Here, in Quebec, power is around $0.05/kWh CAN (about $0.04 US)... and it goes around $0.035 (roughly 0.03$ US) outside peak hours for people who are on bi-energy or similar plans.
Since I only have electric heaters in my apartment, a big PSU (and a PC to match) would give the heaters a break when it is -30C outside with no measurable net effect on the power bill during most of winter. I wouldn't turn that on unless absolutely necessary during summer though.
If PSUs keep getting larger at the current pace, some people will have to consider getting tri-phase power in the not so distant future... or at least rewire their 110-120V houses for 220-240V.
Yes, OEMs get their Windows cheap. XP Home for bulk OEMs is somewhere around $50 (closer to $40 for major OEMs) instead of around $100 for small (mom&pop shop) quantities and $200 for retail. IIRC, XP Pro is only slightly more expensive for bulk OEMs and around $150 for small OEMs.
The $40 license saved on an Open PC barely covers the parts+labour required to re-image or swap the HDD from bulk-produced and imaged configurations. Dell has thin margins on standard configurations, it comes as no surprise that they are making customers pay for the privilege of substituting parts.
As for the actual article, Microsoft might not look as much like a monopoly when looked at globally but it still undeniably is a practical monopoly as far as desktop OSes are concerned. TFA simply got distracted by Microsoft diversification efforts and attempts at gnawing a chunk off others' quasi-monopolies.
IP and copyrights do create more jobs... at least for lawyers, judges, law enforcement and entertainment lobbyists.
For the rest of us who have to solve real-world problems and produce stuff that is actually useful, the current trend towards wholesale IP warfare is a major hinderance.
As others have said, M$ can simply boost European prices and pass some of the costs to the local and world-wide consumers, substantially undermining the fine's effective cost.
Did the fine include a clause prohibiting price changes? Or maybe a clause to scale the fine exponentially with any price changes? This would certainly put the brakes on pricing manipulations to shift the financial burden.
I find those things distracting too... but not doing so would be free publicity for whoever's generic product (bleach, gasoline, oil, etc.) I happened to be using and if I was producing a show, I would probably be tempted to do the same. By covering labels, they are telling manufacturers "you're not on the show unless you directly sponsor us", which seems perfectly fair for a documentary/reality-style entertainment/science show.
If they did not cover product labels, people with vested interests in a product category with few competitors could be tempted to start pushing/fabricating myths involving their products to get some free publicity. Sponsorships usually cost more than the freebies unless we get into expensive or burdensome freebies like concrete trucks.
I wonder about product placement in Mythbusters... if you have ever watched the show, they use "Mythbusters-branded" (either stick a Mythbuster print over the original labels or paint over them) everything. Cola, bug spray, gasoline, etc.
For many myths, this would work fine as long as they can get sponsorships for all the front-row stuff required by the myth. "Today, we are going to test wether or not fried chicken provided by KFC has more penetration power than thawed and frozen chicken provided by XYZ using a modified 20gal 200psi tank from MNO, 12" dia. 1/4" thick 10' long steel pipe from PQR, glass pannels from..." Sounds pretty burdensome given that they do the show primarily for fun.
Good patents are good... but as TFA says, at least 80% of patents are worthless (too obvious or impractical) and 95% of them are never used - though I suspect these ratios are rising.
I cannot remember ever reading a genuinely innovative and useful patent... all those I have read to date (mostly software and digital system stuff) were either 1) expensive practical jokes, 2) obvious, 3) excessively broad blankets obviously intended for patent warfare or 4) technically interesting but practically useless.
The patent system needs a thorough reform to reduce the number of unnecessary/redundant/parasitic/etc. patents that get granted. Inventing stuff should not require threading through a legal/patent minefield.
Biographies lack insight (incomplete facts, incorrect interpretations, etc.) while autobiographies are more likely to be biased to beautify reality. Beyond fundamental facts though, biographies leave all the details to the writer/publisher's impression of what is important and their respective spin on those points... it may be less likely to be biased but the subject may disagree.
Biographies, auto- or otherwise, are always subjective to some non-negligible extent and never inherently unbiased nor complete. Someone trying to draw a complete picture would at least need both to compare what the world remembered VS what the subject considered to be important.
"An autobiography (from the Greek auton, 'self', bios, 'life' and graphein, 'write') is a biography written by the subject or composed conjointly with a collaborative writer (styled "as told to" or "with")."
An "autobiography" where the subject is not allowed to intervene is not an autobiography anymore.
TFA says the alarm was caused by a cylinder wedged inside a transport tube by a defective switch and that it took them three weeks (presumably of trying everything available in-house) to come up with the robot idea.
The/. article could have been titled "Mighty Mouse strikes back" - TFA says the robot used was called "Mighty Mouse 2".
A centimeter worth of lead would stop most of the lower-energy stuff.
The story is kinda ironic: the irradiated cobalt was intended to test electronics against radiation. So, the robotics lab that lended the robot got a free test-run to verify their radiation tolerance calculations.
Note to would-be evil geniuses: put your bombs in shells made of irradiated cobalt isotopes, it may disable would-be bomb-disposal robots and personnel before they can do anything about it. Radiation labs will get a free test of their security measures and delivery tubes out of the deal.
I'm amazed (sort of) that the "ability to read eMail on an hand-held device" is patentable... I can read/write mail on my cell phone, PDA, laptop, etc. Similar abilities have been available since well before the Blackberry, the BB only made it more convenient. I seriously hope RIM's appeal about NTP's patent validity goes to RIM... AFAIK, simple aggregation is not patentable.
Software patents are harmful and these cases are only the tip of the IP-warfare iceberg. Those jokes about IP enforcement creating milions of new jobs - mostly lawyers - will not be so funny in the near future if things keep going the way they have been headed for the last few years and particularly since software patents.
Software and similar patents are almost like patenting oxygen.
How many car accidents involve large quantities of high-pressure gas? Not many because only a tiny fraction of people travel with such tanks.
Dive tanks are nearly impossible to rupture because they are outrageously over-engineered to take a fair amount of abuse and mishandling. They also lack the storage capacity to maintain their internal pressure if any fracture gets through.
For bulk mobile high-pressure tanks, things go double-walled and these require 'hazard' clearance in many jurisdictions even for non-toxic non-flammable stuff, same for cryogenic gases. Dive tanks are hardly bulk when considered individually, though I would not be surprised if some local authorities placed additional limits on how much can be onboard a single vehicle at any given time.
Tell that to everyone within the blast radius if you ever are involved in an accident where one of those tanks ruptures... either from impact damage or post-impact fire. Also, a single large-capacity tank would provide for far more spectacular catastrophic failures. (And now that I think about it, 4gal is rather small and does not fit with my vague memories of the demo video where the seats sit on the L-shaped tank... must have been closer to 10gal.)
As others have pointed out, there are two primary flavours for hybrids: electric-assisted and electric-drive. Electric-assisted is the one where the electric motor provides extra torque while electric-drive is the one where the electric motor is driving the wheels and the gas engine supplements the battery power.
In my post, I referred specifically to the electric-drive variant since with improved battery technology like the upcoming nano-structured plates, it has the most MPG potential. The assist variant is better from the cost-to-benefits point of view with current technology since it places lower average demand on batteries and electrical components.
I seriously doubt the tribrid idea would fly... electric-drive hybrids run off batteries most of the time with the combustion engine kicking off only during strong-ish accelerations or when the batteries go below a certain point. Most of the time, the gas engine would not run long enough to release enough heat and generate usable amounts of steam. For applications where the gas engine does run long enough for that, this implies sustained heavy load where electric hybrids are not quite worth the trouble.
I saw a video once of people doing 0-100km/h in ~10s with a ~1ton vehicle and 18HP engine... instead of coupling the engine to the wheels, they used it to drive a compressor to pressurize a 4gal 2000psi tank and pneumatic motors. The pneumatic motors were also used for regenerative braking, allowing the car to do repeated non-stop 0-100 runs. You're not going to see this on the streets any time soon though since riding on a 2000psi 4gal tank is very much like riding on a big pipe bomb.
Exactly. I should have been more specific and said "isolated point-to-point links".
Trackers are essential only to publish contacts (sources), everything else about them is purely for convenience and statistics. Nothing stops peers from contacting the tracker only once to get an initial contact list and have its IP:port published.
If the P4 was reworked for pure multithreaded performance, it would have close to double the throughput for half the watts. What's holding back desktop CPUs is obsession with single-thread performance for legacy software which leads to barely manageable complexity - the P4 is one fine example of this complexity getting out of hands.
As for the T1, I have known about it for about a year but I have been thinking that quad-threaded CPUs would be the sweet spot since the first time I heard about multi-threading chip theory. With a single thread, one needs complex logic extract ILP to consistently use more than 40% of the chip's execution resources but the extra logic uses more power, uses more logic, adds latency and wastes work when results from speculative execution (missed branch prediction... and many other things with Netburst) are discarded. With dual-threaded core and those complex tricks, utilisation can reach close to 100%. A pure quad-threading core should be able to stick close to 100% execution unit load with nothing more than minimal/basic instruction reordering, something like 16 instructions deep instead of 96+ on modern single-threaded-performance-centric CPUs - and that is indeed pretty close to where the US-T1 is at.
My 3-4 years guesstimate is the time I think the mainstream (desktop) market will need to shift from single-threaded performance to multi-threaded with enough pressure on performance/watt to make AMD/Intel revisit multi-threading for maximizing throughput/core and throughput/watt. Yes, this is what Sun has been plotting for the last few years but US-T1 is not and never will be a mainstream chip.
The 360's PPCs are stripped-down cores, they are very un-powerful for anything that has non-linear code/data dependencies. Among things that were stripped from the PPC is out-of-order execution and branch prediction, which would kill performance for stuff that has non-linear code or data dependencies. For games, most of the processing is bulk geometric and physics transforms, the dual-threaded simplified PPC cores can deal with that reasonably well but do not expect branchier stuff (like relational DBs) to fare well.
Simpler cores have lower per-thread performance (shared resources) but simple multi-threading cores should scale much better in troughput/core and throughput/watt. (less/no wasted work, fewer sub-circuits having idle cycles, less duplicated logic) I personally believe that such quad-threading CPUs will be the sweet spot 3-4 years from now.
That was true for the original XBox which was based on PC-ish and off-the-shelf hardware. For the 360, it would be more like a souped-up Mac/(whatever else uses PPC chips).
Since the custom PPC and chipset/GPU for the 360 were tweaked specifically for M$ from the very start instead of quick hacks of off-the-shelf designs to make them less interoperable, chances are that the 360 will be much harder to crack.
Ah, yes... I can imagine it now. Traditional Klingon mariage on their homeworld where all the witnesses have their Humpty-Dumpty/Frito-Lay/etc. snacks in hand along with Coca-Cola/Pepsi/etc. drink in the other rather than their more traditional Blood Wine and gross live food. I guess they could make the bride kill the groom for offering a stupid/worthless diamond/rock ring.
Well, whatever. I have nearly stopped watching TV anyway, more people doing the same would hurt the medias quite a bit more than TiVo.
Even stuff designed to be relatively secure becomes largely pointless as soon as ONE person finds a practical work-around. I'm looking forward to the Blu-Ray/HD-DVD's DeCSS episode.
Many people appear to have very funny ideas about how BT works and how trackers "control" how peers interact... in reality, BT peers form ad-hoc point-to-point links and the tracker's only true purpose is to provide source propagation. Trackers have absolutely no control over who uploads how much to whoever and peers are free to report completely bogus tracking information - tracker stats should only be considered as ranging from rough to completely bogus estimations. The only trustable ratio enforcement is one based on local history like eMule's network-unique secure peer IDs with credit system where upload priority is given to higher-ranked peers, not practical for BT's disposable per-torrent/session IDs.
The tracker ratio uses client-supplied upload and download figures. Nothing stops people from hacking their BT client to report arbitrary upload and download figures so tracker-based ratios are pointless against moderately smart leeches - if you want to mod Azureus into a leech mod, simply find the place where the tracker update format string is and change "upload=%i" to "upload=%i0", that's all there is to it - this will report a 0.1 share ratio as ~1 and set people free from ratios on most "1331" trackers.
(Well, continuously reporting [0-9]*0 uploads and constant factors would be suspicious, that's why my initial suggestion was more complex.)
The upload and download figures on the BT tracker protocol are only for statistics, they should not be trusted/expected to be reliable since modifying clients to skew the results is trivial.
The BC people could simply generate bogus agents or no agent at all.
Trying to enforce artificial tracker-centric limits is pointless. Leeches and "freeloaders" can mod Azureus and other OSS BT clients to bypass/ignore artificial restrictions like 'private flag' and ratios. For ratios/leech-ban, anyone can mod an OSS client to make it report an arbitrary yet plausible upload count, a simple form of which could be U(n) = U(n-1) + k1 * (0.5+rnd()) * (D(n) - D(n-1) + k0). (where 0 = rnd() = 1 and k0/k1 are used to tweak the simulated traffic profile)... or even simpler than that, append an extra random digit to the upload volume and the ratio magically goes from 1:10 to 1:1.
BT is one of the dumbest KISS protocols out there, keeping things locked up is one of the many things it was not designed/intended to do. Banning BC will only cause people to mod other BT clients to achieve their desired results and send BC-ban-happy sites back to the drawing boards.
Mercury medical thermometers contain ~50X as much mercury as typical mercury/fluourescent lamps and a toxicologically significant dose is over 50X more than a thermometer. One would have to sniff hundreds of freshly broken lamps to get mercury poisoning symptoms from that alone. People should not lose sleep about the compact-fluourescent or similar tubes in their home/apartment/workplace/etc. unless thousands of them routinely shatter at once whenever ventilation breaks down.
Slashdot Mirror
Here, in Quebec, power is around $0.05/kWh CAN (about $0.04 US)... and it goes around $0.035 (roughly 0.03$ US) outside peak hours for people who are on bi-energy or similar plans.
Since I only have electric heaters in my apartment, a big PSU (and a PC to match) would give the heaters a break when it is -30C outside with no measurable net effect on the power bill during most of winter. I wouldn't turn that on unless absolutely necessary during summer though.
If PSUs keep getting larger at the current pace, some people will have to consider getting tri-phase power in the not so distant future... or at least rewire their 110-120V houses for 220-240V.
Yes, OEMs get their Windows cheap. XP Home for bulk OEMs is somewhere around $50 (closer to $40 for major OEMs) instead of around $100 for small (mom&pop shop) quantities and $200 for retail. IIRC, XP Pro is only slightly more expensive for bulk OEMs and around $150 for small OEMs.
The $40 license saved on an Open PC barely covers the parts+labour required to re-image or swap the HDD from bulk-produced and imaged configurations. Dell has thin margins on standard configurations, it comes as no surprise that they are making customers pay for the privilege of substituting parts.
As for the actual article, Microsoft might not look as much like a monopoly when looked at globally but it still undeniably is a practical monopoly as far as desktop OSes are concerned. TFA simply got distracted by Microsoft diversification efforts and attempts at gnawing a chunk off others' quasi-monopolies.
IP and copyrights do create more jobs... at least for lawyers, judges, law enforcement and entertainment lobbyists.
For the rest of us who have to solve real-world problems and produce stuff that is actually useful, the current trend towards wholesale IP warfare is a major hinderance.
As others have said, M$ can simply boost European prices and pass some of the costs to the local and world-wide consumers, substantially undermining the fine's effective cost.
Did the fine include a clause prohibiting price changes? Or maybe a clause to scale the fine exponentially with any price changes? This would certainly put the brakes on pricing manipulations to shift the financial burden.
I find those things distracting too... but not doing so would be free publicity for whoever's generic product (bleach, gasoline, oil, etc.) I happened to be using and if I was producing a show, I would probably be tempted to do the same. By covering labels, they are telling manufacturers "you're not on the show unless you directly sponsor us", which seems perfectly fair for a documentary/reality-style entertainment/science show.
If they did not cover product labels, people with vested interests in a product category with few competitors could be tempted to start pushing/fabricating myths involving their products to get some free publicity. Sponsorships usually cost more than the freebies unless we get into expensive or burdensome freebies like concrete trucks.
I wonder about product placement in Mythbusters... if you have ever watched the show, they use "Mythbusters-branded" (either stick a Mythbuster print over the original labels or paint over them) everything. Cola, bug spray, gasoline, etc.
For many myths, this would work fine as long as they can get sponsorships for all the front-row stuff required by the myth. "Today, we are going to test wether or not fried chicken provided by KFC has more penetration power than thawed and frozen chicken provided by XYZ using a modified 20gal 200psi tank from MNO, 12" dia. 1/4" thick 10' long steel pipe from PQR, glass pannels from..." Sounds pretty burdensome given that they do the show primarily for fun.
Good patents are good... but as TFA says, at least 80% of patents are worthless (too obvious or impractical) and 95% of them are never used - though I suspect these ratios are rising.
I cannot remember ever reading a genuinely innovative and useful patent... all those I have read to date (mostly software and digital system stuff) were either 1) expensive practical jokes, 2) obvious, 3) excessively broad blankets obviously intended for patent warfare or 4) technically interesting but practically useless.
The patent system needs a thorough reform to reduce the number of unnecessary/redundant/parasitic/etc. patents that get granted. Inventing stuff should not require threading through a legal/patent minefield.
Biographies lack insight (incomplete facts, incorrect interpretations, etc.) while autobiographies are more likely to be biased to beautify reality. Beyond fundamental facts though, biographies leave all the details to the writer/publisher's impression of what is important and their respective spin on those points... it may be less likely to be biased but the subject may disagree.
Biographies, auto- or otherwise, are always subjective to some non-negligible extent and never inherently unbiased nor complete. Someone trying to draw a complete picture would at least need both to compare what the world remembered VS what the subject considered to be important.
What's so wrong about editing one's AUTObiography?
"An autobiography (from the Greek auton, 'self', bios, 'life' and graphein, 'write') is a biography written by the subject or composed conjointly with a collaborative writer (styled "as told to" or "with")."
An "autobiography" where the subject is not allowed to intervene is not an autobiography anymore.
TFA says the alarm was caused by a cylinder wedged inside a transport tube by a defective switch and that it took them three weeks (presumably of trying everything available in-house) to come up with the robot idea.
/. article could have been titled "Mighty Mouse strikes back" - TFA says the robot used was called "Mighty Mouse 2".
The
A centimeter worth of lead would stop most of the lower-energy stuff.
The story is kinda ironic: the irradiated cobalt was intended to test electronics against radiation. So, the robotics lab that lended the robot got a free test-run to verify their radiation tolerance calculations.
Note to would-be evil geniuses: put your bombs in shells made of irradiated cobalt isotopes, it may disable would-be bomb-disposal robots and personnel before they can do anything about it. Radiation labs will get a free test of their security measures and delivery tubes out of the deal.
I'm amazed (sort of) that the "ability to read eMail on an hand-held device" is patentable... I can read/write mail on my cell phone, PDA, laptop, etc. Similar abilities have been available since well before the Blackberry, the BB only made it more convenient. I seriously hope RIM's appeal about NTP's patent validity goes to RIM... AFAIK, simple aggregation is not patentable.
Software patents are harmful and these cases are only the tip of the IP-warfare iceberg. Those jokes about IP enforcement creating milions of new jobs - mostly lawyers - will not be so funny in the near future if things keep going the way they have been headed for the last few years and particularly since software patents.
Software and similar patents are almost like patenting oxygen.
How many car accidents involve large quantities of high-pressure gas? Not many because only a tiny fraction of people travel with such tanks.
Dive tanks are nearly impossible to rupture because they are outrageously over-engineered to take a fair amount of abuse and mishandling. They also lack the storage capacity to maintain their internal pressure if any fracture gets through.
For bulk mobile high-pressure tanks, things go double-walled and these require 'hazard' clearance in many jurisdictions even for non-toxic non-flammable stuff, same for cryogenic gases. Dive tanks are hardly bulk when considered individually, though I would not be surprised if some local authorities placed additional limits on how much can be onboard a single vehicle at any given time.
Tell that to everyone within the blast radius if you ever are involved in an accident where one of those tanks ruptures... either from impact damage or post-impact fire. Also, a single large-capacity tank would provide for far more spectacular catastrophic failures. (And now that I think about it, 4gal is rather small and does not fit with my vague memories of the demo video where the seats sit on the L-shaped tank... must have been closer to 10gal.)
As others have pointed out, there are two primary flavours for hybrids: electric-assisted and electric-drive. Electric-assisted is the one where the electric motor provides extra torque while electric-drive is the one where the electric motor is driving the wheels and the gas engine supplements the battery power.
In my post, I referred specifically to the electric-drive variant since with improved battery technology like the upcoming nano-structured plates, it has the most MPG potential. The assist variant is better from the cost-to-benefits point of view with current technology since it places lower average demand on batteries and electrical components.
I seriously doubt the tribrid idea would fly... electric-drive hybrids run off batteries most of the time with the combustion engine kicking off only during strong-ish accelerations or when the batteries go below a certain point. Most of the time, the gas engine would not run long enough to release enough heat and generate usable amounts of steam. For applications where the gas engine does run long enough for that, this implies sustained heavy load where electric hybrids are not quite worth the trouble.
I saw a video once of people doing 0-100km/h in ~10s with a ~1ton vehicle and 18HP engine... instead of coupling the engine to the wheels, they used it to drive a compressor to pressurize a 4gal 2000psi tank and pneumatic motors. The pneumatic motors were also used for regenerative braking, allowing the car to do repeated non-stop 0-100 runs. You're not going to see this on the streets any time soon though since riding on a 2000psi 4gal tank is very much like riding on a big pipe bomb.
Exactly. I should have been more specific and said "isolated point-to-point links".
Trackers are essential only to publish contacts (sources), everything else about them is purely for convenience and statistics. Nothing stops peers from contacting the tracker only once to get an initial contact list and have its IP:port published.
If the P4 was reworked for pure multithreaded performance, it would have close to double the throughput for half the watts. What's holding back desktop CPUs is obsession with single-thread performance for legacy software which leads to barely manageable complexity - the P4 is one fine example of this complexity getting out of hands.
As for the T1, I have known about it for about a year but I have been thinking that quad-threaded CPUs would be the sweet spot since the first time I heard about multi-threading chip theory. With a single thread, one needs complex logic extract ILP to consistently use more than 40% of the chip's execution resources but the extra logic uses more power, uses more logic, adds latency and wastes work when results from speculative execution (missed branch prediction... and many other things with Netburst) are discarded. With dual-threaded core and those complex tricks, utilisation can reach close to 100%. A pure quad-threading core should be able to stick close to 100% execution unit load with nothing more than minimal/basic instruction reordering, something like 16 instructions deep instead of 96+ on modern single-threaded-performance-centric CPUs - and that is indeed pretty close to where the US-T1 is at.
My 3-4 years guesstimate is the time I think the mainstream (desktop) market will need to shift from single-threaded performance to multi-threaded with enough pressure on performance/watt to make AMD/Intel revisit multi-threading for maximizing throughput/core and throughput/watt. Yes, this is what Sun has been plotting for the last few years but US-T1 is not and never will be a mainstream chip.
The 360's PPCs are stripped-down cores, they are very un-powerful for anything that has non-linear code/data dependencies. Among things that were stripped from the PPC is out-of-order execution and branch prediction, which would kill performance for stuff that has non-linear code or data dependencies. For games, most of the processing is bulk geometric and physics transforms, the dual-threaded simplified PPC cores can deal with that reasonably well but do not expect branchier stuff (like relational DBs) to fare well.
Simpler cores have lower per-thread performance (shared resources) but simple multi-threading cores should scale much better in troughput/core and throughput/watt. (less/no wasted work, fewer sub-circuits having idle cycles, less duplicated logic) I personally believe that such quad-threading CPUs will be the sweet spot 3-4 years from now.
That was true for the original XBox which was based on PC-ish and off-the-shelf hardware. For the 360, it would be more like a souped-up Mac/(whatever else uses PPC chips).
Since the custom PPC and chipset/GPU for the 360 were tweaked specifically for M$ from the very start instead of quick hacks of off-the-shelf designs to make them less interoperable, chances are that the 360 will be much harder to crack.
Ah, yes... I can imagine it now. Traditional Klingon mariage on their homeworld where all the witnesses have their Humpty-Dumpty/Frito-Lay/etc. snacks in hand along with Coca-Cola/Pepsi/etc. drink in the other rather than their more traditional Blood Wine and gross live food. I guess they could make the bride kill the groom for offering a stupid/worthless diamond/rock ring.
Well, whatever. I have nearly stopped watching TV anyway, more people doing the same would hurt the medias quite a bit more than TiVo.
Exactly.
Even stuff designed to be relatively secure becomes largely pointless as soon as ONE person finds a practical work-around. I'm looking forward to the Blu-Ray/HD-DVD's DeCSS episode.
Many people appear to have very funny ideas about how BT works and how trackers "control" how peers interact... in reality, BT peers form ad-hoc point-to-point links and the tracker's only true purpose is to provide source propagation. Trackers have absolutely no control over who uploads how much to whoever and peers are free to report completely bogus tracking information - tracker stats should only be considered as ranging from rough to completely bogus estimations. The only trustable ratio enforcement is one based on local history like eMule's network-unique secure peer IDs with credit system where upload priority is given to higher-ranked peers, not practical for BT's disposable per-torrent/session IDs.
The tracker ratio uses client-supplied upload and download figures. Nothing stops people from hacking their BT client to report arbitrary upload and download figures so tracker-based ratios are pointless against moderately smart leeches - if you want to mod Azureus into a leech mod, simply find the place where the tracker update format string is and change "upload=%i" to "upload=%i0", that's all there is to it - this will report a 0.1 share ratio as ~1 and set people free from ratios on most "1331" trackers.
(Well, continuously reporting [0-9]*0 uploads and constant factors would be suspicious, that's why my initial suggestion was more complex.)
The upload and download figures on the BT tracker protocol are only for statistics, they should not be trusted/expected to be reliable since modifying clients to skew the results is trivial.
The BC people could simply generate bogus agents or no agent at all.
Trying to enforce artificial tracker-centric limits is pointless. Leeches and "freeloaders" can mod Azureus and other OSS BT clients to bypass/ignore artificial restrictions like 'private flag' and ratios. For ratios/leech-ban, anyone can mod an OSS client to make it report an arbitrary yet plausible upload count, a simple form of which could be U(n) = U(n-1) + k1 * (0.5+rnd()) * (D(n) - D(n-1) + k0). (where 0 = rnd() = 1 and k0/k1 are used to tweak the simulated traffic profile)... or even simpler than that, append an extra random digit to the upload volume and the ratio magically goes from 1:10 to 1:1.
BT is one of the dumbest KISS protocols out there, keeping things locked up is one of the many things it was not designed/intended to do. Banning BC will only cause people to mod other BT clients to achieve their desired results and send BC-ban-happy sites back to the drawing boards.
Heh.
Mercury medical thermometers contain ~50X as much mercury as typical mercury/fluourescent lamps and a toxicologically significant dose is over 50X more than a thermometer. One would have to sniff hundreds of freshly broken lamps to get mercury poisoning symptoms from that alone. People should not lose sleep about the compact-fluourescent or similar tubes in their home/apartment/workplace/etc. unless thousands of them routinely shatter at once whenever ventilation breaks down.