my own experience and those of people I've talked to they are the best of a set of imperfect alternatives.
Interestingly, we've recently had Verizon giving us the hard sell on getting FIOS. The TOS for their residential connection prohibits running a server of any sort and "excessive" use of bandwidth (not that I have any idea what constitutes "excessive" on a 15/2 Mbps connection). When I mentioned this to the guy trying to sell me the service he said he was "surprised" to hear that's in the TOS and that he had been selling the service to people who wanted to run servers (he seemed to be trying to imply it wasn't true, but you can read for yourself). I guess that outlines the degree of duplicity of ISPs.
I think that's true in many cases. Like I said in reply to the parent, that's why I think the Netshare program my ISP (Speakeasy) runs is cool. The only problems are that a) Speakeasy isn't cheap and b) you then have to get your neighbors interested/convince them it's not some kind of scam.
Interestingly, if your neighbors had their broadband through Speakeasy you could be doing this on the up and up through their Netshare program, because it some cases this probably is a practical solution.
Unfortunately, innovative offerings come through competition, and with high speed broadband moving increasingly toward cable and phone companies' offerings (i.e., cable and fiber) you'll be dealing with local monopolies who have no real incentive to offer you those sorts of options.
I'm not sure the number of possible outcomes has anything to do with it. These quantum algorithms can in principle give the same wrong answer over and over again (though this is extremely improbable), so it's not like you exhaust the set of wrong answers or something. In any case, it is certainly true that you can put a probability distribution on the time it will take the algorithm to get the answer, and you ask how long it will take to get a solution on average or that it will take less than T operations 99.999% of the time. Clearly, if you're comparing a polynomial to an exponential time growth, at any level of confidence the quantum polynomial algorithm will win asymptotically.
But that's all beside the point, because my question is really more about the classical problem. By my (admittedly not thorough) reading, it looks like all they're requiring of the quantum algorithm is to find the solution vector up to an relative error epsilon in the 2-norm. (i.e., it appears epsilon is a distance in the vector space, not a probability) To wit:
What I'm wondering is whether similarly relaxing the requirements on the classical computation (so that you only require such an approximate solution in the same sense, not an exact solution) will reduce the classical computational complexity of the problem significantly.
It's true that something like Shor's algorithm is probabilistic, and so, yes, you're always sort of comparing apples to oranges. So in that situation you can define a probability distribution on the amount of time it will take to get the correct answer. To compare apples to apples you simply compare some number like the mean time for solution, or the 95th percentile longest time for solution (there's also the question of best- and worst-case scenarios) for the two algorithms. If the classical algorithm is deterministic, it's probability distribution is just simpler. I assume that in the case of factoring there's no faster non-deterministic classical algorithm so that it's a moot point.
In this case it looks like they're talking about solving the system to within accuracy epsilon and comparing that to solving exactly, which seems a bit different than the question of solving exactly but sometimes being wrong. Mostly, in this case you're "wrong" most of the time but only a little wrong, and you don't go back and correct it. So my question is whether in this case there is a faster classical algorithm if we let it also get the answer only to within epsilon; this is the classical algorithm we would wish to compare to.
So, I think you're right that you're always sort of comparing apples to oranges, but it seems like here we're doing so to an unnecessary degree. In all likelihood, relaxing the condition on the classical solutions doesn't admit any known faster solution, and the authors just assumed the reader knows this. But I don't know it, and I am hoping someone on /. does.
This looks like a really interesting result, but having look a little at the paper it's not 100% clear to me what the quantum algorithm is being compared to. First a caveat, I study quantum physics but not CS, so my knowledge of algorithms and complexity theory is quite limited. Anyway, in solving problems on a good old classical computer, you can seek to solve a linear system exactly (up to the bounds of finite arithmetic) or you can seek to solve it approximately to within some error.
So the thing I'm wondering is what classical algorithm are they comparing their result to, and is this comparing apples to apples? They say
So it looks like the authors' algorithm gives an approximate solution to the linear system and they're comparing it to a classical algorithm that gives an exact solution to the problem. This seems like comparing apples to oranges. Perhaps someone who knows more about the features of the various classical algorithms can comment on whether it looks like the correct comparison to make and why.
I bring this up because I recall that given a matrix representing the density matrix of a bipartite quantum system, determining whether it represents an entangled or separable quantum state is in general NP-Hard, but IIRC there exist semi-definite programming techniques to get the answer probabilistically in polynomial time. The point is that in that case there's a big gain for accepting an answer that will be wrong every once in a while. I was just curious if settling for an approximate answer in solving linear systems changes the complexity of that problem drastically as well.