Have you *ever* watched a local government at work? Doesn't sound like it. You can vote out town council members sometimes; you can't vote out the Building Department. And local governments aren't elected because they're telecomm visionaries, they're elected because they promise to do good things with schools or beautify downtown or improve housing or kill Wallmart. "Wouldn't say no to a deal that was profitable"? Cities don't *make* profits, and existing suburbs often don't have the right economics to run conduit under residential streets, though you can do it in new developments. My town in Silicon Valley wasn't willing to let anybody install fiber access rings to businesses without a guaranteed cut of the profits, independent of whatever technology got deployed on the fiber. And the places that you can cross Highway 101 and 237 are few and extremely far between, compared to the high density of new business construction on the bay side during the boom, and *forget* dealing with the Santa Clara Valley Light Rail people.
A couple decades ago, my church in New Jersey built a building (we'd been without one for a while.) "Somehow" the township bureaucrats "forgot" to tell the county road department that we wanted to use the nearest sewer connection, so when they paved the street they didn't leave us access to it, and we're not allowed to break open the road for five years. "What a shame - you'll just have to connect to the next connection a couple hundred feet away and pump uphill." So we got a bid of $4000 to do this basically unnecessary work, and found that "No, what a shame, that contractor's not approved by *our* township, you've got to use an *approved* contractor", and it seems the two approved contractors wanted $12K and $25K for it. At that point our building committee guy started talking about going to the town council and maybe the state about it and walked out, and the building department met him out in the parking lot and said they might be able to come to some arrangement about it. They never explicitly asked for a bribe, and we absolutely never volunteered to pay one, which had a lot to do with why they'd caused this problem in the first place, though I suspect our contractor might have had to kick back a bit of money in return for a special permit. And everything else we did with the building got lots of extra inspection, and we ended up having to buy an amazingly fancy fire alarm system to handle the new regulations that kept being discovered, because we never did offer them a bribe or ask them which contractors we "really ought to" be dealing with.
So yes, local governments absolutely will reject deals that might be "profitable", and yes, they can stop utilities (regulated or unregulated) from using existing right-of-way, and even if they're not personally corrupt or greedy for town government revenues, they're still typically not competent at telecomm.
Being a respected figure at a reputable industry source does add some credibility, mostly because editors sometimes make sure fact-checking gets done and basic writing skills get used, and a certain amount of Darwinism gets rid of many of the less capable and less insightful writers, but so what? It's an open-source news industry these days, and if some blogger says something insightful and interesting, it makes sense for Slashdot to pick it up, and if some well-respected pundit at a reputable trade rag says something lame and uninteresting, it makes sense for Slashdot to ignore it (unless somebody's writing an article about how lame most of the industry mouthpieces are.)
In this case, I don't think the article has much depth to it - the main concept is appealing, but I don't see enough thought behind it to really win. But even so, I'd mod you -1 Flamebait:-)
The author asserts that the "network is the computer" approach failed because the networks weren't fast enough to support client/server, and now they are. But the main use for network-based computing is web browsing, and there's already enough bandwidth. Sure, instead of a browser client on your PC, you could have an X Windows server on your PC and a browser app running on your ISP's server, and for faster DSL and cable modem users that might perform tolerably well, but I'm not convinced that it improves security or reliability significantly; for laptop users on slower connections it still doesn't perform well enough. Gamers are a different class of user, who really do need local CPUs to generate graphics fast enough. Network bandwidth doesn't scale for that.
The real differences are between "X terminal" approaches and "diskless workstation" and "local storage" approaches, and local storage seems to keep winning.
Yup. That orphaned open parentheses in my article was supposed to be followed by a comment that the year I moved here I ended up spending more on electricity heating my condo in the Bay Area than I had spent on oil heating my standalone house in New Jersey, where we actually had something resembling winter, as opposed to a moderately cool rainy season. It's gotten a bit better, and my electricity costs haven't gone up as fast as oil, and for a few years we had a cheaper-at-night pricing system that cut my electric bills down. (I think that died when now-Terminated governor Grey Davis was mismangling our electric system.)
On the other hand, I lived in California for a year in the late 70s, back when they were building those electrically-heated buildings, in a gas-heated one-bedroom apartment that never used more than $13 for gas+electricity in the worst of the winter, which was less than the electric bill alone for the gas-heated two-bedroom townhouse in New Jersey the next year.
According to the article, they're talking about spending $175 billion to build this edifice, but it's only going to bring in $135 billion in toll revenue over 50 years. That means that either the economics fail, and a toll road company would be unlikely to invest in it, or else they'd have to pour in at least $40B in tax funding, possibly a lot more if they're doing big favors for their politically-well-connected good old boys in the toll road business. Surprisingly, that's only $2000 per Texan worth of pork barrel, until you get to the hidden costs that probably inflate it way above the $175B costs or find that the toll revenues aren't close to the predicted $135B, so the shortfall's even larger.
32 bits? 10 Years ago? I've been using 32-bit machines for 25 years, though the Vaxen did hide some memory limitations inside that made it hard to do more than about 26 bits of address space some years (or was it 22?) and some various Motorola things and DSPs tended to be 24 bits of usable space.
But yes, yesterday's supercomputer, the Cray-1, was about the speed of a Pentium-133. Too slow to run Microsoftware on these days, though if you ran Win95 and Netscape 2 and Excel 5 you'd probably think it was pretty responsive. And X Windows ran ok on my 33 MHz 386 box, though Open Look was a bit slow, so if you want a modern Linux on a P133 (which you do) you probably need a bunch of extra RAM and a lightweight window manager - fvtwm or blackbox or something is a lot friendlier than KDE/Gnome.
Back when I had a _big_ database machine, I think we'd installed about 4 or 8 of the 16 CPUs it could hold, because that was enough for most of the customer benchmarks we wanted to do. This was at NCR in ~1994-95, so they were starting to be Pentiums, and I think we also made 32-CPU machines. (That doesn't count the Teradatas which had several hundred CPUs, one per disk drive plus a few master CPUs:-) We were using the Unix System V Release 4 multiprocessor support, which did a decent job of balancing loads on them.
But for home, since I'm not a gamer, I've never had a good reason to buy the current year's CPU. Sure, I recently upgraded from 233 MHz to 2.4 GHz, but that was mostly because my old motherboard fried itself and 2.4 GHz was $10 more than 2 GHz. It's an embarrasingly large amount of CPU, and the only reasons that I perceive it as slow are that Windows XP isn't very good at switching users (probably more RAM would help), plus I'm usually burning CPU with Folding@Home and running BitTorrent uploads in the background. When I get to use it as a geek machine and boot it into Linux, it's far faster than it needs to be, even if I'm running bloatware window managers.
Sure, my office has air conditioning, and PC power consumption really does make a difference there, but my home doesn't need it except maybe 1-2 weeks a year (and doesn't get it then.) Fans and high ceilings do the job just fine (or if they don't, I _could_ go to the office instead of telecommuting.) And I've already got electric heat, so my heat bills don't go up significantly just because the heat's coming from a computer instead of a baseboard heater. (
Noise, on the other hand, is a real issue, and having a silent or very quiet machine would be a pleasant change.
A while back I received some mail about it. At this point I forget if it was paper junk mail or email spam, I think paper, but if you had registered contact information with Microsoft when you bought your software during the appropriate years, I think they probably had to use that. The mail wasn't done very well in terms of identifying who really sent it, but it did look like it probably came from some internet-clueless law firm that probably really was managing that part of the class action lawsuit, as opposed to some random scammer (of course, I've also received lots of those:-) It had some web sites listed for more information, and of course you couldn't tell from the web sites or the whois records for the domain names if they were the genuine article, but it looked like they were.
Why I declined to participate: The anti-trust lawsuit purports that Microsoft did something wrong by becoming big and successful at selling people software and was so mean and nasty and twisted everybody's arms and that they ought to be spanked for doing so. That's a bogus ripoff of Microsoft. They were allegedly especially mean and nasty because they gave away Internet Exploiter for FREE, interfering with the sales of those nice friendly Netscape people who used to give us Netscape for free, and that people wouldn't install Netscape instead of using the pre-installed IE, in spite of the fact that we used to happily install Netscape. (Do *you* use IE? No? Then don't complain. Or, if yes, then don't complain.) Anti-trust suits are almost always a combination of appeals to blatant greed and a tool for some companies to use government to attack their bigger competition (though often it's a bad way to balance out help that the government has been giving the bigger competition), and this was no exception. If the State of California thinks for some reason that Microsoft ought to be spanked, they know where to buy Macintoshes or Linux or {Free/Net/Open/etc.}BSD, and the BSD source code even has lots of acknowledgements about University of California copyrights that they can feel proud of.
Sure, Microsoft operating systems cost too much and the quality is too low. It's not like they fooled me by selling me the stuff - the price tag was on the box and it said "Microsoft" in big friendly letters, so I pretty much knew what the quality was going to be like. (The one exception was that I bought a Win98SE upgrade edition specifically because Internet Connection Sharing was supposed to let me actually share my Internet connection, and it was way too broken to actually do that, but umm, mumble, that just means that I didn't feel any guilt about installing it on a couple of Win95 machines, so I think we probably broke even on that one...) Did I feel annoyed about having to buy WinXP to install on my mother-in-law's machine because she'd lost the original Win98 installation disk and "helpful" backup software that Compaq had added had left the machine too hosed to successfully clean up after we'd removed all the random spyware and IE browser extensions and invasive popup things that had accumulated in the year since we'd last cleaned it up? Yup. Bought it anyway, installed it and installed a new copy of AOL, and she's back to happily IMing with her friends and forwarding cutesy little greeting cards, and at least the Compaq stuff is gone:-) Did I feel annoyed about having to buy XP for my home Windows machine because my Windows ME CD wasn't bootable any more? Yes, but Fry's lets you buy the OEM version cheap if you're buying computer hardware, like the $5 fan with blinky-lights, and I didn't feel all that bad about not running WinME any more... And you can think of the price of the operating system as really being part of the price of TurboTax, because that's what I really use Windows for.
So while I don't particularly like most of Microsoft's products, I don't think the State has any business telling me that I was stupid for buying them and deserve my money back, because there was no dishonesty involved. (Y'all can tell me I was stupid for buying them, but you won't be telling me anything new, and The State of California Anti-Trust Lawyers Are Known To Be Harmful To Software Businesses.)
IM? Nah, the other guy would need to read, unless there's a standard for IM-to-Text-to-speech. It's a job for VOIP. But you really do want to be able to tell the idiot to do something less idiotic, and you need some kind of naming schema to support it. The privacy-invasive mechanism, of course, is to broadcast your license plate, True Name, driver's licenses and passports read by RFID right out of your wallet by the sensor on the seatback, etc., or at least have some kind of unique identifier built into the MAC address. But that requires other drivers' User Interfaces to present too much info to process quickly - the system really needs to support a broadcast interface with multiple naming structures, so you can yell "Hey, Red Beamer! Move over!" and have your car VOIP the message to the nearest car that accepts a broadcast message for a red BMW, and also so you can yell "Hey Everybody" and have everybody nearby get your message. (OK, technically you want to have the car's voice recognition system detect whether you're on a cellphone, and also interpret "Oh, Sh1t!" as a request to broadcast to everybody.) It'd be nice to also be able to yell "Hey, Stupid!" and have the car figure out who's acting stupid, but that takes Artificial Intelligence as opposed to mere voice recognition:-)
IPv6 is an obvious choice for a protocol, because it supports autoconfiguration based on a network number (which could be standardized for this application) and a MAC address (your wireless card's), so IP applications can work. There's still the problem of protecting privacy, since you don't want to be broadcasting ARPs as you drive by every detector on the street - in reality, the Home Office and Homeland Security and KGB and Stasi and Local ticket-revenue-generating cops will all want it, so it'll probably happen if this sort of thing gets deployed, but people should start thinking about what kinds of protocols can do the useful work while still protecting privacy. Downloadable MACs and randomly settable IP addresses, or something like them, are definitely needed, and setting up the car to answer broadcasts for its address as opposed to always advertising it is pretty much required. But what else do you need? What requirements are legitimate, or interesting? How do you prevent spamming (e.g. drive-by ads)?
Tanks are fine for semi-open country, at least when nobody's throwing tactical nukes or shooting with depleted uranium shells, but they have some vulnerabilities in cities. During the 1956 Hungarian Revolution, the Russians were invading Budapest, and the locals would take out a strip of cobblestone street, pour gasoline in it, and light in when a tank crossed, with negative results to the tank. Probably newer tanks are a bit better protected, but it's still a risk.
Bram's distinctly not responsible for any copyright violations, any more than the people who created ftp or implementations of ftp servers are. But Bittorrent is heavily used not only because it's a great tool for hauling around movies and other CD-sized data, but also because there are sites that host torrents for that kind of content. And unlike Bram's work, their work is enough closer to abetting copyright infringement that they can be sued (and this weekend's news is that Supernova and a couple of other sites just shut down.) They may or may not be guilty, but they're at least close enough to sue.
Here in Silicon Valley we almost never have that silly white stuff falling from the sky, and when we do it usually doesn't stick around very long, especially because it usually gets rained on, but I seem to remember from when I lived in other parts of the world that it sticks around on roofs for a while. And solar cells probably don't like having you climb around on them with a broom.
When I was in college, there was a leak in the watercooling system for the mainframe - it sprayed all over the motherboard and it was down for a couple of weeks. (Imagine a Beowulf cluster of IBM 360s:-) We also lost service for a week or two when they tried to add the fourth megabyte of RAM to the mainframe, which had previously only had 3 Mbps. Seems like small potatoes now, but back in the mid 70s that was a lot - being able to define a 1 Megabyte virtual machine let us solve much bigger math and engineering problems that we'd been able to before then, and we didn't waste much of that RAM on GUIs (though most people used PL/C, the local PL/I variant, which was a bit more bloated than the Watfor and Watfiv versions of F0RTRAN and had much better print statements.)
My work PC runs Win2K in "You're not the Administrator" mode, with a software installer widget that sometimes works. When Mozilla updates come out, they've got support for all the plugins and don't seem to lose them, so I can upgrade it in one swell foop and it Just Works. Firefox upgrades seem to lose information, require reinstallation of plugins and lots of retraining, and are basically much more trouble.
For a while I switched back to Mozilla on my home machine also, because Firefox was crashing way too often, but it's getting a bit better and Mozilla's occasionally crashing now.
One reason Ian and Nikita are "rockstars" is that they've been successful at cracking a number of important cryptographic algorithms or communications protocols that weren't designed carefully or weren't designed in a way that could be implemented unambiguously or were implemented poorly. Here they're not only trying to provide enough detail that you can implement the protocol in actual code, they're also trying to provide enough description that you can tell if the protocol's buggy.
For instance, in your suggestion, how do you "require your peer to disclose" anything? What if he doesn't comply? You can hang up, but you're already busted.... Meanwhile, if you've disclosed the keys, or if you're using a protocol that requires both players to know the key in order to verify a message, the other guy can't claim that you signed your message with information only you could have known and he can't - the information you're using to sign the message is information that only you and he and optionally anybody you or he have disclosed it to can know, so he can validate for himself that you wrote the message but he can't prove to anybody else that you wrote it and he didn't.
My wife's previous company valued the options that way when they issued them to employees, and we had to use the of that calculation when she exercised the options as for alternative minimum tax calculations. (She'd bought some at 10 cents, some 25 cents, etc., but the VC valuation was $3.00.) But they never ended up going public, so those options never were really worth that much as stock, and eventually they gave us 5 cents a share for them when they sold off the last bits of the company.
GPS works just fine for jetliners, and these blimps will probably blow around a bit (and will probably use GPS to keep themselves roughly in one place.)
Also, to navigate, you need to see at least two of them if you can measure angles precisely, or three if you can't but you're at a known altitude, or four if you're not at a known altitude. You're typically only going to be able to see one of these at a time, because they're not satellites, they're near-ground.
I did, in fact, read the end of the article:-) The risk with unsubscribing isn't whether they'll ignore your efforts to unsubscribe - it's that some fraction of them will sell the lists to other spammers and the other spammers will spam you more. In my case, there's probably not much downside risk - I've had my main email address for a decade, and it's splattered all over mailing list archives where harvesters regularly find it.
I have had spammers remove my names from lists - Scotty Richter's OptInRealBig gang were particularly diligent about it, since they were trying to promote an image of legitimacy and responsibility. (yeah, right...) I didn't directly request that they unsubscribe, but I sent spam complaints to some of their ISPs from one of my addresses, and that address showed a noticable drop in spam while the other addresses they were spamming didn't. I didn't bother complaining to their direct ISP, which was an obvious front, but I complained to their upstream, who should have known better than to "send complaints to them for resolution", but eventually got the hint that they really needed to get rid of the problem spammer, not just get the problem spammer to remove addresses.
Well, duh, a "legitimate" company is one that doesn't spam people, but actually got your name on their list by mistake.
The real question is how you recognize a "lazy" company, which doesn't bother selling your unsubscribed name to other spammers. Unfortunately, the main method I know for doing that is to use their unsubscribe link and see if you get spam. You can get fancy and unsubscribe your dummy spambait address, but many of the unsubscribe links I've seen have your email information encoded in them rather than listed transparently, e.g. "http://spammer.biz/unsubscribe/3485093285489035" as opposed to "http://spammer.biz/unsubscribe/addr=yourname@exam ple.net" so they're really just using the URL and not the address you unsubscribed from.
Columbia's announcement says they've increased their internet connectivity to ~300 Mbps, from their previous 150 Mbps. Seems kind of wimpy - I'd have expected an announcement that says they're running at least gigabit connections to the other participants in the study. Maybe they're doing that and the press release just doesn't mention it? Or did they just start cheap and reuse their existing OC3 cards, while maybe some of the other players have fast connections to each other?
If you wade through the piles of documentation, it looks like they've got dark fiber routes from each of the participants to racks at a couple of hub locations where they can meet with each other and Nysernet and also crossconnect to other carriers at a carrier-neutral facility. That means they could be running whatever combination they want of DWDM or CWDM, 10 Gig Ether, 1 Gig Ether, or traditional SONET (155 Mbps x 1,4,16,64) depending on how much they want to spend on CPE. I couldn't tell how many fiber pairs they were deploying per customer, but they're using fairly new high-end fiber that supports almost anything. The cheapest way to light up the stuff is with GigE fiber connections, since you can get by with a pretty small router, and cheap cards for short-distance hops, but CWDM is coming down in price (Coarse Wavelength Division Multiplexing doesn't get as many channels per fiber as Dense Wavelength Division Multiplexing, but the hardware's a lot cheaper) so you should be able to run multiple GigEs or whatever else you feel like. It looks like hardware costs for the CWDM versions are on the order of $5-10K per FDX GigE channel.
In case it's not slow enough, it does run on Bochs, so you could probably running on Linux running on your X-Box, and that might get you USB support:-)
Yes, IE is buggy, dangerous, and unreliable. Yes, MS encourages users to write IE-specific web pages that can't be rendered on other browsers. Yes, IE tolerates some kinds of buggy HTML that better browsers reject, and IE-oriented web publishers publish buggy pages and don't care.
And yes, reading Slashdot is much easier on a tabbed web browser, so IE loses. But at least IE appears to render all of Slashdot's web pages correctly the first time, without the need for kluges like hitting reload N times or trying to resize the text twice. Arrrgh!
I only had to hit reload once to get Mozilla to render this comment-posting page well enough to post this comment. Aaaargh!
Slashdot Mirror
A couple decades ago, my church in New Jersey built a building (we'd been without one for a while.) "Somehow" the township bureaucrats "forgot" to tell the county road department that we wanted to use the nearest sewer connection, so when they paved the street they didn't leave us access to it, and we're not allowed to break open the road for five years. "What a shame - you'll just have to connect to the next connection a couple hundred feet away and pump uphill." So we got a bid of $4000 to do this basically unnecessary work, and found that "No, what a shame, that contractor's not approved by *our* township, you've got to use an *approved* contractor", and it seems the two approved contractors wanted $12K and $25K for it. At that point our building committee guy started talking about going to the town council and maybe the state about it and walked out, and the building department met him out in the parking lot and said they might be able to come to some arrangement about it. They never explicitly asked for a bribe, and we absolutely never volunteered to pay one, which had a lot to do with why they'd caused this problem in the first place, though I suspect our contractor might have had to kick back a bit of money in return for a special permit. And everything else we did with the building got lots of extra inspection, and we ended up having to buy an amazingly fancy fire alarm system to handle the new regulations that kept being discovered, because we never did offer them a bribe or ask them which contractors we "really ought to" be dealing with.
So yes, local governments absolutely will reject deals that might be "profitable", and yes, they can stop utilities (regulated or unregulated) from using existing right-of-way, and even if they're not personally corrupt or greedy for town government revenues, they're still typically not competent at telecomm.
In this case, I don't think the article has much depth to it - the main concept is appealing, but I don't see enough thought behind it to really win. But even so, I'd mod you -1 Flamebait :-)
The real differences are between "X terminal" approaches and "diskless workstation" and "local storage" approaches, and local storage seems to keep winning.
On the other hand, I lived in California for a year in the late 70s, back when they were building those electrically-heated buildings, in a gas-heated one-bedroom apartment that never used more than $13 for gas+electricity in the worst of the winter, which was less than the electric bill alone for the gas-heated two-bedroom townhouse in New Jersey the next year.
According to the article, they're talking about spending $175 billion to build this edifice, but it's only going to bring in $135 billion in toll revenue over 50 years. That means that either the economics fail, and a toll road company would be unlikely to invest in it, or else they'd have to pour in at least $40B in tax funding, possibly a lot more if they're doing big favors for their politically-well-connected good old boys in the toll road business. Surprisingly, that's only $2000 per Texan worth of pork barrel, until you get to the hidden costs that probably inflate it way above the $175B costs or find that the toll revenues aren't close to the predicted $135B, so the shortfall's even larger.
But yes, yesterday's supercomputer, the Cray-1, was about the speed of a Pentium-133. Too slow to run Microsoftware on these days, though if you ran Win95 and Netscape 2 and Excel 5 you'd probably think it was pretty responsive. And X Windows ran ok on my 33 MHz 386 box, though Open Look was a bit slow, so if you want a modern Linux on a P133 (which you do) you probably need a bunch of extra RAM and a lightweight window manager - fvtwm or blackbox or something is a lot friendlier than KDE/Gnome.
But for home, since I'm not a gamer, I've never had a good reason to buy the current year's CPU. Sure, I recently upgraded from 233 MHz to 2.4 GHz, but that was mostly because my old motherboard fried itself and 2.4 GHz was $10 more than 2 GHz. It's an embarrasingly large amount of CPU, and the only reasons that I perceive it as slow are that Windows XP isn't very good at switching users (probably more RAM would help), plus I'm usually burning CPU with Folding@Home and running BitTorrent uploads in the background. When I get to use it as a geek machine and boot it into Linux, it's far faster than it needs to be, even if I'm running bloatware window managers.
Noise, on the other hand, is a real issue, and having a silent or very quiet machine would be a pleasant change.
Why I declined to participate: The anti-trust lawsuit purports that Microsoft did something wrong by becoming big and successful at selling people software and was so mean and nasty and twisted everybody's arms and that they ought to be spanked for doing so. That's a bogus ripoff of Microsoft. They were allegedly especially mean and nasty because they gave away Internet Exploiter for FREE, interfering with the sales of those nice friendly Netscape people who used to give us Netscape for free, and that people wouldn't install Netscape instead of using the pre-installed IE, in spite of the fact that we used to happily install Netscape. (Do *you* use IE? No? Then don't complain. Or, if yes, then don't complain.) Anti-trust suits are almost always a combination of appeals to blatant greed and a tool for some companies to use government to attack their bigger competition (though often it's a bad way to balance out help that the government has been giving the bigger competition), and this was no exception. If the State of California thinks for some reason that Microsoft ought to be spanked, they know where to buy Macintoshes or Linux or {Free/Net/Open/etc.}BSD, and the BSD source code even has lots of acknowledgements about University of California copyrights that they can feel proud of.
Sure, Microsoft operating systems cost too much and the quality is too low. It's not like they fooled me by selling me the stuff - the price tag was on the box and it said "Microsoft" in big friendly letters, so I pretty much knew what the quality was going to be like. (The one exception was that I bought a Win98SE upgrade edition specifically because Internet Connection Sharing was supposed to let me actually share my Internet connection, and it was way too broken to actually do that, but umm, mumble, that just means that I didn't feel any guilt about installing it on a couple of Win95 machines, so I think we probably broke even on that one...) Did I feel annoyed about having to buy WinXP to install on my mother-in-law's machine because she'd lost the original Win98 installation disk and "helpful" backup software that Compaq had added had left the machine too hosed to successfully clean up after we'd removed all the random spyware and IE browser extensions and invasive popup things that had accumulated in the year since we'd last cleaned it up? Yup. Bought it anyway, installed it and installed a new copy of AOL, and she's back to happily IMing with her friends and forwarding cutesy little greeting cards, and at least the Compaq stuff is gone :-) Did I feel annoyed about having to buy XP for my home Windows machine because my Windows ME CD wasn't bootable any more? Yes, but Fry's lets you buy the OEM version cheap if you're buying computer hardware, like the $5 fan with blinky-lights, and I didn't feel all that bad about not running WinME any more... And you can think of the price of the operating system as really being part of the price of TurboTax, because that's what I really use Windows for.
So while I don't particularly like most of Microsoft's products, I don't think the State has any business telling me that I was stupid for buying them and deserve my money back, because there was no dishonesty involved. (Y'all can tell me I was stupid for buying them, but you won't be telling me anything new, and The State of California Anti-Trust Lawyers Are Known To Be Harmful To Software Businesses.)
IPv6 is an obvious choice for a protocol, because it supports autoconfiguration based on a network number (which could be standardized for this application) and a MAC address (your wireless card's), so IP applications can work. There's still the problem of protecting privacy, since you don't want to be broadcasting ARPs as you drive by every detector on the street - in reality, the Home Office and Homeland Security and KGB and Stasi and Local ticket-revenue-generating cops will all want it, so it'll probably happen if this sort of thing gets deployed, but people should start thinking about what kinds of protocols can do the useful work while still protecting privacy. Downloadable MACs and randomly settable IP addresses, or something like them, are definitely needed, and setting up the car to answer broadcasts for its address as opposed to always advertising it is pretty much required. But what else do you need? What requirements are legitimate, or interesting? How do you prevent spamming (e.g. drive-by ads)?
Tanks are fine for semi-open country, at least when nobody's throwing tactical nukes or shooting with depleted uranium shells, but they have some vulnerabilities in cities. During the 1956 Hungarian Revolution, the Russians were invading Budapest, and the locals would take out a strip of cobblestone street, pour gasoline in it, and light in when a tank crossed, with negative results to the tank. Probably newer tanks are a bit better protected, but it's still a risk.
Sure, in Soviet Russia, giant mecha robots build humans in their backyards, but basically, giant mecha robots are popular in Japan!
Bram's distinctly not responsible for any copyright violations, any more than the people who created ftp or implementations of ftp servers are. But Bittorrent is heavily used not only because it's a great tool for hauling around movies and other CD-sized data, but also because there are sites that host torrents for that kind of content. And unlike Bram's work, their work is enough closer to abetting copyright infringement that they can be sued (and this weekend's news is that Supernova and a couple of other sites just shut down.) They may or may not be guilty, but they're at least close enough to sue.
Here in Silicon Valley we almost never have that silly white stuff falling from the sky, and when we do it usually doesn't stick around very long, especially because it usually gets rained on, but I seem to remember from when I lived in other parts of the world that it sticks around on roofs for a while. And solar cells probably don't like having you climb around on them with a broom.
When I was in college, there was a leak in the watercooling system for the mainframe - it sprayed all over the motherboard and it was down for a couple of weeks. (Imagine a Beowulf cluster of IBM 360s :-) We also lost service for a week or two when they tried to add the fourth megabyte of RAM to the mainframe, which had previously only had 3 Mbps. Seems like small potatoes now, but back in the mid 70s that was a lot - being able to define a 1 Megabyte virtual machine let us solve much bigger math and engineering problems that we'd been able to before then, and we didn't waste much of that RAM on GUIs (though most people used PL/C, the local PL/I variant, which was a bit more bloated than the Watfor and Watfiv versions of F0RTRAN and had much better print statements.)
For a while I switched back to Mozilla on my home machine also, because Firefox was crashing way too often, but it's getting a bit better and Mozilla's occasionally crashing now.
For instance, in your suggestion, how do you "require your peer to disclose" anything? What if he doesn't comply? You can hang up, but you're already busted.... Meanwhile, if you've disclosed the keys, or if you're using a protocol that requires both players to know the key in order to verify a message, the other guy can't claim that you signed your message with information only you could have known and he can't - the information you're using to sign the message is information that only you and he and optionally anybody you or he have disclosed it to can know, so he can validate for himself that you wrote the message but he can't prove to anybody else that you wrote it and he didn't.
My wife's previous company valued the options that way when they issued them to employees, and we had to use the of that calculation when she exercised the options as for alternative minimum tax calculations. (She'd bought some at 10 cents, some 25 cents, etc., but the VC valuation was $3.00.) But they never ended up going public, so those options never were really worth that much as stock, and eventually they gave us 5 cents a share for them when they sold off the last bits of the company.
Also, to navigate, you need to see at least two of them if you can measure angles precisely, or three if you can't but you're at a known altitude, or four if you're not at a known altitude. You're typically only going to be able to see one of these at a time, because they're not satellites, they're near-ground.
(Need to replace Verisign/Netsol's broken proposals for internationalised domain names with something better to support se~nor.mobi ....)
I have had spammers remove my names from lists - Scotty Richter's OptInRealBig gang were particularly diligent about it, since they were trying to promote an image of legitimacy and responsibility. (yeah, right...) I didn't directly request that they unsubscribe, but I sent spam complaints to some of their ISPs from one of my addresses, and that address showed a noticable drop in spam while the other addresses they were spamming didn't. I didn't bother complaining to their direct ISP, which was an obvious front, but I complained to their upstream, who should have known better than to "send complaints to them for resolution", but eventually got the hint that they really needed to get rid of the problem spammer, not just get the problem spammer to remove addresses.
The real question is how you recognize a "lazy" company, which doesn't bother selling your unsubscribed name to other spammers. Unfortunately, the main method I know for doing that is to use their unsubscribe link and see if you get spam. You can get fancy and unsubscribe your dummy spambait address, but many of the unsubscribe links I've seen have your email information encoded in them rather than listed transparently, e.g. "http://spammer.biz/unsubscribe/3485093285489035" as opposed to "http://spammer.biz/unsubscribe/addr=yourname@exam ple.net" so they're really just using the URL and not the address you unsubscribed from.
If you wade through the piles of documentation, it looks like they've got dark fiber routes from each of the participants to racks at a couple of hub locations where they can meet with each other and Nysernet and also crossconnect to other carriers at a carrier-neutral facility. That means they could be running whatever combination they want of DWDM or CWDM, 10 Gig Ether, 1 Gig Ether, or traditional SONET (155 Mbps x 1,4,16,64) depending on how much they want to spend on CPE. I couldn't tell how many fiber pairs they were deploying per customer, but they're using fairly new high-end fiber that supports almost anything. The cheapest way to light up the stuff is with GigE fiber connections, since you can get by with a pretty small router, and cheap cards for short-distance hops, but CWDM is coming down in price (Coarse Wavelength Division Multiplexing doesn't get as many channels per fiber as Dense Wavelength Division Multiplexing, but the hardware's a lot cheaper) so you should be able to run multiple GigEs or whatever else you feel like. It looks like hardware costs for the CWDM versions are on the order of $5-10K per FDX GigE channel.
In case it's not slow enough, it does run on Bochs, so you could probably running on Linux running on your X-Box, and that might get you USB support :-)
And yes, reading Slashdot is much easier on a tabbed web browser, so IE loses. But at least IE appears to render all of Slashdot's web pages correctly the first time, without the need for kluges like hitting reload N times or trying to resize the text twice. Arrrgh!
I only had to hit reload once to get Mozilla to render this comment-posting page well enough to post this comment. Aaaargh!