At one level, there's still a lot of naivety. This is true. Naivete is surprisingly hard to kill; call them fools or optimists, but a lot of people seem to love to hold onto unrealistic expectations of others far beyond what is rational or predictable.
I think this is one of the main reasons why so much security policy is reactive rather than proactive. Nobody wants to be the person to call out everyone else for being potential criminals, even though everyone rationally knows that it's true.
Were these counties all named things like Microsoft-land, Microsoft-world, Microsoftia and so on? No, but some of them were countries that probably had bigger issues than ODF versus OOXML, like say feeding themselves. It was pretty clear that some of them were in it for the cold, hard cash, and couldn't give a crap about what they were voting on.
Maybe they could make voting membership in a computer-standards committee contingent on having some sort of viable technology industry or something. (Of course, in a few decades that would probably knock out the United States, the way we're going...)
Their bylaws probably prevent them from doing this except by a vote of all the P-class members.
I've seen this sort of thing happen before, to smaller organizations. You get a huge influx of members for some reason, but then they stop participating. If you didn't anticipate this possibility when drafting your constitution or bylaws, and you have some rule in there that says "changes to the bylaws must be ratified by 50% of the membership" or something similar, you're screwed. You can't change the rules, because nobody shows up, and you can't do anything, because nobody shows up.
Maybe the ISO Standards Committee should dissolve itself and reform under a slightly different name, with a better set of bylaws...
We think of sunspots as following a fixed, 11-year cycle, but this may only be one part of the story.
I don't think the 'experts' necessarily know anything more at this point, either; just a few years ago, NASA was predicting that the next cycle would be the strongest ever, and that got a lot of people (especially folks that do a lot of shortwave/HF radio) very excited. Now, it looks like we may have a very small cycle, or no cycle at all -- it's anybody's guess.
The dead spot on some sunspot charts from 1650-1700 is called the "Maunder Minimum". During that period, rather than talking about sunspots, observers of the day would write about the appearance of a particular sunspot (very much singular!). Unfortunately, the data prior to the beginning of the minimum is pretty sparse, and exactly when it started is under some dispute.
There was also another minimum in the early 19th century, called the Dalton Minimum, although it wasn't as severe and it only lasted about 25 years.
So that's two minima separated by a 150-year gap. But at 150 years after the 1800 minimum, rather than another minimum, we actually get a maximum in 1950. There's just not enough historical data to make a good prediction, because we don't know how complex the cycle is. But it's clearly more complex than just 11 years.
I can't find a link to it online, but I heard a talk recently about a group that was using geological evidence to try and track the sunspot cycle further back than we have human observations. Not sure quite what the method is, or if it's yielded any results. But that would certainly be interesting, if you could get some real historical perspective instead of the piddling 7 centuries (at most) that you can find written records of. That might give us some idea of what's been going on, on very long timescales, as well as perhaps filling in the gaps in the historical record in more recent times (not sure what kind of resolution you can get).
To use a water analogy, the 11-year cycles might be waves lapping at the shore, but there might be scores of other forces acting on them at higher levels, like tides, wind, and the seasons, all on vastly different time-scales.
All in all, for something that we spend the majority of our waking lives under, our understanding of the sun is surprisingly poor. Particularly given how much modern technology (radio communications is the obvious one, but there are others) can be affected by the solar cycle, it seems to be ignored until it does something unexpected.
I don't think you can connect to it directly using an IP address, since it uses the domain portion of the URL to figure out what page you're requesting. Although all.nyud.net sites resolve to 132.239.17.225 for me (which happens to be the node in Princeton), you can't just connect to that and get the cached page.
If you're still blocked you'll need to tunnel out to a proxy.
Humm. It seems like the $69 price for Leopard (desktop, single-seat) is only available at some EDUs. Lots of other people are going in and seeing significantly higher prices.
I wonder what the basis for the difference is? Apparently some institutions are a little more equal than others. I wonder if it represents those that have big Apple site-licenses or something?
A geek working at the Apple store here told me that they will be dropping an upgrade DVD into the box for MACs from the time that Leopard is released up until the units begin arriving with Leopard preloaded. This makes sense and is consistent with what Apple has done previously. Either they will toss a retail disc into the box of units sold in the transitional period (between release but before they ship with the new OS), or there will be an upgrade program where you can get a retail disc using a proof-of-purchase coupon and the cost of shipping.
Basically, this is just because Apple doesn't want people to put off purchasing computers once they release the new OS, because they're waiting on ones that come with it preloaded. (I suspect that people do wait anyway, but it's their effort to mitigate this.)
It's also helped that each release has felt faster, so buying a new copy of OSX also replaced a hardware boost I typically underwent with Windows updates. Agreed. This is the thing that concerns me about Leopard, though. I'm a little worried that since they've brought out so many hardware performance improvements, that it will allow more bloat to sneak in than the past few years have.
Up until last year I was running OS X on a 400MHz G4; it got progressively faster with each release and was still totally usable. (It's still in use, too; I just had to give it to a family member whose computer died rather suddenly and spectacularly, giving me an opportunity to buy myself a nice present.) Shelling out a few hundred bucks every couple of years was totally worth it to keep getting more performance out of a 1999 computer.
Some of the features in Leopard are neat, but I don't see anything that would justify any type of performance hit. So I'm waiting on upgrading until I hear whether this release keeps up the gains in "teh snappy" that I've gotten used to.
I'll just get Leopard with a new Mac sometime in the distant future and put it on my older machines then. This won't work. Although Apple doesn't do serialization or verification, the discs that come with a computer are different from the retail box versions of the OS. They're not the crummy 'software restore' discs like you get with some PCs -- they do have a regular OS installer on them -- but the installer is fixed so that it looks for the machine ID and refuses to run on a different model computer.
The retail versions, by contrast, will run on any machine that's listed as capable of running the software. (Which sometimes is slightly different than the machines that are *actually* capable of running the software; Apple specs systems that are capable of running the OS comfortably, but some people have found acceptable results after forcing it onto older machines.)
If you wait around until the next paid-upgrade OS release though, you can get the older version, in retail packaging, quite cheap. Either eBay or some of the used-Mac stores like Smalldog regularly have new-old-stock retail OS packages.
You do realize that 298 of those 1195 SEK are tax, right? So subtracting that out, you get a real price of 897 SEK, which is only 68 SEK more than the US price, or about $10.60 USD.
I doubt that you'd be able to order a US version and have it shipped to Sweden for less than $10 in shipping.
Seems like a pretty fair price to me. Maybe you should vote for politicians who support lower taxes if you don't like it?
Thanks for playing, though. I don't think --link-dest does binary diffs/deltas, though, does it? If you modify a large file, it will create new copies of them rather than keeping track of the deltas. While this might be a good thing from a reliability standpoint, I could imagine that it could quickly cause your backups to get quite large.
I can't quite figure out how Time Machine works, but I think its operation is closer to rdiff-backup than rsync --link-dest. If it's not, the space requirements for it are going to be huge.
Depending on how you're using Ekiga (seriously, could they have found a more difficult to pronounce name?) you could still be sending your traffic through a CALEA-compliant (and thus snoopable) network.
Ekiga is just the client program that runs on your computer; it is to VoIP what Firefox is to the WWW. The client program isn't (generally) where tapping occurs; law enforcement does that in the network, where it's harder to detect. So the question isn't whether you use Ekiga, it's who do you use for an ISP, and who do you use for SIP-to-POTS service?
If you use one of the major US ISP's, you have to look no further than them for the authorities' way into your traffic, since they can provide ways to tap the SIP traffic directly if it's unencrypted (and I don't think most SIP implementations support encryption, more shame on them). It's not terribly hard to pick out the right packets if you're sniffing the connection. So even if you're calling from your computer to your friend's computer completely via SIP, never going near the POTS network, you could still be snooped.
If you use a SIP-to-POTS gateway service (allowing you to call landline numbers from your SIP phone), that provider -- if it's based in the US -- almost certainly is CALEA-compliant and will provide a way to tap. I'm fairly certain that whoever does Ekiga.net's gateway service (Diamondcard.us?) is either CALEA-compliant themselves, or whoever they buy POTS circuits from is.
To be frank, it's a fool's errand to try and avoid CALEA by switching providers. If you want to talk with people in the U.S., your traffic is going to go through a network that's CALEA-compliant and can be snooped on. It may be marginally more difficult right now for authorities to snoop on a completely IP-based, SIP conversation than on a POTS one, but this will probably change as law enforcement becomes more comfortable with the technology.
The solution, IMO, for a person desiring private communication, is not to rely on the security of the data channel, but to create that security using encryption for the conversation itself. Zfone is Phil Zimmerman's modern update to PGPfone, and works in conjunction with most SIP clients at the protocol stack level to encrypt the SIP stream. It looks pretty slick, although I haven't played with it much myself. There seem to be versions available for Mac, Linux, and Windows. Sadly, the code is not GPL, so it will probably never appear in mainstream Linux distributions.
The rule is one rifle (AK-47 or similar) per household for protection, no heavy weapons, explosives, or caches of weapons.
The military isn't so stupid as to ban civilian ownership of all weapons; it would just make the population more exposed -- not just to foreign hostiles, but also to sectarian violence, and the usual criminal elements -- rather than safer.
This won't work for the reasons that other people have noted.
The best security precaution is continual awareness. If you're intimately familiar with all of your hardware and software, it's a lot harder for someone to install a keylogger. Would you know if someone came into your office and moved something around? You should. It requires an effort, though, to start paying attention to little things, so that you'll notice if something is amiss. And if you have a bad feeling, you need to act on it immediately.
Would you notice if someone swapped your keyboard with one of an identical make and model and approximate age? And if you did notice something odd -- maybe a little stiffness in the keys that wasn't there before, a difference in the wear patterns from where your fingers normally lie -- would you just shrug it off or would you immediately stop using it? How often do you actually look behind your desk to see if someone has shoved one of these in between your keyboard and CPU? Those are the things you have to take into consideration.
It's similar with software. A while back I read about a guy who only discovered he'd been rooted because of an oddly misbehaving "ls" command when it was invoked with certain switches. Lots of other intrusions are only discovered because of similar, very subtle, signs. (Most of which boil down to the intruder making a mistake somewhere.)
Most people don't want to have to pay attention to security, and thus look for easy ways out. This is generally where they become most vulnerable. Automated and procedural security is good, but ultimately any 'fire and forget' approach is fatally flawed. There's no replacement for vigilance.
If you have OS X 10.4, you can make your own certificates. Yes, you can do this. However, it's a pretty poor idea.
S/MIME is designed to work with centralized Certificate Authorities. If you roll your own CA and issue yourself a self-signed certificate, you'll be able to sign stuff, but people who receive your messages will get a big "BAD SIGNATURE" error or warning, because they won't have your CA in their trusted chain. In order to get it to work, you'd need to get them the CA certificate, and they'd need to import it into their trusted root database. (Which is a security risk -- you do not want to encourage clueless users to start importing certs from every idiot they want to talk to into their Trusted Root.)
It is much better to just get a personal certificate from Thawte or several of the other places online that give them out. Thawte is aimed at people who want authenticated communication; it's not anonymous and in fact they require some form of Government ID in order to issue one. If you want to use S/MIME anonymously or pseudonymously, you're better off going to OpenCA and getting one through them. (Their CA cert isn't included by default in most browsers and OSes like Thawte's is, but at least your correspondents only need to import one additional certificate to recognize yours, and it comes from a basically legitimate institution. That's a lot better than importing random people's CA certs into your root DB.)
Yes, but the US pricing structure also doesn't distinguish between the cost of making and receiving a call. Right. You pay for the circuit over the cellular network. The circuit needs to be opened regardless of which direction the call is going in.
By your logic, receiving a call should be at least marginally less expensive. The bulk of the cost of a cellular telephone call is the transit over the cellular network. Using the POTS network is cheap. Cellphone companies negotiate rates with the POTS network that are so low, compared to the cellular airtime, that it's not broken out per-call. Should it be slightly more expensive to make an outgoing call from your cellphone than receive an incoming one? Sure. (And at one point it was; very early cellphone plans didn't include long-distance...so if you called California from your Massachusetts cellphone, you'd get charged airtime plus a fee for the long-distance landline call.) But POTS time is cheap, and the cellphone companies basically just build the cost to transfer the outgoing calls over the POTS network to anyplace in the country into your bill as part of the base rate.
I also don't like having to pay for the "privilege" of paying for my boss to call me while I'm at home. If you want to get in touch with me, you should pay for it! You're the one with the cellphone. Your boss doesn't know when he dials that number that it's a mobile and not a fixed line; there's no reason why he should pay extra for it. If you don't like paying for a cellphone, don't have one. Or if you specifically don't want to pay to talk to him, just don't answer the phone -- that's what CNID is for. (You can always let the call roll to voicemail and then pick up the VM via a landline, which doesn't require any minutes at all.)
Can I do anything about them snooping in my email - regardless if it's encrypted or not? This is where I think you are wrong. There is strong evidence to suggest that modern, widely-available encryption techniques provide a substantial barrier to snooping, and make the process of snooping far more difficult than it would otherwise be. It's certainly possible that someone has the capability of decrypting 2048-bit ElGamal or other modern PK encryption, if they do it's a closely guarded secret, unavailable to the vast majority of would-be snoopers. (I.e., if the NSA does have some unimaginably powerful quantum computer in its basement, which I frankly don't think they do, they're only going to use it on very high-value targets; anything more risks revealing their capability. It's not a tool you could use for the most oppressive kinds of mass surveillance.)
Therefore the aggregate effect of large numbers of people using encryption would be to render large-scale electronic surveillance systems useless, since they are only practical for plaintext traffic. (In fact, you don't really even need to be using state-of-the-art crypto; if everyone were using even keys that took a few days to break on a supercomputer, it would prevent most types of high-speed/real-time analysis and force authorities to take much more fine-grained, targeted approaches.
Your argument against taking an individual step to prohibit mass surveillance is the same argument that many people make against voting: your action, taken singularly, has virtually no effect. It is only as part of a group that it is significant. But just as many people deciding to vote the same way can change a government, a large number of people deciding to make the snoopers' jobs (even slightly more) difficult would quickly outpace their resources available for the task.
I don't think the solution is either-or, personally. As concerned citizens, we need to vote. As people with technological knowledge and capabilities, we have a responsibility to not make it easy for those in power to abuse it, through our passivity.
I have the capability of using both S/MIME and GPG for email (using Apple Mail, it's a matter of installing gpg, getting the Sente Software gpg addon for Mail, and getting a S/MIME certificate to activate the built-in S/MIME support), but overall I think S/MIME is probably better positioned to succeed in the marketplace. It's more idiotproof.
As much as I really despise the centralized philosophy behind S/MIME and x.509, there's something to be said for avoiding the 'web of trust' models that lie underneath GPG as its currently used, because most users just don't want to have to deal with it.
Getting people to use encryption is always a tough sell, because most people, to be perfectly frank, lead lives that are so completely boring that nobody would ever want to read their mail, and they know it. Therefore, they're not going to expend much effort getting it working. Either it works all automagically, or they don't use it at all.
I've yet to see a GPG implementation that comes as close to being foolproof as some S/MIME implementations (like Apple's), once you get the certificates set up. Once you've received a signed message from someone, you have their public key. Once you have that, the encryption button is magically enabled, and you can send encrypted stuff to them. Even Sente's Mail frontend to GPG isn't that easy to use.
The U.S. system seems screwy for text messages, but it makes sense for voice calls. The caller pays for the cost of the call on the POTS system to whatever exchange the cellular number is in. Then the person with the cellphone pays for the airtime to transmit that call over the cellular network to their handset. (And they pay for the airtime whether the call is outgoing or incoming; what they're paying for is the circuit, not really the 'call.') This means, if the call originates from the same area that the cellphone's number is in, the caller pays next to nothing, since it's a local call. In fact, they have no way of knowing, just by looking at the number, whether it's a cell or landline. There's no difference in the U.S. between a "cellular number" and a "regular number."
It doesn't strike me as illogical. If it cost people more to call cellphones than landlines, the uptake of cellphones would have been a lot slower. I certainly wouldn't be able to use a cellphone as my primary business line, since it would be obnoxious to charge people more (and, hence, discourage them from calling me) because I want the ability to take calls on the road.
The U.S. pricing structure means that text messages are a bad deal (which is why they're little used here compared to in Europe), but it also sped the adoption of cell phones to many people who wouldn't have bought them otherwise, particularly business users, and it prevented people from consciously avoiding making calls to cell phones because of the expense. It puts the expense of owning a cellphone on the person who wants the convenience of being mobile, rather than on the caller.
I'm thinking about forwarding PGP encrypted copies of my real mail over to Gmail for backup/archival purposes This is a good idea in theory, but in practice it's a PITA to retrieve from. At least to my knowledge, there's no way to directly integrate PGP with Gmail's web interface, and when you use Gmail with a real desktop mail program, it's POP only -- no IMAP -- so you can't easily browse archived messages.
So you could encrypt messages and send them to Gmail, but when it came time to get a message out, you'd have to go into Gmail and send it to your regular desktop account for decryption (or copy the encrypted text out some other way) before you could do it. That strikes me as somewhat inconvenient, compared to other backup options.
If Gmail offered IMAP you could do some pretty slick stuff, because then you could use a GPG-aware desktop client to browse the encrypted messages stored on Gmail's servers, decrypting on-the-fly to browse, but with only POP access you're more limited.
Because people voted to elect people that decided to use that machine? Sure. But come on, we both know that's a trite answer. There are limits on what sort of stupidity you can vote into place. I'm personally in favor of a minimalist central government, with basically politically autonomous states that only go to the Federal level in order to resolve conflicts between them.
But there are some things that are a legitimate concern of the Federal government, because they have effects that aren't limited purely to the residents of one state. If the residents of one state decide to do something that's going to undermine confidence in the results of the entire national political system, that is a problem for the entire nation. It's not something that can be dealt with purely in isolation. One state's seemingly arbitrary, indecipherable results (particularly if it's a 'tiebreaker' state) could cast doubts on an entire administration and destroy public confidence in the democratic process generally.
So while I don't think that the Federal government should be specifying exactly how elections must be done (specifically what machines can be used, etc.), I don't think there's any problem if they specify certain minimum standards for auditability and transparency. How the states satisfy those requirements is up to them, and how the states conduct elections for local or state vacancies is completely up to them, since it's not a national issue.
You doubt the entire thing will weigh (okay, mass) 4000 pounds? Look, I was with your back-of-the-envelope numbers up to that point, but 2000 pounds for 10 megawatts of solar panels, plus meteorite shielding, control/propulsion systems, and the microwave transmitter to beam the power back down? No way. 5000 pounds is a fair weight estimate for a modern communications satellite, and they're a whole lot simpler.
Do you even have an idea of how many square feet of PV cells you need for 10MW? There's a system in Portugal that's that big, you can see a photo of it here. Even figuring that you might get slightly more efficient cells and by putting them in orbit might be able to get more power out of each, you're still talking about a *huge* station.
I strongly suspect you are talking about a Shuttle launch or using one of the Russian or European heavy-lift rockets (I think an Ariane 5 can lift something like 10,000 kilos to geostationary orbit), and that's assuming you can lift it in one shot to begin with.
I think this is neat technology too, but let's not understate the difficulty here. This is an immense undertaking.
A country should never let a large proportion of it's food production all be outsourced. What happens in case of war or political/trade fallout? Here is the problem: When the Australians, Japanese and Europeans do exactly that, the US screams "subsidies, subsidies, subsidies...", as if the US is any innocent. The U.S. is the bigger market, therefore it can negotiate trade agreements that are in its favor. Chances are, the Australians want access to the U.S. market a lot more badly than the U.S. wants access to Australia's; therefore, the U.S. can keep its subsidies and make other nations get rid of theirs.
I'd just like to say that I think your post was one of the most sensible and well-reasoned expositions on the e-voting issue that I've heard in a long time.
If I thought it would make any damn difference at all, I'd say that you should mail that to your senators and representatives, but I'm not that naive anymore.
Computerized voting machines were a solution looking for a problem. By and large, I don't know anyone that really had a huge issue with any of the existing methods of voting, except perhaps the punch-card systems. In particular, this last election, I heard lots of people lamenting the demise of the venerable old mechanical lever-based machines in favor of touch-screen ones that they found far more complicated and less intuitive.
I've never really understood what's been driving the electronic voting madness. At one point I thought it was just a weird obsession with computerization and eliminating the human element, pushed by people who didn't really understand "computers" and thought they were infallible. Sometimes I think it was a perceived need to get the official results as quickly as possible (ignoring that the Constitution provides a fairly long period for the tabulation of the official results, for good reason). And in my more paranoid moments, it's not hard to imagine a few good conspiracy and vote-rigging theories. At any rate, it's high time that we rectified out mistakes and got rid of those ridiculous machines.
Sometimes, throwing more technology at the problem isn't the solution. Sometimes, the best solution is to either leave well enough alone, or pick the best existing solution and use it, rather than developing something new to fill a need that doesn't exist.
I look forward to seeing a glut of surplus touch-screens on eBay.
Uh, please give me one good reason why municipalities should be given the option of using highly insecure, no-physical-record, easily hacked "voting" machines in elections that influence the entire nation?
By your logic, we should allow states to allocate their delegates to the Electoral College by coin toss, cockfight, or single combat, if a bunch of political appointees in that state think it's a bright idea.
I think we should rigorously enforce some sort of minimum standard of quality for elections. Above and beyond that, sure, states can choose what brand and type of machines they want. But we all have an interest in making sure that elections are fair, unbiased, and transparent. Auditless electronic voting systems prohibit that by design, and for that reason they ought to be illegal. Leave them for supermarket taste-tests where they belong.
Slashdot Mirror
I think this is one of the main reasons why so much security policy is reactive rather than proactive. Nobody wants to be the person to call out everyone else for being potential criminals, even though everyone rationally knows that it's true.
Maybe they could make voting membership in a computer-standards committee contingent on having some sort of viable technology industry or something. (Of course, in a few decades that would probably knock out the United States, the way we're going...)
Their bylaws probably prevent them from doing this except by a vote of all the P-class members.
I've seen this sort of thing happen before, to smaller organizations. You get a huge influx of members for some reason, but then they stop participating. If you didn't anticipate this possibility when drafting your constitution or bylaws, and you have some rule in there that says "changes to the bylaws must be ratified by 50% of the membership" or something similar, you're screwed. You can't change the rules, because nobody shows up, and you can't do anything, because nobody shows up.
Maybe the ISO Standards Committee should dissolve itself and reform under a slightly different name, with a better set of bylaws...
We think of sunspots as following a fixed, 11-year cycle, but this may only be one part of the story.
I don't think the 'experts' necessarily know anything more at this point, either; just a few years ago, NASA was predicting that the next cycle would be the strongest ever, and that got a lot of people (especially folks that do a lot of shortwave/HF radio) very excited. Now, it looks like we may have a very small cycle, or no cycle at all -- it's anybody's guess.
The dead spot on some sunspot charts from 1650-1700 is called the "Maunder Minimum". During that period, rather than talking about sunspots, observers of the day would write about the appearance of a particular sunspot (very much singular!). Unfortunately, the data prior to the beginning of the minimum is pretty sparse, and exactly when it started is under some dispute.
There was also another minimum in the early 19th century, called the Dalton Minimum, although it wasn't as severe and it only lasted about 25 years.
So that's two minima separated by a 150-year gap. But at 150 years after the 1800 minimum, rather than another minimum, we actually get a maximum in 1950. There's just not enough historical data to make a good prediction, because we don't know how complex the cycle is. But it's clearly more complex than just 11 years.
I can't find a link to it online, but I heard a talk recently about a group that was using geological evidence to try and track the sunspot cycle further back than we have human observations. Not sure quite what the method is, or if it's yielded any results. But that would certainly be interesting, if you could get some real historical perspective instead of the piddling 7 centuries (at most) that you can find written records of. That might give us some idea of what's been going on, on very long timescales, as well as perhaps filling in the gaps in the historical record in more recent times (not sure what kind of resolution you can get).
To use a water analogy, the 11-year cycles might be waves lapping at the shore, but there might be scores of other forces acting on them at higher levels, like tides, wind, and the seasons, all on vastly different time-scales.
All in all, for something that we spend the majority of our waking lives under, our understanding of the sun is surprisingly poor. Particularly given how much modern technology (radio communications is the obvious one, but there are others) can be affected by the solar cycle, it seems to be ignored until it does something unexpected.
For the Coral Cache you just append .nyud.net to the end of the domain.
.nyud.net sites resolve to 132.239.17.225 for me (which happens to be the node in Princeton), you can't just connect to that and get the cached page.
E.g., "http://www.meangene.com/google/design_for_google.html" becomes "http://www.meangene.com.nyud.net/google/design_for_google.html"
I don't think you can connect to it directly using an IP address, since it uses the domain portion of the URL to figure out what page you're requesting. Although all
If you're still blocked you'll need to tunnel out to a proxy.
Humm. It seems like the $69 price for Leopard (desktop, single-seat) is only available at some EDUs. Lots of other people are going in and seeing significantly higher prices.
I wonder what the basis for the difference is? Apparently some institutions are a little more equal than others. I wonder if it represents those that have big Apple site-licenses or something?
Basically, this is just because Apple doesn't want people to put off purchasing computers once they release the new OS, because they're waiting on ones that come with it preloaded. (I suspect that people do wait anyway, but it's their effort to mitigate this.)
Up until last year I was running OS X on a 400MHz G4; it got progressively faster with each release and was still totally usable. (It's still in use, too; I just had to give it to a family member whose computer died rather suddenly and spectacularly, giving me an opportunity to buy myself a nice present.) Shelling out a few hundred bucks every couple of years was totally worth it to keep getting more performance out of a 1999 computer.
Some of the features in Leopard are neat, but I don't see anything that would justify any type of performance hit. So I'm waiting on upgrading until I hear whether this release keeps up the gains in "teh snappy" that I've gotten used to.
The retail versions, by contrast, will run on any machine that's listed as capable of running the software. (Which sometimes is slightly different than the machines that are *actually* capable of running the software; Apple specs systems that are capable of running the OS comfortably, but some people have found acceptable results after forcing it onto older machines.)
If you wait around until the next paid-upgrade OS release though, you can get the older version, in retail packaging, quite cheap. Either eBay or some of the used-Mac stores like Smalldog regularly have new-old-stock retail OS packages.
You do realize that 298 of those 1195 SEK are tax, right? So subtracting that out, you get a real price of 897 SEK, which is only 68 SEK more than the US price, or about $10.60 USD.
I doubt that you'd be able to order a US version and have it shipped to Sweden for less than $10 in shipping.
Seems like a pretty fair price to me. Maybe you should vote for politicians who support lower taxes if you don't like it?
Thanks for playing, though. I don't think --link-dest does binary diffs/deltas, though, does it? If you modify a large file, it will create new copies of them rather than keeping track of the deltas. While this might be a good thing from a reliability standpoint, I could imagine that it could quickly cause your backups to get quite large.
I can't quite figure out how Time Machine works, but I think its operation is closer to rdiff-backup than rsync --link-dest. If it's not, the space requirements for it are going to be huge.
Depending on how you're using Ekiga (seriously, could they have found a more difficult to pronounce name?) you could still be sending your traffic through a CALEA-compliant (and thus snoopable) network.
Ekiga is just the client program that runs on your computer; it is to VoIP what Firefox is to the WWW. The client program isn't (generally) where tapping occurs; law enforcement does that in the network, where it's harder to detect. So the question isn't whether you use Ekiga, it's who do you use for an ISP, and who do you use for SIP-to-POTS service?
If you use one of the major US ISP's, you have to look no further than them for the authorities' way into your traffic, since they can provide ways to tap the SIP traffic directly if it's unencrypted (and I don't think most SIP implementations support encryption, more shame on them). It's not terribly hard to pick out the right packets if you're sniffing the connection. So even if you're calling from your computer to your friend's computer completely via SIP, never going near the POTS network, you could still be snooped.
If you use a SIP-to-POTS gateway service (allowing you to call landline numbers from your SIP phone), that provider -- if it's based in the US -- almost certainly is CALEA-compliant and will provide a way to tap. I'm fairly certain that whoever does Ekiga.net's gateway service (Diamondcard.us?) is either CALEA-compliant themselves, or whoever they buy POTS circuits from is.
To be frank, it's a fool's errand to try and avoid CALEA by switching providers. If you want to talk with people in the U.S., your traffic is going to go through a network that's CALEA-compliant and can be snooped on. It may be marginally more difficult right now for authorities to snoop on a completely IP-based, SIP conversation than on a POTS one, but this will probably change as law enforcement becomes more comfortable with the technology.
The solution, IMO, for a person desiring private communication, is not to rely on the security of the data channel, but to create that security using encryption for the conversation itself. Zfone is Phil Zimmerman's modern update to PGPfone, and works in conjunction with most SIP clients at the protocol stack level to encrypt the SIP stream. It looks pretty slick, although I haven't played with it much myself. There seem to be versions available for Mac, Linux, and Windows. Sadly, the code is not GPL, so it will probably never appear in mainstream Linux distributions.
The rule is one rifle (AK-47 or similar) per household for protection, no heavy weapons, explosives, or caches of weapons.
The military isn't so stupid as to ban civilian ownership of all weapons; it would just make the population more exposed -- not just to foreign hostiles, but also to sectarian violence, and the usual criminal elements -- rather than safer.
This won't work for the reasons that other people have noted.
The best security precaution is continual awareness. If you're intimately familiar with all of your hardware and software, it's a lot harder for someone to install a keylogger. Would you know if someone came into your office and moved something around? You should. It requires an effort, though, to start paying attention to little things, so that you'll notice if something is amiss. And if you have a bad feeling, you need to act on it immediately.
Would you notice if someone swapped your keyboard with one of an identical make and model and approximate age? And if you did notice something odd -- maybe a little stiffness in the keys that wasn't there before, a difference in the wear patterns from where your fingers normally lie -- would you just shrug it off or would you immediately stop using it? How often do you actually look behind your desk to see if someone has shoved one of these in between your keyboard and CPU? Those are the things you have to take into consideration.
It's similar with software. A while back I read about a guy who only discovered he'd been rooted because of an oddly misbehaving "ls" command when it was invoked with certain switches. Lots of other intrusions are only discovered because of similar, very subtle, signs. (Most of which boil down to the intruder making a mistake somewhere.)
Most people don't want to have to pay attention to security, and thus look for easy ways out. This is generally where they become most vulnerable. Automated and procedural security is good, but ultimately any 'fire and forget' approach is fatally flawed. There's no replacement for vigilance.
S/MIME is designed to work with centralized Certificate Authorities. If you roll your own CA and issue yourself a self-signed certificate, you'll be able to sign stuff, but people who receive your messages will get a big "BAD SIGNATURE" error or warning, because they won't have your CA in their trusted chain. In order to get it to work, you'd need to get them the CA certificate, and they'd need to import it into their trusted root database. (Which is a security risk -- you do not want to encourage clueless users to start importing certs from every idiot they want to talk to into their Trusted Root.)
It is much better to just get a personal certificate from Thawte or several of the other places online that give them out. Thawte is aimed at people who want authenticated communication; it's not anonymous and in fact they require some form of Government ID in order to issue one. If you want to use S/MIME anonymously or pseudonymously, you're better off going to OpenCA and getting one through them. (Their CA cert isn't included by default in most browsers and OSes like Thawte's is, but at least your correspondents only need to import one additional certificate to recognize yours, and it comes from a basically legitimate institution. That's a lot better than importing random people's CA certs into your root DB.)
Therefore the aggregate effect of large numbers of people using encryption would be to render large-scale electronic surveillance systems useless, since they are only practical for plaintext traffic. (In fact, you don't really even need to be using state-of-the-art crypto; if everyone were using even keys that took a few days to break on a supercomputer, it would prevent most types of high-speed/real-time analysis and force authorities to take much more fine-grained, targeted approaches.
Your argument against taking an individual step to prohibit mass surveillance is the same argument that many people make against voting: your action, taken singularly, has virtually no effect. It is only as part of a group that it is significant. But just as many people deciding to vote the same way can change a government, a large number of people deciding to make the snoopers' jobs (even slightly more) difficult would quickly outpace their resources available for the task.
I don't think the solution is either-or, personally. As concerned citizens, we need to vote. As people with technological knowledge and capabilities, we have a responsibility to not make it easy for those in power to abuse it, through our passivity.
I have the capability of using both S/MIME and GPG for email (using Apple Mail, it's a matter of installing gpg, getting the Sente Software gpg addon for Mail, and getting a S/MIME certificate to activate the built-in S/MIME support), but overall I think S/MIME is probably better positioned to succeed in the marketplace. It's more idiotproof.
As much as I really despise the centralized philosophy behind S/MIME and x.509, there's something to be said for avoiding the 'web of trust' models that lie underneath GPG as its currently used, because most users just don't want to have to deal with it.
Getting people to use encryption is always a tough sell, because most people, to be perfectly frank, lead lives that are so completely boring that nobody would ever want to read their mail, and they know it. Therefore, they're not going to expend much effort getting it working. Either it works all automagically, or they don't use it at all.
I've yet to see a GPG implementation that comes as close to being foolproof as some S/MIME implementations (like Apple's), once you get the certificates set up. Once you've received a signed message from someone, you have their public key. Once you have that, the encryption button is magically enabled, and you can send encrypted stuff to them. Even Sente's Mail frontend to GPG isn't that easy to use.
The U.S. system seems screwy for text messages, but it makes sense for voice calls. The caller pays for the cost of the call on the POTS system to whatever exchange the cellular number is in. Then the person with the cellphone pays for the airtime to transmit that call over the cellular network to their handset. (And they pay for the airtime whether the call is outgoing or incoming; what they're paying for is the circuit, not really the 'call.') This means, if the call originates from the same area that the cellphone's number is in, the caller pays next to nothing, since it's a local call. In fact, they have no way of knowing, just by looking at the number, whether it's a cell or landline. There's no difference in the U.S. between a "cellular number" and a "regular number."
It doesn't strike me as illogical. If it cost people more to call cellphones than landlines, the uptake of cellphones would have been a lot slower. I certainly wouldn't be able to use a cellphone as my primary business line, since it would be obnoxious to charge people more (and, hence, discourage them from calling me) because I want the ability to take calls on the road.
The U.S. pricing structure means that text messages are a bad deal (which is why they're little used here compared to in Europe), but it also sped the adoption of cell phones to many people who wouldn't have bought them otherwise, particularly business users, and it prevented people from consciously avoiding making calls to cell phones because of the expense. It puts the expense of owning a cellphone on the person who wants the convenience of being mobile, rather than on the caller.
So you could encrypt messages and send them to Gmail, but when it came time to get a message out, you'd have to go into Gmail and send it to your regular desktop account for decryption (or copy the encrypted text out some other way) before you could do it. That strikes me as somewhat inconvenient, compared to other backup options.
If Gmail offered IMAP you could do some pretty slick stuff, because then you could use a GPG-aware desktop client to browse the encrypted messages stored on Gmail's servers, decrypting on-the-fly to browse, but with only POP access you're more limited.
Because people voted to elect people that decided to use that machine? Sure. But come on, we both know that's a trite answer. There are limits on what sort of stupidity you can vote into place. I'm personally in favor of a minimalist central government, with basically politically autonomous states that only go to the Federal level in order to resolve conflicts between them.
But there are some things that are a legitimate concern of the Federal government, because they have effects that aren't limited purely to the residents of one state. If the residents of one state decide to do something that's going to undermine confidence in the results of the entire national political system, that is a problem for the entire nation. It's not something that can be dealt with purely in isolation. One state's seemingly arbitrary, indecipherable results (particularly if it's a 'tiebreaker' state) could cast doubts on an entire administration and destroy public confidence in the democratic process generally.
So while I don't think that the Federal government should be specifying exactly how elections must be done (specifically what machines can be used, etc.), I don't think there's any problem if they specify certain minimum standards for auditability and transparency. How the states satisfy those requirements is up to them, and how the states conduct elections for local or state vacancies is completely up to them, since it's not a national issue.
You doubt the entire thing will weigh (okay, mass) 4000 pounds? Look, I was with your back-of-the-envelope numbers up to that point, but 2000 pounds for 10 megawatts of solar panels, plus meteorite shielding, control/propulsion systems, and the microwave transmitter to beam the power back down? No way. 5000 pounds is a fair weight estimate for a modern communications satellite, and they're a whole lot simpler.
Do you even have an idea of how many square feet of PV cells you need for 10MW? There's a system in Portugal that's that big, you can see a photo of it here. Even figuring that you might get slightly more efficient cells and by putting them in orbit might be able to get more power out of each, you're still talking about a *huge* station.
I strongly suspect you are talking about a Shuttle launch or using one of the Russian or European heavy-lift rockets (I think an Ariane 5 can lift something like 10,000 kilos to geostationary orbit), and that's assuming you can lift it in one shot to begin with.
I think this is neat technology too, but let's not understate the difficulty here. This is an immense undertaking.
Life is not fair.
I'd just like to say that I think your post was one of the most sensible and well-reasoned expositions on the e-voting issue that I've heard in a long time.
If I thought it would make any damn difference at all, I'd say that you should mail that to your senators and representatives, but I'm not that naive anymore.
Computerized voting machines were a solution looking for a problem. By and large, I don't know anyone that really had a huge issue with any of the existing methods of voting, except perhaps the punch-card systems. In particular, this last election, I heard lots of people lamenting the demise of the venerable old mechanical lever-based machines in favor of touch-screen ones that they found far more complicated and less intuitive.
I've never really understood what's been driving the electronic voting madness. At one point I thought it was just a weird obsession with computerization and eliminating the human element, pushed by people who didn't really understand "computers" and thought they were infallible. Sometimes I think it was a perceived need to get the official results as quickly as possible (ignoring that the Constitution provides a fairly long period for the tabulation of the official results, for good reason). And in my more paranoid moments, it's not hard to imagine a few good conspiracy and vote-rigging theories. At any rate, it's high time that we rectified out mistakes and got rid of those ridiculous machines.
Sometimes, throwing more technology at the problem isn't the solution. Sometimes, the best solution is to either leave well enough alone, or pick the best existing solution and use it, rather than developing something new to fill a need that doesn't exist.
I look forward to seeing a glut of surplus touch-screens on eBay.
Uh, please give me one good reason why municipalities should be given the option of using highly insecure, no-physical-record, easily hacked "voting" machines in elections that influence the entire nation?
By your logic, we should allow states to allocate their delegates to the Electoral College by coin toss, cockfight, or single combat, if a bunch of political appointees in that state think it's a bright idea.
I think we should rigorously enforce some sort of minimum standard of quality for elections. Above and beyond that, sure, states can choose what brand and type of machines they want. But we all have an interest in making sure that elections are fair, unbiased, and transparent. Auditless electronic voting systems prohibit that by design, and for that reason they ought to be illegal. Leave them for supermarket taste-tests where they belong.