Nerve impulses are incredibly slow: they max out at 120 m/s, and are often much slower (10 m/s) depending on the nerve. Contrast with electrical connections, where even with transmission-line effects, the signal goes at a good fraction of the speed of light (let's say 300000 m/s).
Our brain's power comes from the complexity and subtlety of the interconnect, not from the raw speed of processing.
That's what any scientific "law" is, from Ohm's Law, to Kepler's Law of Equal Areas in Equal Times, to the Law of Gravitation, etc.: an observation that this quantity appears to always match that quantity. They're laws in the sense of rules, in the sense of "rule of thumb" (though typically much more accurate than a rule of thumb). They're not laws in the legal sense of something that's imposed from outside. This is another place where English confusingly uses the same words for two different concepts, much like the "free as in beer vs. free as in speech" distinction that is already familiar to any slashdot reader.
Of course, the submitter (and original article) are still saying something other than what they think they are saying. It's not Moore's Law that has a limit; it's physics that has a limit. We don't know what that limit is yet, but it seems pretty likely that we won't be able to keep up with the Moore's Law trend indefinitely. So far so good, but someday...
Hey, it's not like all those millions of users, and Google themselves, aren't paying for that bandwidth. If Verizon's business model relies on people not using the service they're paying for, well, too bad for them.
"Neutronium" is a common name for the collapsed matter of a neutron star. If I used my Implausitron to keep the stuff stable in smaller-than-star quantities, it'd still be neutronium, but no longer a neutron star.
Yep. And fixing that would be really easy, since you just have to find big blocks of duplicated text. It's annoying and time-consuming to dig through Google News for the 2 or 3 actual distinct stories among the dozens of reposted wire articles.
Dragonball refers to (at least?) two CPU architectures. The Dragonball is a line of embeddable CPU chips from Motorola (now Freescale); it's what PalmPilots have been using since day one. But "dragonball" is a product line or brand name, not an architecture. The older chips were based on the Motorola 68k architecture; the newer ones are based on the ARM920T architecture.
The Treo 300 used an m68k processor; the 600 and 650 both use ARMs.
More than ten by now. Hundreds of Kuiper belt objects are known. The ones that make the news are ones that are interesting for some reason: larger than usual (Quaoar, Sedna, Xena/Gabrielle 2003 UB313, etc.) or in unusual orbits (this one).
Sorry, not true. This particular kind of DRM doesn't break the CD spec. Other kinds of "copy protection", such as SunComm's MediaCloQ, do violate the CD specification and are technically not CDs. But schemes which are based on autorun trojans, like the scheme in the article or like SunComm's MediaMax, don't violate the CD specification at all.
One of Freenet's primary goals may be resistance to a powerful, organized opposition, but what keeps such an opposition to simply label anyone running a freenet node as a criminal?
Well, that's essentially the situation now with freenet, and anonymous remailers, and whatnot. People are afraid to use them because they don't want to stick out, even if they're innocent of any crime. One of the goals of the cypherpunk types was to get enough people using anonymizing software that the mere use of such software would not be enough to look unusual. PGP, ssh, and SSL have more or less been accepted, but more esoteric things like mixmaster, freenet, and onion routing have a harder time.
Network flow as a method of automatically determining trust networks and optimizing connections is an interesting idea. I think there are some problems which need to be worked out, though.
Clique formation. Self-reinforcing reputation systems have a tendency to form into small isolated systems which don't talk to outsiders, even if the outsiders would otherwise be considered interesting to talk to. This can be hard to notice, because from inside the clique everything seems fine --- you can't see most of the network, but you don't notice this because everyone you can talk to is in the same boat as you. IIRC, Bram Cohen looked at something like this for trackerless bittorrent and noted this problem. Advogato has this problem. Real-world societies have this problem. In order to avoid it, you need to continually exchange data with nodes that you don't think you like.
Vulnerability to traffic analysis. One of Freenet's primary goals is resistance to powerful, organized opposition (such as an oppressive government, or the Mob, or whatever). If the network clusters traffic based on content, then it becomes extremely vulnerable to traffic analysis. It doesn't matter that all the traffic is encrypted: all I need to do is find one dissident or undercover cop (or plant an infiltrator) and then watch which nodes get pulled into the dissident's traffic cluster. Those nodes are also operated by dissidents, so now I know who to assassinate / torture / imprison.
Non-universal ideas of "goodness" vs "badness" Let's say I fire up STAY+Freenet and spend a while using it for GPL'd software. It determines its trust metric based on that, excluding from my network all nodes that don't traffic in GPL'd software. Then I decide to look at some BSD-licensed software. Ooops! All of that software has fallen over the "trust horizon" and is invisible to me.
It's also worth noting that Freenet doesn't require any nodes to be "trustworthy"; everything's checked against hashes. This is similar to BitTorrent: the torrent file allows you to verify whether a peer is feeding you good data, and if a peer gives you bad data you just drop it and connect to a different one.
That's likely just it: the evidence is manifestly circumstantial, and might not result in the type of punishment sought, or indeed, even a conviction.
So what you're saying is, the guy wouldn't be convicted in an honest court, and this justifies keeping him out of the courts? According to your argument, he's probably innocent of any crime (where "crime" here means "something that's actually illegal", and "innocent" includes "unless proven guilty"). You're advocating imprisoning innocent people based on the completely un-appealable say-so of some unknown government official. I guess you might think this is a good idea, until it happens to you... me, I still remember the "liberty and justice for all" part of the Pledge, even if I think the "under God" part is a piece of tripe.
Me too. I've also thought it would be fun to stencil or poster that pattern around, just to confuse people who might take a picture with the pattern in the background.
I think there's a very, very good chance that all these printers have in-the-field-upgradeable firmware in Flash, so all you need to do is reverse engineer one printer's firmware, find the "download new firmware" command in it, and use that to produce a software-only printer crack which updates any printer of that model so that it won't print the tracking dots. Instead, it will just print goatse on every third page.
The firmware (which you can download from MobiBlu's website) has "StationZ" and "MP3PZ" scattered all through it. So I'm guessing the hardware is pretty closely related to the StationZ "pendant" MP3 players.
Unfortunately I'm not leet enough to recognize what cpu it's using from a hex dump, so I can't guess anything more from the firmware file. (FYI, 'strings' only gets about half of the text in the file, since the rest is in UTF16 or some other 2-byte character set.)
The equivalent of open source hardware is the commodity beige-box hardware that the PC world is known for. Remember when PC hardware prices fell through the floor as the number of producers increased (early/mid 90's, IIRC)?
Things like motherboards are already effectively open source in some ways. The chipset manufacturer publishes a reference design which they say the chips should work with. Some motherboards are basically just the reference design put into mass production. Other mobo vendors do more of their own (re)design of course. (Think BSD-style open source here, not GPL.)
And re the Xbox -- isn't Microsoft still losing money on each sale? Selling hardware at a loss and making up the difference in software sales doesn't make them a hardware company!
It's an interesting phrase. Most often, in discussions of economics, the "tragedy" is assumed to be the overuse by one person of a common resource --specifically, the overgrazing of a village's common land because each farmer figures they can add a few more sheep to their flock. This tragedy scenario is used to argue that common resources will be destroyed by use, and therefore must be removed from common ownership and owned by some particular person or organization in order to preserve them.
Except that the whole idea is historically inaccurate. In reality, this scenario was avoided by a complex set of social norms. Everyone in the village had a stake in keeping the commons useful and generally managed to keep it so, despite the theories of economists. This worked until the land was enclosed -- divided up and put under private ownership, less practical on a small scale, and generally forcing small farms out, or forcing them to rent from a few giant landowners. In general, the few large landowners profited, and the many smaller landowners became poorer.
It's been noticed by many that the copyright of music and other intellectual "property" is the same kind of enclosure. It takes valuable things -- Beethoven's music, say -- out of common hands and places them into the hands of a few giant corporations. And in the case of the IP commons, we don't even have Hardin's argument for enclosure: no matter how many people listen to a piece of music, or run a piece of software, it's not going to get used up or worn out.
I think of this as "cargo-cult management". (Check wikipedia if you're not familiar with cargo cults.) People notice that many successful companies have a coherent core values and missions and whatnot, and certain kinds of documentation and processes, and decide that those things cause success, when in many cases it's much more likely that some third quality is causing both success and the observable "cargo" of mission statements.
Seems to me that one of the big benefits of inlined functions is actually not eliminating the function-call overhead, but giving the optimizer (register allocator, CSE, instruction scheduler) a crack at what would otherwise have been inter-procedural optimizations.
Yeah, I do this sometimes when somebody's giving an especially content-free report or something at a meeting. Sometimes it works. Sometimes it doesn't work; when it fails it usually fails in one of a few ways:
Other guy keeps responding with no information; everyone else is annoyed and just wants to move on so they can leave the meeting.
Other guy responds with actual information, but everyone else really just wants to get the meeting over with so they're annoyed at the slowdown.
I just look like a dumbass for not pretending to know what the other guy means by "some granular issues which we are pinging back over the table" or whatever.
Despite its occasional failures, I think it's a worthwhile approach.
I don't see why the EU's patent regime would make companies want to move development offshore. If they develop software in India or China, it's still unpatented in the EU. So this victory might make it less profitable to sell software in the EU, due to increased competition; but it will not make it more costly to develop software in the EU.
Slashdot Mirror
For modern, commercially available magnetic memory, you could use FRAM (ferroelectric RAM).
Nerve impulses are incredibly slow: they max out at 120 m/s, and are often much slower (10 m/s) depending on the nerve. Contrast with electrical connections, where even with transmission-line effects, the signal goes at a good fraction of the speed of light (let's say 300000 m/s).
Our brain's power comes from the complexity and subtlety of the interconnect, not from the raw speed of processing.
Of course, the submitter (and original article) are still saying something other than what they think they are saying. It's not Moore's Law that has a limit; it's physics that has a limit. We don't know what that limit is yet, but it seems pretty likely that we won't be able to keep up with the Moore's Law trend indefinitely. So far so good, but someday...
Hey, it's not like all those millions of users, and Google themselves, aren't paying for that bandwidth. If Verizon's business model relies on people not using the service they're paying for, well, too bad for them.
Indeed, it's different from all of those, read TFA. Here's a basic explanation of the device from RPI.
"Neutronium" is a common name for the collapsed matter of a neutron star. If I used my Implausitron to keep the stuff stable in smaller-than-star quantities, it'd still be neutronium, but no longer a neutron star.
And their recent changes have made it worse and worse.
Yep. And fixing that would be really easy, since you just have to find big blocks of duplicated text. It's annoying and time-consuming to dig through Google News for the 2 or 3 actual distinct stories among the dozens of reposted wire articles.
That's why computer programming is difficult, especially security-related computer programming: you have to deal with people doing unexpected things.
Dragonball refers to (at least?) two CPU architectures. The Dragonball is a line of embeddable CPU chips from Motorola (now Freescale); it's what PalmPilots have been using since day one. But "dragonball" is a product line or brand name, not an architecture. The older chips were based on the Motorola 68k architecture; the newer ones are based on the ARM920T architecture.
The Treo 300 used an m68k processor; the 600 and 650 both use ARMs.
More than ten by now. Hundreds of Kuiper belt objects are known. The ones that make the news are ones that are interesting for some reason: larger than usual (Quaoar, Sedna, Xena/Gabrielle 2003 UB313, etc.) or in unusual orbits (this one).
Sorry, not true. This particular kind of DRM doesn't break the CD spec. Other kinds of "copy protection", such as SunComm's MediaCloQ, do violate the CD specification and are technically not CDs. But schemes which are based on autorun trojans, like the scheme in the article or like SunComm's MediaMax, don't violate the CD specification at all.
Well, that's essentially the situation now with freenet, and anonymous remailers, and whatnot. People are afraid to use them because they don't want to stick out, even if they're innocent of any crime. One of the goals of the cypherpunk types was to get enough people using anonymizing software that the mere use of such software would not be enough to look unusual. PGP, ssh, and SSL have more or less been accepted, but more esoteric things like mixmaster, freenet, and onion routing have a harder time.
God speaks to me personally and tells me that Freenet is used by Chinese dissidents to exchange Biblical child pornography.
- Clique formation. Self-reinforcing reputation systems have a tendency to form into small isolated systems which don't talk to outsiders, even if the outsiders would otherwise be considered interesting to talk to. This can be hard to notice, because from inside the clique everything seems fine --- you can't see most of the network, but you don't notice this because everyone you can talk to is in the same boat as you. IIRC, Bram Cohen looked at something like this for trackerless bittorrent and noted this problem. Advogato has this problem. Real-world societies have this problem. In order to avoid it, you need to continually exchange data with nodes that you don't think you like.
- Vulnerability to traffic analysis. One of Freenet's primary goals is resistance to powerful, organized opposition (such as an oppressive government, or the Mob, or whatever). If the network clusters traffic based on content, then it becomes extremely vulnerable to traffic analysis. It doesn't matter that all the traffic is encrypted: all I need to do is find one dissident or undercover cop (or plant an infiltrator) and then watch which nodes get pulled into the dissident's traffic cluster. Those nodes are also operated by dissidents, so now I know who to assassinate / torture / imprison.
- Non-universal ideas of "goodness" vs "badness" Let's say I fire up STAY+Freenet and spend a while using it for GPL'd software. It determines its trust metric based on that, excluding from my network all nodes that don't traffic in GPL'd software. Then I decide to look at some BSD-licensed software. Ooops! All of that software has fallen over the "trust horizon" and is invisible to me.
It's also worth noting that Freenet doesn't require any nodes to be "trustworthy"; everything's checked against hashes. This is similar to BitTorrent: the torrent file allows you to verify whether a peer is feeding you good data, and if a peer gives you bad data you just drop it and connect to a different one.Me too. I've also thought it would be fun to stencil or poster that pattern around, just to confuse people who might take a picture with the pattern in the background.
Reminds me of the comp.basilisk FAQ...
I think there's a very, very good chance that all these printers have in-the-field-upgradeable firmware in Flash, so all you need to do is reverse engineer one printer's firmware, find the "download new firmware" command in it, and use that to produce a software-only printer crack which updates any printer of that model so that it won't print the tracking dots. Instead, it will just print goatse on every third page.
Re currency detectors, check out the "EURion constellation": PDF, etc. (slashdot, freedom to tinker)
The firmware (which you can download from MobiBlu's website) has "StationZ" and "MP3PZ" scattered all through it. So I'm guessing the hardware is pretty closely related to the StationZ "pendant" MP3 players. Unfortunately I'm not leet enough to recognize what cpu it's using from a hex dump, so I can't guess anything more from the firmware file. (FYI, 'strings' only gets about half of the text in the file, since the rest is in UTF16 or some other 2-byte character set.)
The equivalent of open source hardware is the commodity beige-box hardware that the PC world is known for. Remember when PC hardware prices fell through the floor as the number of producers increased (early/mid 90's, IIRC)? Things like motherboards are already effectively open source in some ways. The chipset manufacturer publishes a reference design which they say the chips should work with. Some motherboards are basically just the reference design put into mass production. Other mobo vendors do more of their own (re)design of course. (Think BSD-style open source here, not GPL.) And re the Xbox -- isn't Microsoft still losing money on each sale? Selling hardware at a loss and making up the difference in software sales doesn't make them a hardware company!
It's an interesting phrase. Most often, in discussions of economics, the "tragedy" is assumed to be the overuse by one person of a common resource --specifically, the overgrazing of a village's common land because each farmer figures they can add a few more sheep to their flock. This tragedy scenario is used to argue that common resources will be destroyed by use, and therefore must be removed from common ownership and owned by some particular person or organization in order to preserve them.
Except that the whole idea is historically inaccurate. In reality, this scenario was avoided by a complex set of social norms. Everyone in the village had a stake in keeping the commons useful and generally managed to keep it so, despite the theories of economists. This worked until the land was enclosed -- divided up and put under private ownership, less practical on a small scale, and generally forcing small farms out, or forcing them to rent from a few giant landowners. In general, the few large landowners profited, and the many smaller landowners became poorer.
It's been noticed by many that the copyright of music and other intellectual "property" is the same kind of enclosure. It takes valuable things -- Beethoven's music, say -- out of common hands and places them into the hands of a few giant corporations. And in the case of the IP commons, we don't even have Hardin's argument for enclosure: no matter how many people listen to a piece of music, or run a piece of software, it's not going to get used up or worn out.
I think of this as "cargo-cult management". (Check wikipedia if you're not familiar with cargo cults.) People notice that many successful companies have a coherent core values and missions and whatnot, and certain kinds of documentation and processes, and decide that those things cause success, when in many cases it's much more likely that some third quality is causing both success and the observable "cargo" of mission statements.
Seems to me that one of the big benefits of inlined functions is actually not eliminating the function-call overhead, but giving the optimizer (register allocator, CSE, instruction scheduler) a crack at what would otherwise have been inter-procedural optimizations.
- Other guy keeps responding with no information; everyone else is annoyed and just wants to move on so they can leave the meeting.
- Other guy responds with actual information, but everyone else really just wants to get the meeting over with so they're annoyed at the slowdown.
- I just look like a dumbass for not pretending to know what the other guy means by "some granular issues which we are pinging back over the table" or whatever.
Despite its occasional failures, I think it's a worthwhile approach.I don't see why the EU's patent regime would make companies want to move development offshore. If they develop software in India or China, it's still unpatented in the EU. So this victory might make it less profitable to sell software in the EU, due to increased competition; but it will not make it more costly to develop software in the EU.