The stuff that cable head ends use costs around $10k/channel. Or more, depending on the model. Yeah, I know it is sometimes done cheaper, but broadcast-grade gear isn't cheap. I think it could be done (at consumer grade) cheaply enough in a cable box, if it were for say one channel at a time, and mass-produced.
My original point was that the students didn't take everything they should have into account.
The students' proposal was for broadcasting unused, as well as watched, channels over the 802.11g. What you suggest is more sensible, but would require the cable box to act as a smarter server. I was pointing out that the student paper did not address this issue appropriately.
In North America, channels are 6 Mbps wide; 8 Mbps is European-standard, both video and DOCSIS. Digital Cable takes a 6 Mbps channel, runs QAM in it to get about 20 Mbps, and then feeds about ten MPEG streams through it (bit rate of each can be adjusted depending on content).
Real cable is typically a mix of analog and digital channels. Digitizing is costly, so putting analog channels onto digital wireless would take some effort.
The proposal has everything on the cable sent over 802.11g all the time, so even TVs that aren't on will receive the signal. Well, lessee, a cable can support about a hundred (depending on the plant) downstream channels, which digitized are around 30 mbps apiece, so that's 3 Gbps... sure, that'll fit into 802.11g, not.
And the 2.4 GHz band is already congested. It's home of the microwave oven, for one thing, so reception would be mighty bad while the corn is popping. The last thing it needs is a wideband access point in every home.
First off, I do this type of thing for a living, as a consultant to the CLEC and ISP trades, so I know a thing or two about the DSL market. Please, please, ignore the consensus of the Slashdot crowd who want you to pull Ethernet! They imagine that they'd want the better speed, but as a provider, you have to face reality. DSL has real advantages:
1) It lets you control the top speed. I suggest that the top speed to a user be less than half of your feed speed. A company I work very closely with has almost 200 DSL lines in a luxury condominium. They feed it with only two T1s. That's quite adequate! They have to pay for that bandwidth -- backbone ISP service isn't cheap, and the T1 loops into the condo aren't free either. Of course they only provide 700 kbps service. Sure, people might like more, but the competition is dial-up, and price matters.
2) DSL tolerates long wire. It can go a few miles, after all -- even a sprawling condo complex is a short hop for DSL. Ethernet tends to be pickier.
3) ADSL can share wire with telephone. You might be able to piggyback onto the phone wire. (A CLEC can; whether you can is a different issue.)
4) DSL is cheap! Lots of providers tanked, leaving good working gear on the secondary market. A 500-line Lucent Stinger can be had for $12k; a 200-line ADSL DSLAM is maybe half that. SDSL needs its own wire pair (can't share phone like ADSL) but the DSLAMs are a glut on the market, much cheaper than even that. Check eBay, telephone.com, etc.
I'd be happy to talk more about this offline (isdnip at netscape dot net)....
The article was half right. SMTP is in need of replacement. But so is the entire TCP/IP suite.
Sure, most Internet users simply assume that it's good, because the Internet is cool, and uses it. But TCP/IP was a lab research project from the 1970s, designed for closed government networks with a small number of time-sharing computers on it. Misbehavior could be dealt with easily, because connections were not open to the public. And the backbone links went at 50 kbps; most sites got on at 9600 bps.
People are now using TCP/IP for anything and everything. Voice, video, radio, spam. It's flexible enough to handle it all, but not efficiently! There are many technical flaws in the protocol suite. IPv6, btw, does nothing to fix it; it just makes matters worse by having even more overhead. NATs today are a security feature, not a bug; apps like FTP that put the address in the application layer (hard on NATs) are BROKEN! BTW, FTP did that because it saved a little code in the Pluribus IMP print routine in 1973. Don't know what that was? Good -- but don't foist its workarounds on the future.
A new protocol suite should be developed that handles today's high speeds (as well as slow links, which will always exist), resists spam and identity spoofing, allows multihoming, handles voice and streaming with connected-mode QoS, and doesn't have TCP/IPs overhead. It can be done. Stamp out TCP/IP fundamentalism!
Okay, so the previous article had "Intel" in the title, but it covered the same ground. Five days ago. I guess on/. that means it's forgotten.
And what I said there is still true. 802.16 is a technology spec, not a frequency band. Long range and high speed are for licensed users. Unlicensed users get short range (5 GHz band). And the 25+ GHz frequencies are very sensitive to rain fade. Even with high licensed power, most non-desert areas will have noticeable outages several times a year during strong rain on any link over 2-3 miles.
WiMAX is a technique, not bound to a specific frequency like 802.11b/g/a etc. It is mainly aimed at licensed providers. It could be used in the LMDS (29 GHz) range, where reliable coverage is 2-3 miles (because of rain fade), at 39 GHz (similar), or at other licensed frequencies. It could be used in the licensed 2.6 GHz MMDS band, whose rules are the topic of a currently-open FCC docket in the USA. (Among the options is opening up some more unlicensed spectrum there.) Or it could be used, with low power and thus lower range, in the 5.8 GHz unlicensed band.
The technology is better for a public MAN than any of the 802.11 family. But let's not expect miracles from shared, unlicensed frequencies.
The original UToronto release talks about evanescent waves, apparently a fairly critical part of the equation, and leads to the conclusion that the laws of physics are not actually being broken. Rather, the whole idea is that it is possible to create a lens with a negative index of refraction without anything exceeding the speed of light. Fancy footwork, yes, and perhaps still only a theoretical possibility rather than product nearly ready for sale. But not quite as dramatic as it sounds.
Note that the initial article came from a British paper, indeed a very good one which I, like many on the left side of the puddle, read regularly. The article refers to Office Depot's UK stores, which have adopted the policy. It also suggests that US stores haven't yet done so, though they might at some point in the future.
I don't know how autonomous the different Office Depot divisions are, but many companies give a lot of autonomy to national divisions.
I sit two doors away from a Staples so I don't really go into an Office Depot much anyway....
Roxio's "Napster" isn't the real thing. It won't have any peer-to-peer. It'll be just another resale of MusicNet and/or Pressplay. "Napster" will have downloads for computer-only play that expire when the subscription is no longer paid up. There will be premium-priced "burns" for a price that makes CDs look cheap, but competitive for "singles" (e.g., around a buck a track). Roxio's value is to integrate it better into Roxio's software.
Napster Fanning himself? He's just a figurehead. George Foreman does more appliance design than Fanning will do with the actual running of this service.
Of course it'll flop, but that's what the record industry wants.
No, mr. Subsidized Telco. The problem is not that Covad can't do voice. The problem is that the ruling is not about voice. It says that the CLEC has to rent the FULL loop, not just the high frequencies which the ILEC DSL operation uses.
In non-Bell territories, the full loop often costs $30-$100 PER MONTH! Verizon Illinois is $42/month for a plain old copper pair. Compare to around $3 in Chicago for Ameritech (typical Bell ranges are $7-15/month, somewhat higher in a few rural areas).
So what this ruling does is take away DSL from rural areas. Sounds like CenturyTel (scum-o'-the-earth) or the like got to Martin big time.
More complex than that. Democrat Copps specifically said in his statement that he disagreed. But he did not make it a *formal* dissent, which would have turned it 3-2 the other way, because he horse-traded his vote on DSL line sharing for other changes Martin wouldn't have given him. Probably UNE Platform, which is what MCI's The Neighborhood is, and which Powell wanted to kill fast.
Some of the press, in their post-9/11 fealty to King George the Commander-in-Thief, reported in headlines that "Bush won the recount". But the fine print said that it was his *if* they followed Boies'/Gore's recommendation of recounting Democratic counties, or following certain standards. Had the entire state been recounted by a common standard, as the Republicans had originally seemed to accept but Boies didn't, then Gore would have won.
Yes, confusing. I am going by memory but mainly the point I remember was that the Democrats guessed wrong what strategy would have won.
You've got the right idea. SMTP is woefully obsolete. It was invented for a closed-to-the-public ARPAnet. Woe befell the idiot DEC salesbozo who invented spam when he sent a new-product announcement to *@*! (That was before DNS; with the HOSTS table, it worked.)
What's needed is some kind of "digital postage stamp", voluntarily issued among ISPs and users (not the postal authorities, so please don't bring up mythical "Bill 602P"), which has to be there before mail gets relayed or, more importantly, *accepted*. No stamp, no receipt. Every retail ISP user will get hundreds of stamps a month, and bulk users can buy them (say, for a corporate email gateway) by the myriad, for something in the penny order of magnitude. That wouldn't be noticeable to anybody but a spammer, who depends on extreme volume.
The trick is to make it work securely without too much of a performance hit.
Ark Linux web site is a farce with pr0n on it
on
Ark Linux
·
· Score: 2, Informative
I'm surprised fewer Slashdotters even looked at Ark Linux' web site before engaging in the usual jabber about easy-to-use Linux distros. The site itself has no info about the distro. But if you go to the FAQ, lo and behold you get a bunch of references to "goatsex", an ASCII art dirty picture, a few hundred lines repeating the word "faggot", and a dirty story. Not much about Linux though.
I suspect the whole thing is a put-on by somebody with a serious lack of taste. Certainly if it were a real distro, and this stuff were hacked onto the site, then its server security wouldn't be adequate.
So what they're proposing, in effect, is a shift of money from Prius drivers to Hummer drivers. Go buy that huge Luxury Truck ("SUV" to the marketeers), folks; the price of gas will fall, and Ford Excursions will pay the same mileage-based rate as Honda Impacts.
This proposal is incredibly counterproductive. I think the gas tax should be raised, regularly (e.g., 5c/year), to discourage heavy consumption. And btw lightweight fuel-efficient vehicles wear out roads less than huge testosterone trucks.
I like his adjustable skis. But how do you set the speed? It's not practical to bend over and tweak a knob on top while going down a hill; at least, it would not be good style.
But if the control circuit used, say, Bluetooth, and the control buttons were in, say, the pole handles, then the skier would have good control while staying in motion. This could be quite nice.
(If somebody tries to patent the idea with a later date than today, remember, you heard it first on Slashdot, making it potentially Prior Art.)
Re:How About Permitting _Real_ Competition?
on
DSL Amidst Phone Wars
·
· Score: 5, Informative
Whoa! Let's clarify the rules -- it's a lot more subtle. More like Microsoft's monopoly than the old Bell System.
Anybody can open a local phone company nowadays. Lots of us have... but if you want to reach residential subscribers, you can't afford to pull new copper wire (old technology, very capital-intensive) to each house. That is a true "natural monopoly", which means that the cost to a competitor would be much higher than the cost to the incumbent, making competition unworkable.
However, there's no natural monopoly to various other aspects of the business. So for instance Bell competitors (CLECs) can rent the wire, at a regulated rate, and provide dialtone and/or DSL over it. Covad does this, for instance. The current regulations (which may change; the Bells are doing a full-frontal attack on them at the FCC, and chairman Powell's their lapdog) also require the Bells to rent their switches at wholesale to CLECs. So a CLEC can lease Bell copper loops and switches, and thus provide service with none of its own equipment. This is called the "UNE Platform" and is how most (though not all) non-cable residential telephone "competition" works. Note that if it's called UNE-P, the CLEC sets its own prices quite freely, vs. the declining-in-popularity so-called "total service Resale" where the CLEC is just taking a commission on Bell's regular rates. So UNE-P lets New York City and Chicago subscribers get flat rate service, without Zone charges; it powers MCI's "Neighborhood" too. Note that Bells are not required to provide DSL to CLECs as a wholesale service, so it isn't part of UNE-P.
The controversy: If SBC (or another Bell) provides the dialtone, then they will also sell their ADSL atop it. The price for residential ADSL is held down because the dialtone line is paying for the loop; SBC's own ADSL "business" gets the line for basically free. So can Covad -- they can rent the "high frequency element" of an SBC dialtone line for near zero. BUT if the SBC dialtone line is being provisioned on a UNE-P or Resale basis, so MCI or AT&T (etc.) is the end user's phone company of record, then SBC as a matter of policy chooses not to provide its DSL service. This is pure spite, not technology --the UNE-P line is identical to an SBC-service line, and the UNE-P CLEC is already paying for the loop.
The nice thing about UNE-P is that you can switch carriers without really touching anything -- it's a computer entry, of who gets billed how much by whom. But because SBC refuses to sell DSL atop UNE-P, they "lock in" voice subscribers by threatening to take away the ADSL. They're gambling that they'll make more money by keeping voice subscriber than they'll lose by having UNE-P subscribers switch to other DSL providers. And, more likely, they are just such monopolists at heart that they don't give a rat's ass about maximizing their own profit, if they can thumb their nose at a competitor.
This all has interesting antitrust connotations (no, Bells are not exempt from antitrust any more) but that will take years to play out.
It is worth stating again: ISPs are not common carriers.
Telephone companies are common carriers. ISPs, including cable modem operators, are technically "information service providers". A DSL ISP service has two separable components, the DSL (common carrier) and the ISP. This distinction matters in various areas of law and regulation.
Now it is true that legally, ISPs are often not responsible for content put out by users. But sometimes they are, and they have certain responsibilities. I don't think that Cablevision (ISP) is responsible for copyright violations by peer-to-peer subscribers, but they could be asked to "take down" certain users or content. This type of thing keeps lawyers busy.
Ricochet's best feature is that it is not 802.11. It runs at a slower speed, which allows it to have a longer range. I do find it odd that so many people are trying to blanket the country with Wi-Fi, a local area network technology. Well, of course Wi-Fi has the volume behind it, which makes the kit dirt cheap, but it's not really up to the job. Ricochet's range is nothing great; it needs a lot of lamp poles, but it is better suited to its purpose than Wi-Fi.
The poor performance you cite applies in the South Bay area, where Ricochet was first rolled out using low-speed 900 MHz radios. They later developed 2.4 GHz radios with a higher data rate. Thus the rough speed they were selling went from 28 to 128 kbps, with peak speeds higher.
A release candidate is a place to look for the last few critical bugs, no? In this case, the bugs that were going to be squashed were squashed, and everything was ready to go. They just had to do a final build and let it go.
But oops, the build process went awry, and the binaries were built against the wrong sources. So the RC process wouldn't have worked, because it would have put fixes into the tree, which would have been ignored in the erroneous build.
BTW, 1.1 was very unstable for me under Linux, while 1.2 (as released) is rock solid.
I know that sounds funny, but it's not implausible. Knoppix can install itself onto a hard drive using the knx-install script. It just asks a few questions, then does the install in a few minutes.
I suppose if you found the sources to that script (by Christian Perle, not Knopper himself), then you could modify it. I frankly wouldn't trash the store's WinXP, which would have serious legal repercussions; it'd however make sense to automagically repartition the drive, putting Linux after Windows, with a pretty lilo or grub boot selector.
BIDMC is a big place, too; two adjacent campuses (the old Beth Israel and Deaconess hospitals) and a lot of legacy stuff from pre-merger days. The articles are shy on details but from what I can tell, they had a mix of routable IP and non-routable protocols. The old ones (like LAT, or IPX if you don't route it) depend on bridging, and the routers try to be bridges too, and that's just not something they're good at.
Indeed, <b>bridging does not scale well</b>. Campus-wide (both campuses, actually) support for any non-routing protocol is hazardous to a network's health. It's tempting to have a little bridged network and just add a little more, and a little more, but when it tips, it tips fast.
But if you read the Felten papers about the SDMI crack, then it's understandable how they have a "strong" watermark that survives MP3 or other encoding. It's not lossless -- the watermark is in the audio, just not so prominent as to be psychoacoustically noticeable.
The basic idea, IIRC, was that it took narrow frequency bands in the high treble, and then raised and lowered adjacent bands (think: play with a graphic equalizer) rapidly in a pattern. A dB here, a dB there, and the detector can notice the changes in the relative levels. Your ears probably won't.
Slashdot Mirror
The stuff that cable head ends use costs around $10k/channel. Or more, depending on the model. Yeah, I know it is sometimes done cheaper, but broadcast-grade gear isn't cheap. I think it could be done (at consumer grade) cheaply enough in a cable box, if it were for say one channel at a time, and mass-produced.
My original point was that the students didn't take everything they should have into account.
The students' proposal was for broadcasting unused, as well as watched, channels over the 802.11g. What you suggest is more sensible, but would require the cable box to act as a smarter server. I was pointing out that the student paper did not address this issue appropriately.
In North America, channels are 6 Mbps wide; 8 Mbps is European-standard, both video and DOCSIS. Digital Cable takes a 6 Mbps channel, runs QAM in it to get about 20 Mbps, and then feeds about ten MPEG streams through it (bit rate of each can be adjusted depending on content).
Real cable is typically a mix of analog and digital channels. Digitizing is costly, so putting analog channels onto digital wireless would take some effort.
The proposal has everything on the cable sent over 802.11g all the time, so even TVs that aren't on will receive the signal. Well, lessee, a cable can support about a hundred (depending on the plant) downstream channels, which digitized are around 30 mbps apiece, so that's 3 Gbps... sure, that'll fit into 802.11g, not.
And the 2.4 GHz band is already congested. It's home of the microwave oven, for one thing, so reception would be mighty bad while the corn is popping. The last thing it needs is a wideband access point in every home.
First off, I do this type of thing for a living, as a consultant to the CLEC and ISP trades, so I know a thing or two about the DSL market. Please, please, ignore the consensus of the Slashdot crowd who want you to pull Ethernet! They imagine that they'd want the better speed, but as a provider, you have to face reality. DSL has real advantages:
1) It lets you control the top speed. I suggest that the top speed to a user be less than half of your feed speed. A company I work very closely with has almost 200 DSL lines in a luxury condominium. They feed it with only two T1s. That's quite adequate! They have to pay for that bandwidth -- backbone ISP service isn't cheap, and the T1 loops into the condo aren't free either. Of course they only provide 700 kbps service. Sure, people might like more, but the competition is dial-up, and price matters.
2) DSL tolerates long wire. It can go a few miles, after all -- even a sprawling condo complex is a short hop for DSL. Ethernet tends to be pickier.
3) ADSL can share wire with telephone. You might be able to piggyback onto the phone wire. (A CLEC can; whether you can is a different issue.)
4) DSL is cheap! Lots of providers tanked, leaving good working gear on the secondary market. A 500-line Lucent Stinger can be had for $12k; a 200-line ADSL DSLAM is maybe half that. SDSL needs its own wire pair (can't share phone like ADSL) but the DSLAMs are a glut on the market, much cheaper than even that. Check eBay, telephone.com, etc.
I'd be happy to talk more about this offline (isdnip at netscape dot net)....
The article was half right. SMTP is in need of replacement. But so is the entire TCP/IP suite.
Sure, most Internet users simply assume that it's good, because the Internet is cool, and uses it. But TCP/IP was a lab research project from the 1970s, designed for closed government networks with a small number of time-sharing computers on it. Misbehavior could be dealt with easily, because connections were not open to the public. And the backbone links went at 50 kbps; most sites got on at 9600 bps.
People are now using TCP/IP for anything and everything. Voice, video, radio, spam. It's flexible enough to handle it all, but not efficiently! There are many technical flaws in the protocol suite. IPv6, btw, does nothing to fix it; it just makes matters worse by having even more overhead. NATs today are a security feature, not a bug; apps like FTP that put the address in the application layer (hard on NATs) are BROKEN! BTW, FTP did that because it saved a little code in the Pluribus IMP print routine in 1973. Don't know what that was? Good -- but don't foist its workarounds on the future.
A new protocol suite should be developed that handles today's high speeds (as well as slow links, which will always exist), resists spam and identity spoofing, allows multihoming, handles voice and streaming with connected-mode QoS, and doesn't have TCP/IPs overhead. It can be done. Stamp out TCP/IP fundamentalism!
http://slashdot.org/article.pl?sid=03/04/08/195524 8&tid=193
/. that means it's forgotten.
Okay, so the previous article had "Intel" in the title, but it covered the same ground. Five days ago. I guess on
And what I said there is still true. 802.16 is a technology spec, not a frequency band. Long range and high speed are for licensed users. Unlicensed users get short range (5 GHz band). And the 25+ GHz frequencies are very sensitive to rain fade. Even with high licensed power, most non-desert areas will have noticeable outages several times a year during strong rain on any link over 2-3 miles.
WiMAX is a technique, not bound to a specific frequency like 802.11b/g/a etc. It is mainly aimed at licensed providers. It could be used in the LMDS (29 GHz) range, where reliable coverage is 2-3 miles (because of rain fade), at 39 GHz (similar), or at other licensed frequencies. It could be used in the licensed 2.6 GHz MMDS band, whose rules are the topic of a currently-open FCC docket in the USA. (Among the options is opening up some more unlicensed spectrum there.) Or it could be used, with low power and thus lower range, in the 5.8 GHz unlicensed band.
The technology is better for a public MAN than any of the 802.11 family. But let's not expect miracles from shared, unlicensed frequencies.
The original UToronto release talks about evanescent waves, apparently a fairly critical part of the equation, and leads to the conclusion that the laws of physics are not actually being broken. Rather, the whole idea is that it is possible to create a lens with a negative index of refraction without anything exceeding the speed of light. Fancy footwork, yes, and perhaps still only a theoretical possibility rather than product nearly ready for sale. But not quite as dramatic as it sounds.
Note that the initial article came from a British paper, indeed a very good one which I, like many on the left side of the puddle, read regularly. The article refers to Office Depot's UK stores, which have adopted the policy. It also suggests that US stores haven't yet done so, though they might at some point in the future.
I don't know how autonomous the different Office Depot divisions are, but many companies give a lot of autonomy to national divisions.
I sit two doors away from a Staples so I don't really go into an Office Depot much anyway....
Roxio's "Napster" isn't the real thing. It won't have any peer-to-peer. It'll be just another resale of MusicNet and/or Pressplay. "Napster" will have downloads for computer-only play that expire when the subscription is no longer paid up. There will be premium-priced "burns" for a price that makes CDs look cheap, but competitive for "singles" (e.g., around a buck a track). Roxio's value is to integrate it better into Roxio's software.
Napster Fanning himself? He's just a figurehead. George Foreman does more appliance design than Fanning will do with the actual running of this service.
Of course it'll flop, but that's what the record industry wants.
No, mr. Subsidized Telco. The problem is not that Covad can't do voice. The problem is that the ruling is not about voice. It says that the CLEC has to rent the FULL loop, not just the high frequencies which the ILEC DSL operation uses.
In non-Bell territories, the full loop often costs $30-$100 PER MONTH! Verizon Illinois is $42/month for a plain old copper pair. Compare to around $3 in Chicago for Ameritech (typical Bell ranges are $7-15/month, somewhat higher in a few rural areas).
So what this ruling does is take away DSL from rural areas. Sounds like CenturyTel (scum-o'-the-earth) or the like got to Martin big time.
More complex than that. Democrat Copps specifically said in his statement that he disagreed. But he did not make it a *formal* dissent, which would have turned it 3-2 the other way, because he horse-traded his vote on DSL line sharing for other changes Martin wouldn't have given him. Probably UNE Platform, which is what MCI's The Neighborhood is, and which Powell wanted to kill fast.
Expect to see this one in court, fast.
No; the previous guy was right.
Some of the press, in their post-9/11 fealty to King George the Commander-in-Thief, reported in headlines that "Bush won the recount". But the fine print said that it was his *if* they followed Boies'/Gore's recommendation of recounting Democratic counties, or following certain standards. Had the entire state been recounted by a common standard, as the Republicans had originally seemed to accept but Boies didn't, then Gore would have won.
Yes, confusing. I am going by memory but mainly the point I remember was that the Democrats guessed wrong what strategy would have won.
You've got the right idea. SMTP is woefully obsolete. It was invented for a closed-to-the-public ARPAnet. Woe befell the idiot DEC salesbozo who invented spam when he sent a new-product announcement to *@*! (That was before DNS; with the HOSTS table, it worked.)
What's needed is some kind of "digital postage stamp", voluntarily issued among ISPs and users (not the postal authorities, so please don't bring up mythical "Bill 602P"), which has to be there before mail gets relayed or, more importantly, *accepted*. No stamp, no receipt. Every retail ISP user will get hundreds of stamps a month, and bulk users can buy them (say, for a corporate email gateway) by the myriad, for something in the penny order of magnitude. That wouldn't be noticeable to anybody but a spammer, who depends on extreme volume.
The trick is to make it work securely without too much of a performance hit.
I'm surprised fewer Slashdotters even looked at Ark Linux' web site before engaging in the usual jabber about easy-to-use Linux distros. The site itself has no info about the distro. But if you go to the FAQ, lo and behold you get a bunch of references to "goatsex", an ASCII art dirty picture, a few hundred lines repeating the word "faggot", and a dirty story. Not much about Linux though.
I suspect the whole thing is a put-on by somebody with a serious lack of taste. Certainly if it were a real distro, and this stuff were hacked onto the site, then its server security wouldn't be adequate.
So what they're proposing, in effect, is a shift of money from Prius drivers to Hummer drivers. Go buy that huge Luxury Truck ("SUV" to the marketeers), folks; the price of gas will fall, and Ford Excursions will pay the same mileage-based rate as Honda Impacts.
This proposal is incredibly counterproductive. I think the gas tax should be raised, regularly (e.g., 5c/year), to discourage heavy consumption. And btw lightweight fuel-efficient vehicles wear out roads less than huge testosterone trucks.
I like his adjustable skis. But how do you set the speed? It's not practical to bend over and tweak a knob on top while going down a hill; at least, it would not be good style.
But if the control circuit used, say, Bluetooth, and the control buttons were in, say, the pole handles, then the skier would have good control while staying in motion. This could be quite nice.
(If somebody tries to patent the idea with a later date than today, remember, you heard it first on Slashdot, making it potentially Prior Art.)
Whoa! Let's clarify the rules -- it's a lot more subtle. More like Microsoft's monopoly than the old Bell System.
Anybody can open a local phone company nowadays. Lots of us have... but if you want to reach residential subscribers, you can't afford to pull new copper wire (old technology, very capital-intensive) to each house. That is a true "natural monopoly", which means that the cost to a competitor would be much higher than the cost to the incumbent, making competition unworkable.
However, there's no natural monopoly to various other aspects of the business. So for instance Bell competitors (CLECs) can rent the wire, at a regulated rate, and provide dialtone and/or DSL over it. Covad does this, for instance. The current regulations (which may change; the Bells are doing a full-frontal attack on them at the FCC, and chairman Powell's their lapdog) also require the Bells to rent their switches at wholesale to CLECs. So a CLEC can lease Bell copper loops and switches, and thus provide service with none of its own equipment. This is called the "UNE Platform" and is how most (though not all) non-cable residential telephone "competition" works. Note that if it's called UNE-P, the CLEC sets its own prices quite freely, vs. the declining-in-popularity so-called "total service Resale" where the CLEC is just taking a commission on Bell's regular rates. So UNE-P lets New York City and Chicago subscribers get flat rate service, without Zone charges; it powers MCI's "Neighborhood" too. Note that Bells are not required to provide DSL to CLECs as a wholesale service, so it isn't part of UNE-P.
The controversy: If SBC (or another Bell) provides the dialtone, then they will also sell their ADSL atop it. The price for residential ADSL is held down because the dialtone line is paying for the loop; SBC's own ADSL "business" gets the line for basically free. So can Covad -- they can rent the "high frequency element" of an SBC dialtone line for near zero. BUT if the SBC dialtone line is being provisioned on a UNE-P or Resale basis, so MCI or AT&T (etc.) is the end user's phone company of record, then SBC as a matter of policy chooses not to provide its DSL service. This is pure spite, not technology --the UNE-P line is identical to an SBC-service line, and the UNE-P CLEC is already paying for the loop.
The nice thing about UNE-P is that you can switch carriers without really touching anything -- it's a computer entry, of who gets billed how much by whom. But because SBC refuses to sell DSL atop UNE-P, they "lock in" voice subscribers by threatening to take away the ADSL. They're gambling that they'll make more money by keeping voice subscriber than they'll lose by having UNE-P subscribers switch to other DSL providers. And, more likely, they are just such monopolists at heart that they don't give a rat's ass about maximizing their own profit, if they can thumb their nose at a competitor.
This all has interesting antitrust connotations (no, Bells are not exempt from antitrust any more) but that will take years to play out.
It is worth stating again: ISPs are not common carriers.
Telephone companies are common carriers. ISPs, including cable modem operators, are technically "information service providers". A DSL ISP service has two separable components, the DSL (common carrier) and the ISP. This distinction matters in various areas of law and regulation.
Now it is true that legally, ISPs are often not responsible for content put out by users. But sometimes they are, and they have certain responsibilities. I don't think that Cablevision (ISP) is responsible for copyright violations by peer-to-peer subscribers, but they could be asked to "take down" certain users or content. This type of thing keeps lawyers busy.
Ricochet's best feature is that it is not 802.11. It runs at a slower speed, which allows it to have a longer range. I do find it odd that so many people are trying to blanket the country with Wi-Fi, a local area network technology. Well, of course Wi-Fi has the volume behind it, which makes the kit dirt cheap, but it's not really up to the job. Ricochet's range is nothing great; it needs a lot of lamp poles, but it is better suited to its purpose than Wi-Fi.
The poor performance you cite applies in the South Bay area, where Ricochet was first rolled out using low-speed 900 MHz radios. They later developed 2.4 GHz radios with a higher data rate. Thus the rough speed they were selling went from 28 to 128 kbps, with peak speeds higher.
A release candidate is a place to look for the last few critical bugs, no? In this case, the bugs that were going to be squashed were squashed, and everything was ready to go. They just had to do a final build and let it go.
But oops, the build process went awry, and the binaries were built against the wrong sources. So the RC process wouldn't have worked, because it would have put fixes into the tree, which would have been ignored in the erroneous build.
BTW, 1.1 was very unstable for me under Linux, while 1.2 (as released) is rock solid.
I know that sounds funny, but it's not implausible. Knoppix can install itself onto a hard drive using the knx-install script. It just asks a few questions, then does the install in a few minutes.
I suppose if you found the sources to that script (by Christian Perle, not Knopper himself), then you could modify it. I frankly wouldn't trash the store's WinXP, which would have serious legal repercussions; it'd however make sense to automagically repartition the drive, putting Linux after Windows, with a pretty lilo or grub boot selector.
BIDMC is a big place, too; two adjacent campuses (the old Beth Israel and Deaconess hospitals) and a lot of legacy stuff from pre-merger days. The articles are shy on details but from what I can tell, they had a mix of routable IP and non-routable protocols. The old ones (like LAT, or IPX if you don't route it) depend on bridging, and the routers try to be bridges too, and that's just not something they're good at.
Indeed, <b>bridging does not scale well</b>. Campus-wide (both campuses, actually) support for any non-routing protocol is hazardous to a network's health. It's tempting to have a little bridged network and just add a little more, and a little more, but when it tips, it tips fast.
I don't know how these are watermarked.
But if you read the Felten papers about the SDMI crack, then it's understandable how they have a "strong" watermark that survives MP3 or other encoding. It's not lossless -- the watermark is in the audio, just not so prominent as to be psychoacoustically noticeable.
The basic idea, IIRC, was that it took narrow frequency bands in the high treble, and then raised and lowered adjacent bands (think: play with a graphic equalizer) rapidly in a pattern. A dB here, a dB there, and the detector can notice the changes in the relative levels. Your ears probably won't.