Many of the 'many factors of lithium battery life' apply to many other battery technologies. Most rechargeable batteries do not like going above 90% of a full charge either: shoving extra ions into the graphite electrode adds mechanical strain and eventually causes the electrode to fracture. This is part of the reason why lithium cells absolutely require high-tech chargers and why advanced chargers are highly recommended for pretty much all battery techs other than lead-acid.
As far as disconnecting the battery goes, this is unnecessary: the charge controller simply needs to drop its end-of-charge output voltage until the battery's net current becomes negligible and the battery remains connected for instant pick-up should primary power go down. This is how most on-line battery systems work.
BTW, forget about having hold capacitors even for 100ms at low voltage because sizes go up insanely fast: just for the heck of it, I sized the capacitor needed for 20ms hold for an hypothetical 100W laptop with 18-14V input tolerance and that already comes out at 33mF - over a cubic inch for a 25V standard capacitor, not exactly the best use of the space and weight budget.
Although there were no outright failures, the Prius' batteries were down to under 40% of their nominal capacity and that had a -20% impact on MPG, dragging it back down to the realm of non-hybrids. Since these were 2002 models, that makes them under four years old when the report was published, there should be some decent mileage left in these cars for people who do less than 20k miles/year and happen to have a mechanics friend who can do an inexpensive 160k miles overhaul - batteries not included, add ~$2000.
So, people with ~120k miles 2002 Prius will have to look into replacement batteries if they wish to get their MPG back, otherwise they're only toting dead (battery) weight. On the other hand, it seems like Civics and Insights are good for a little while longer, with the Insignts apparently having the only visibly over-engineered battery pack (still holding over 80% after five years) of this particular lot.
In the harsher Canadian winter, I expect hybrids to have a much harder time: I doubt hybrid batteries (of the same designs) would fare remotely as well while enduring weeks worth of crazy weather where temperatures can jump from +10C to -40C almost overnight.
At a glance, it seems the patent is about a CPU that breaks down a program among multiple execution cores and resynchronizes the results afterwards. Since the PS3 relies on parallelizing compilers, Cell-optimized libraries and explicit programmer intervention to exploit its parallel hardware rather than have some sort of dispatch engine in Cell to do this seamlessly - assuming this is actually feasible and practical, the case appears to be a baseless patent scam. However, the fact that PS3 devs have such a hard time getting up to speed with Cell programming hints that the PS3 possesses no such miracle hardware - extracting parallelism from Cell is still as hands-on a practice as it has been on all parallel computers up to now: an art in SOFTWARE design.
AFAIK, pretty much all the MP scaling up to now is purely software-driven either by explicit programmer intervention (MPI, OpenMP tagging or explicit threading) or picking MP/MPI-optimized libraries for all the heavy lifting. If that company should sue anyone, the closest thing I've ever heard of that might infringe on the patent would be AMD's fabled reverse-multithreading - a CPU able to seamlessly extract parallelism from a single instruction stream and dispatch execution among multiple processing cores.
BTW, I vaguely remember reading similar health warnings about matrix printers related to ink/ink-ribbon/paper dust generated by the pin impacts in high-speed departmental monsters from the 80's.
Inorganic fine particulates have been a health hazard for as long as we have known of their existence... that toner got five minutes of shame for being such a material today is not much of a surprise.
With lithium batteries, it is charge depth that makes the largest difference: they last longest when they are not allowed to drop much below 70-80%. Actually, this is true for most rechargeable battery technologies. Overcharging (deficient charger designs) is the other major cause of premature failures - my cheap GE cordless phone pretty much killed its OEM NiCd batteries within a year.
If my laptop's battery is any indication of typical li-ion batteries, deep-cycling hurts them really bad - I deep-cycled my laptop's battery many more times than I ever meant to due to the power brick's plug slipping out of the laptop's socket on its own and the charge circuit being inhibited unless the plug is fully inserted. After about a year, I was already down to around 70% battery life even though I used my laptop plugged-in (as far as I knew) 99% of the time.
At the other end of the spectrum, my current cell phone (four years old Nokia 7190) also has a li-ion battery. I plug it in overnight whenever possible and the battery still holds a very decent charge: the phone still indicates a full battery after 3-4 days of (unplugged) standby, seems as good as new.
Since some supermarkets were fraudulently mismarking prices in my area some years ago, a law was passed to force supermarkets to GIVE AWAY items under $10 with mismatched in-store (product label VS cash register) prices. For items above $10, they have to deduce $10 off the lowest price or give it away. So here, reporting supermarket price mismatches is now always to the customer's advantage. With measures like this, store managers are now double-checking their prices to avoid this law getting extended to cover other commerces.
As for ATM incidents, the bank knows about its ATM issues, about the periods where these were in effect, who used it and how so not much is really stopping them from charging/crediting accounts to cover systematic over/under-paying ATM malfunctions. As someone else wrote, banks have insurance so people shouldn't have to pay for the bank's fuck-ups. The only cases where I would see grounds for prosecution is for hacking or knowingly accessing a hacked ATM.
If my experience with Energizer's 15min. charger has taught me anything it is to avoid ultra-fast chargers. The stupid thing ruins (degrades them enough that the charger refuses to charge them) AAA cells in about 50 charges and AA ones in about 100.
I am extremely skeptical about charging lithium cells at 12C currents and claiming improved longevity... lithium cells have a long history of not liking currents above 1C where longevity is concerned. ('C' is the cell's nominal capacity in Ah... 12C for an 1Ah cell = 12A charge/discharge current.) Another problem with lithiums is that much like lead acid batteries, they do not like deep-cycling.
The biggest problem with hybrids (particularly the plug-in types that put extra strain on the batteries) is that the eco-freaks fail to consider the environmental costs of replacing the car's battery every 2-5 years. The carbon and miscellaneous pollutants footprint of manufacturing, recycling and recharging all those batteries probably goes a long way towards closing the gap between the hybrid propaganda and high-efficiency combustion-only vehicles.
I was writing more about scalability than actual performance levels. A monolithic multi-core CPU has fewer interconnect bottlenecks than equivalent multi-die/chip implementations and will therefore allow better scaling with core count. Since multi-threaded processing throughput quickly degrades when CPUs start waiting to sync with each other, keeping all delays to the absolute minimum becomes critical as the total system core count increases.
If we could make a direct comparison between a dual-dual and a native quad (we cannot given the numerous other tweaks that happen between core revisions), we would probably see the native quad lead by 1-20% depending on how much cache trashing and process synchronization the benchmark is generating. There are various ways to avoid costly thread/CPU synchronizations but they require HPC-style tricks few programmers are familiar with. With less costly synchronization, monolithic multi-core CPUs will take a much smaller performance hit from sub-optimal multi-threaded coding.
Also, even though the Core2 CPUs may lead in raw performance, Athlon64/Opterons often take the lead in bandwidth-constrained situations where Core2/Xeons are simply choking on the shared (CPUs+IO+RAM) FSB... no amount of IPC+MHz domination can save Core2 in that situation.
I do not care which one has the upper hand now as I plan to stretch my current PC+laptop until late 2008 unless they die before then... but given how messy AMD's business has become after the ATI acquisition, my next system is most likely going to be Intel-based.
Native quad-core is better than dual-dual-core because more cores can exchange cache snoop data over CPU-speed internal buses instead of low-speed external buses. Cache snooping quickly kills performance scaling on shared FSB architectures like the P3, P4 and Core 1&2. Since the same FSB is also used for memory IO, cache snooping robs some more of the FSB-limited memory performance on P3/P4/Core-1&2 FSB-based SMP architectures.
Shared FSB systems do not scale... even Intel knows that. However, dual-dual-core is more profitable short-term and easily more than enough to give AMD a run for its money for a good while longer. Things will get really interesting after Intel's Nehalem materializes in late 2008... perhaps we'll see AMD get one-up'd like they were when Core2 came out last year.
Deep-cycling is a killer for all battery chemistries. One of my friends used to deep-cycle his cordless phone and his batteries lasted only about a year each. I convinced him to try putting his phone on the base station each night with his new battery and it is still going strong after more than three years of service. Me, I leave my cell phones plugged in whenever I am home and my NiMH batteries lasted at least three years each while my current Li-ion one is nearly four years old and still going strong.
From all the stuff I read about different battery technologies, keeping cells fully charged whenever possible (without overcharging) is the best way of maximizing a battery's useful lifespan... and it seems this is even more important/effective with lithium batteries.
For my usage pattern, inkjet printers are one-time-use devices: since it may be weeks between print jobs, the printheads are often hopelessly clogged by the time I try to use the printer again. With disposable inkjets being a $50-a-pop proposition, my last replacement was one of those $200 (after $350 instant rebate) color laser printers I saw on liquidation last year. By now, it has certainly paid for itself a few times over and I am only about half-way through the OEM toners.
Right now, I am wondering if I should buy a set of replacement cartridges before they go out of production or just replace the printer when its toner runs out... at the current pace, it is going to be over a year past its warranty by the time its OEM cartridges are spent anyhow.
I hate throwing stuff out but we're really living in throw-away consumerism, this is quite a contrast to all the pro-environment face so many try to put on.
If I wrote a custom PHP web app (like a tracker) and did not bother to add function calls to write data to disk, the judge would have to order that these be added before he could claim that evidence was destroyed since the information up to that point was never meant to be preserved - I do not keep transcripts of my phone calls or other oral communications so a judge cannot order me to hand over such transcripts and claim that failure to comply will be considered as destruction of evidence... he'll have to order wiretaps/bugging (logging) and let the authorities work with that afterwards.
For sites like TS, there is pretty much only one practical solution: move operations to somewhere beyond the reach of US legislation with pretty lax copyright/IP enforcement laws.
BTW, the *AA are lobbying to get police funds moved from physical crimes to copyright enforcement... I think the *AA VIPs need a taste of real crimes to set them straight - they need to be reminded that real crimes can cause permanent physical and psychological injuries beyond what can be quantified in dollar value while copyright infringement only cause vague and vastly exaggerated hypothetical revenue loss.
Since the RAM used to process HTTP requests usually gets reallocated and rewritten every couple of miliseconds, the data is usually gone long before you even get a chance to cause a kernel panic or memory dump to prevent further memory alterations.
It is impossible to comply literally with that order: it is fundamentally incompatible with the normal operation of RAM/SRAM which are meant to store a computer's working data set only for as long as it is necessary to accomplish the task at hand... and even within that task, countless memory reallocations and modifications (data destructions) are bound to occur.
Computer data, be it RAM or HDD, gets reallocated/modified/destroyed as a matter of course during normal operation... computers cannot operate without continuously destroying some data. Building computers with WORM memory would be ludicrously impractical and unfeasibly expensive. Compliance with the literal interpretation of this order to any useful extent is physically impossible since each server would need its own storage media factory to supply the ~100k DVDs/day needed to dump all memory writes.
Of course, we all know what the judge really meant to order is the logging of relevant data... he just did not know anything about how real-world computers work and TS failed at educating him. It is quite scary how judges can issue such unreasonably uneducated orders.
May I suggest that you check the fundamental definition of 'animal'... the most basic definition (animant) includes anything capable of moving on its own so humans are simply another kind of animal. Human DNA does not really mysteriously/intrinsically separate humans from the rest of the animal kingdom, what does is our unmatched (at least among Earth's creatures) ability to learn and our unique ability to work with abstractions.
Morals, religions, education, etc. are the only things that truly sets humanity apart. If a human babies somehow survived without any human tutoring, they'd pick up most of the behavior of whatever is caring for them - humans have the most degenerated ancestral memory of all animals so practically all the behavior we end up with is forged from environmental (social and other) exposure.
As far as nature is concerned, humans are simply abnormally successful animals.
The DMCA copyright amendment does not make DRM itself as a technological measure any more copyrightable. While DMCA in the USA is an extension of copyright law, things are different elsewhere. In any case, only the protected content is.
Most current DRM schemes rely on common cryptographic algorithm and secret keys, nothing copyrightable and very little (if anything) patentable there. Uncover the root keys and you can blow all DRM wide-open because pretty much everything else is non-patentable/copyrightable public knowledge.
CSS was never protected under copyright law... it was protected under DMCA and any other relevant law, industrial secrets, patents and licensing. Same thing for AACS. What is copyrighted is the protected content, not the DRM.
The ruling only means the court will not protect obsolete, weak or otherwise largely compromised DRM schemes. This is a good thing as it will prevent companies from seeking DMCA-like protection for daft/dummy/broken DRM schemes.
There was a similar technology using magnesium or some other metal a few months ago.
The catch with all these little alternative is that most promoters conveniently leave out the details of how much power and other costs are involved in manufacturing and recycling the spent metals when addressing the general public in promotional material.
When auto manufacturers and reviewers praise the fuel-efficiency of hybrids, they conveniently leave out the fact that the battery pack needs replacement every few years. Calculate your cost per mile after factoring in the cost of replacement batteries and you get something roughly equivalent to plain gasoline, possibly worse. We do not see much criticism of this yet because there are very few hybrids right now that are much over one year old so their batteries are still close to top-shape. In two or three years from now, performance will start to go down and we will see reports of people complaining about the cost of replacement batteries and how the reseller did not "properly" educate them about this unavoidable issue before closing the deal.
As much as people may not fancy having a tank of liquid hydrogen in their cars, I still think this is the most viable option until major battery technology break-throughs enable higher capacity (kWh/weight/volume), higher performance (higher currents, lower impedances) and longer-lived - or at the very least cheaper - cells.
I proposed 100 years on the basis that should I ever publish something of cultural value, I would like to have at least some level of control on how it gets used at least throughout my lifetime.
If I wrote a book, I want to still be entitled to my fair share of the profits should a movie be made of it 50 years later - only a very small minority of books become overnight successes that may attract the immediate attention of movie studios, less spectacular titles are even less likely to get that attention. For movies, if I distributed a popular movie on Beta/VHS I would most likely want to remaster it if the cost can be justified and sell it again on DVD, HD-DVD, Blu-Ray and anything else that may come in the future for as long as there is a public for it - that's pretty much what Lucas has been doing with Star Wars.
Making a living off writing is extremely tough for those whose names are still unknown - I can easily imagine them going suicidal/homicidal should a movie studio make blockbuster movie off books authors actually lost money on nearly immediately after the copyrights expired. Life + X years at least guarantees that authors will never have to face this situation and it guarantees that the author's heirs will not have to face the insult while still mourning the author.
Your vision of copyright appears to be far too short-sighted. As far as books, movie, music and other similar entertainments are concerned, the content retains tangible immediate (as-is) economic value (without impeding the work of others like perpetual patents would) for several years beyond initial publication.
Well, I'm in Canada... fatal cyclist accidents are currently around 150/year so that makes it more like at least 1.3% of all country-wide fatal bicycle accidents last year happened on that boulevard. Since that boulevard is the only way to cross the few highways and rail tracks (without doing a 5+ km detour only to end up on yet another high-traffic boulevard or highway service lanes) that stand in the way of me getting to work, avoiding it is a practical impossibility. Taking the subway would "work" but the detours (due to lousy PT service outside the core downtown area) it imposes make bicycle-only faster by about 30 minutes. Depending on traffic, bicycle is even competitive with going by car during rush hours: PT(sub+bus)=1h00-2h00 (1h20 typ.), bicycle=0h50, car=0h30-1h00 (0h40 typ.).
Lets continue the argument of perpetual copyright compared to ownership of physical objects... physical objects become damaged, obsolete and eventually disposed of. Works of art aside, physical objects have very much finite existences.
Also, even if copyrights were perpetual, it is very likely that descendants would eventually get rid or forget about their rights to their ancestors's work.
IMO, life (up to the last living actor/director/writer/etc. involved in a team production) + 25 years should be plenty sufficient... or maybe a flat 100 years from initial publications.
Most IR remote controls do not have IR receivers... and even if they do, they are there only to record IR patterns from other remote controls and this function is disabled unless the user triggers a key programming event. They most likely have no signal suitable for reliable remote triggering.
Network signal meters are also not really suitable as remote triggers since you get no control over the trigger mechanism. The whole point of the cell-phone, pager or any remote-controlled approach is to simply wire the speaker/vibrator/whatever to the trigger circuit and call the phone/pager/whatever to trigger. Network (non-)detection would be more like a backup self-destruct trigger in case the signal gets unexpectedly jammed.
Using alternative RF bands will work if the deployed jamming is targeted at specific service frequencies... but if the jamming device is a high-power Tesla coil or better broadband noise sources, the reliability of dumb RF triggers (accidental or failed triggering) would be substantially undermined.
Unregistered prepaid cell phones were convenient off-the-shelf addressable remote triggers. New procurement requirements mean terrorists will have to seek similarly convenient and reliable replacements from other sources - but they certainly could fall back to less reliable previous remote trigger schemes in the meantime like tuned high-Q LC oscillators.
Unknowing people (including you, apparently) gladly lump pot in the same category as cocaine, heroine, amphetamine and other hard drugs... I just find the precious irony and total lack of perspective here delicious. Kind of funny that you should write that... I am starting to think you are too frosty to read properly.
Swap alcohol and pot, swap drinkers and pot-heads, swap legality issues, the picture would not be much/any prettier.
Last time I checked, pot was not classified as a hard drug.
BTW, drug addicts rarely commit crimes while they are stoned... they commit them when they are nearly "sober", have run out of stock and start craving their next fix. They start with petty theft and depending on the addiction's severity, it can escalate all the way to home invasions and homicides - alcoholics rarely go that far.
Cases of lethal pot DUIs are not unheard of
If you consider the totals for alcohol incidents on one side and junkies on the other, the tallies of crimes and fatal incidents sides against the junkies when proportions are considered.
Alcohol-related incidents get more media attention in large part because they are more frequent and sometimes spectacular. Narcotics&all get less attention because fewer addicts/abusers means fewer (sometimes spectacular) cases to report so the whole issue fades into the collective background until the next big junkie case related to serial home invasions.
Do you seriously think the world would be better off if we replaced alcoholics by an equal number of pot addicts?
I do not hear about alcoholics committing robberies to get something to drink remotely as often as I see stories of junkies doing so for crack, heroine and other hard drugs... even though there probably are well over 100 alcoholics (to various degrees) for every hard-drugs junkie out there.
These guys often have a toe or more in organized crime and are far more likely to sell your banking and credit information to get their fix than your average alcoholic.
If I was a terrorist and I had to work around RF jamming, one possibility would be to use IR sensors and IR lasers/diodes. Not quite off-the-shelf like cell phones but it would be unscramblable by anything short of an EMP. The detector itself would be a wired remote to the actual detonator so the payload would not have to lie in plain sight. Add a small microcontroller and you get the ability to remotely program detonation delays so you can dispose of the triggering device before the fireworks start.
The major shortcoming would be the limited triggering range and suitable transmitter locations due to the LoS requirement.
Ultrasonic and spread-spectrum triggers would be more complex but definitely doable with DSP or FPGA+DAC... but the bomb would require somewhat large batteries since they would most likely have to be set several days ahead of the target's arrival.
Slashdot Mirror
Many of the 'many factors of lithium battery life' apply to many other battery technologies. Most rechargeable batteries do not like going above 90% of a full charge either: shoving extra ions into the graphite electrode adds mechanical strain and eventually causes the electrode to fracture. This is part of the reason why lithium cells absolutely require high-tech chargers and why advanced chargers are highly recommended for pretty much all battery techs other than lead-acid.
As far as disconnecting the battery goes, this is unnecessary: the charge controller simply needs to drop its end-of-charge output voltage until the battery's net current becomes negligible and the battery remains connected for instant pick-up should primary power go down. This is how most on-line battery systems work.
BTW, forget about having hold capacitors even for 100ms at low voltage because sizes go up insanely fast: just for the heck of it, I sized the capacitor needed for 20ms hold for an hypothetical 100W laptop with 18-14V input tolerance and that already comes out at 33mF - over a cubic inch for a 25V standard capacitor, not exactly the best use of the space and weight budget.
Although there were no outright failures, the Prius' batteries were down to under 40% of their nominal capacity and that had a -20% impact on MPG, dragging it back down to the realm of non-hybrids. Since these were 2002 models, that makes them under four years old when the report was published, there should be some decent mileage left in these cars for people who do less than 20k miles/year and happen to have a mechanics friend who can do an inexpensive 160k miles overhaul - batteries not included, add ~$2000.
So, people with ~120k miles 2002 Prius will have to look into replacement batteries if they wish to get their MPG back, otherwise they're only toting dead (battery) weight. On the other hand, it seems like Civics and Insights are good for a little while longer, with the Insignts apparently having the only visibly over-engineered battery pack (still holding over 80% after five years) of this particular lot.
In the harsher Canadian winter, I expect hybrids to have a much harder time: I doubt hybrid batteries (of the same designs) would fare remotely as well while enduring weeks worth of crazy weather where temperatures can jump from +10C to -40C almost overnight.
At a glance, it seems the patent is about a CPU that breaks down a program among multiple execution cores and resynchronizes the results afterwards. Since the PS3 relies on parallelizing compilers, Cell-optimized libraries and explicit programmer intervention to exploit its parallel hardware rather than have some sort of dispatch engine in Cell to do this seamlessly - assuming this is actually feasible and practical, the case appears to be a baseless patent scam. However, the fact that PS3 devs have such a hard time getting up to speed with Cell programming hints that the PS3 possesses no such miracle hardware - extracting parallelism from Cell is still as hands-on a practice as it has been on all parallel computers up to now: an art in SOFTWARE design.
AFAIK, pretty much all the MP scaling up to now is purely software-driven either by explicit programmer intervention (MPI, OpenMP tagging or explicit threading) or picking MP/MPI-optimized libraries for all the heavy lifting. If that company should sue anyone, the closest thing I've ever heard of that might infringe on the patent would be AMD's fabled reverse-multithreading - a CPU able to seamlessly extract parallelism from a single instruction stream and dispatch execution among multiple processing cores.
I have a box here with about 3000 sheets left...
BTW, I vaguely remember reading similar health warnings about matrix printers related to ink/ink-ribbon/paper dust generated by the pin impacts in high-speed departmental monsters from the 80's.
Inorganic fine particulates have been a health hazard for as long as we have known of their existence... that toner got five minutes of shame for being such a material today is not much of a surprise.
With lithium batteries, it is charge depth that makes the largest difference: they last longest when they are not allowed to drop much below 70-80%. Actually, this is true for most rechargeable battery technologies. Overcharging (deficient charger designs) is the other major cause of premature failures - my cheap GE cordless phone pretty much killed its OEM NiCd batteries within a year.
If my laptop's battery is any indication of typical li-ion batteries, deep-cycling hurts them really bad - I deep-cycled my laptop's battery many more times than I ever meant to due to the power brick's plug slipping out of the laptop's socket on its own and the charge circuit being inhibited unless the plug is fully inserted. After about a year, I was already down to around 70% battery life even though I used my laptop plugged-in (as far as I knew) 99% of the time.
At the other end of the spectrum, my current cell phone (four years old Nokia 7190) also has a li-ion battery. I plug it in overnight whenever possible and the battery still holds a very decent charge: the phone still indicates a full battery after 3-4 days of (unplugged) standby, seems as good as new.
Since some supermarkets were fraudulently mismarking prices in my area some years ago, a law was passed to force supermarkets to GIVE AWAY items under $10 with mismatched in-store (product label VS cash register) prices. For items above $10, they have to deduce $10 off the lowest price or give it away. So here, reporting supermarket price mismatches is now always to the customer's advantage. With measures like this, store managers are now double-checking their prices to avoid this law getting extended to cover other commerces.
As for ATM incidents, the bank knows about its ATM issues, about the periods where these were in effect, who used it and how so not much is really stopping them from charging/crediting accounts to cover systematic over/under-paying ATM malfunctions. As someone else wrote, banks have insurance so people shouldn't have to pay for the bank's fuck-ups. The only cases where I would see grounds for prosecution is for hacking or knowingly accessing a hacked ATM.
If my experience with Energizer's 15min. charger has taught me anything it is to avoid ultra-fast chargers. The stupid thing ruins (degrades them enough that the charger refuses to charge them) AAA cells in about 50 charges and AA ones in about 100.
I am extremely skeptical about charging lithium cells at 12C currents and claiming improved longevity... lithium cells have a long history of not liking currents above 1C where longevity is concerned. ('C' is the cell's nominal capacity in Ah... 12C for an 1Ah cell = 12A charge/discharge current.) Another problem with lithiums is that much like lead acid batteries, they do not like deep-cycling.
The biggest problem with hybrids (particularly the plug-in types that put extra strain on the batteries) is that the eco-freaks fail to consider the environmental costs of replacing the car's battery every 2-5 years. The carbon and miscellaneous pollutants footprint of manufacturing, recycling and recharging all those batteries probably goes a long way towards closing the gap between the hybrid propaganda and high-efficiency combustion-only vehicles.
I was writing more about scalability than actual performance levels. A monolithic multi-core CPU has fewer interconnect bottlenecks than equivalent multi-die/chip implementations and will therefore allow better scaling with core count. Since multi-threaded processing throughput quickly degrades when CPUs start waiting to sync with each other, keeping all delays to the absolute minimum becomes critical as the total system core count increases.
If we could make a direct comparison between a dual-dual and a native quad (we cannot given the numerous other tweaks that happen between core revisions), we would probably see the native quad lead by 1-20% depending on how much cache trashing and process synchronization the benchmark is generating. There are various ways to avoid costly thread/CPU synchronizations but they require HPC-style tricks few programmers are familiar with. With less costly synchronization, monolithic multi-core CPUs will take a much smaller performance hit from sub-optimal multi-threaded coding.
Also, even though the Core2 CPUs may lead in raw performance, Athlon64/Opterons often take the lead in bandwidth-constrained situations where Core2/Xeons are simply choking on the shared (CPUs+IO+RAM) FSB... no amount of IPC+MHz domination can save Core2 in that situation.
I do not care which one has the upper hand now as I plan to stretch my current PC+laptop until late 2008 unless they die before then... but given how messy AMD's business has become after the ATI acquisition, my next system is most likely going to be Intel-based.
Native quad-core is better than dual-dual-core because more cores can exchange cache snoop data over CPU-speed internal buses instead of low-speed external buses. Cache snooping quickly kills performance scaling on shared FSB architectures like the P3, P4 and Core 1&2. Since the same FSB is also used for memory IO, cache snooping robs some more of the FSB-limited memory performance on P3/P4/Core-1&2 FSB-based SMP architectures.
Shared FSB systems do not scale... even Intel knows that. However, dual-dual-core is more profitable short-term and easily more than enough to give AMD a run for its money for a good while longer. Things will get really interesting after Intel's Nehalem materializes in late 2008... perhaps we'll see AMD get one-up'd like they were when Core2 came out last year.
Deep-cycling is a killer for all battery chemistries. One of my friends used to deep-cycle his cordless phone and his batteries lasted only about a year each. I convinced him to try putting his phone on the base station each night with his new battery and it is still going strong after more than three years of service. Me, I leave my cell phones plugged in whenever I am home and my NiMH batteries lasted at least three years each while my current Li-ion one is nearly four years old and still going strong.
From all the stuff I read about different battery technologies, keeping cells fully charged whenever possible (without overcharging) is the best way of maximizing a battery's useful lifespan... and it seems this is even more important/effective with lithium batteries.
For my usage pattern, inkjet printers are one-time-use devices: since it may be weeks between print jobs, the printheads are often hopelessly clogged by the time I try to use the printer again. With disposable inkjets being a $50-a-pop proposition, my last replacement was one of those $200 (after $350 instant rebate) color laser printers I saw on liquidation last year. By now, it has certainly paid for itself a few times over and I am only about half-way through the OEM toners.
Right now, I am wondering if I should buy a set of replacement cartridges before they go out of production or just replace the printer when its toner runs out... at the current pace, it is going to be over a year past its warranty by the time its OEM cartridges are spent anyhow.
I hate throwing stuff out but we're really living in throw-away consumerism, this is quite a contrast to all the pro-environment face so many try to put on.
If I wrote a custom PHP web app (like a tracker) and did not bother to add function calls to write data to disk, the judge would have to order that these be added before he could claim that evidence was destroyed since the information up to that point was never meant to be preserved - I do not keep transcripts of my phone calls or other oral communications so a judge cannot order me to hand over such transcripts and claim that failure to comply will be considered as destruction of evidence... he'll have to order wiretaps/bugging (logging) and let the authorities work with that afterwards.
For sites like TS, there is pretty much only one practical solution: move operations to somewhere beyond the reach of US legislation with pretty lax copyright/IP enforcement laws.
BTW, the *AA are lobbying to get police funds moved from physical crimes to copyright enforcement... I think the *AA VIPs need a taste of real crimes to set them straight - they need to be reminded that real crimes can cause permanent physical and psychological injuries beyond what can be quantified in dollar value while copyright infringement only cause vague and vastly exaggerated hypothetical revenue loss.
Since the RAM used to process HTTP requests usually gets reallocated and rewritten every couple of miliseconds, the data is usually gone long before you even get a chance to cause a kernel panic or memory dump to prevent further memory alterations.
It is impossible to comply literally with that order: it is fundamentally incompatible with the normal operation of RAM/SRAM which are meant to store a computer's working data set only for as long as it is necessary to accomplish the task at hand... and even within that task, countless memory reallocations and modifications (data destructions) are bound to occur.
Computer data, be it RAM or HDD, gets reallocated/modified/destroyed as a matter of course during normal operation... computers cannot operate without continuously destroying some data. Building computers with WORM memory would be ludicrously impractical and unfeasibly expensive. Compliance with the literal interpretation of this order to any useful extent is physically impossible since each server would need its own storage media factory to supply the ~100k DVDs/day needed to dump all memory writes.
Of course, we all know what the judge really meant to order is the logging of relevant data... he just did not know anything about how real-world computers work and TS failed at educating him. It is quite scary how judges can issue such unreasonably uneducated orders.
May I suggest that you check the fundamental definition of 'animal'... the most basic definition (animant) includes anything capable of moving on its own so humans are simply another kind of animal. Human DNA does not really mysteriously/intrinsically separate humans from the rest of the animal kingdom, what does is our unmatched (at least among Earth's creatures) ability to learn and our unique ability to work with abstractions.
Morals, religions, education, etc. are the only things that truly sets humanity apart. If a human babies somehow survived without any human tutoring, they'd pick up most of the behavior of whatever is caring for them - humans have the most degenerated ancestral memory of all animals so practically all the behavior we end up with is forged from environmental (social and other) exposure.
As far as nature is concerned, humans are simply abnormally successful animals.
The DMCA copyright amendment does not make DRM itself as a technological measure any more copyrightable. While DMCA in the USA is an extension of copyright law, things are different elsewhere. In any case, only the protected content is.
Most current DRM schemes rely on common cryptographic algorithm and secret keys, nothing copyrightable and very little (if anything) patentable there. Uncover the root keys and you can blow all DRM wide-open because pretty much everything else is non-patentable/copyrightable public knowledge.
CSS was never protected under copyright law... it was protected under DMCA and any other relevant law, industrial secrets, patents and licensing. Same thing for AACS. What is copyrighted is the protected content, not the DRM.
The ruling only means the court will not protect obsolete, weak or otherwise largely compromised DRM schemes. This is a good thing as it will prevent companies from seeking DMCA-like protection for daft/dummy/broken DRM schemes.
There was a similar technology using magnesium or some other metal a few months ago.
The catch with all these little alternative is that most promoters conveniently leave out the details of how much power and other costs are involved in manufacturing and recycling the spent metals when addressing the general public in promotional material.
When auto manufacturers and reviewers praise the fuel-efficiency of hybrids, they conveniently leave out the fact that the battery pack needs replacement every few years. Calculate your cost per mile after factoring in the cost of replacement batteries and you get something roughly equivalent to plain gasoline, possibly worse. We do not see much criticism of this yet because there are very few hybrids right now that are much over one year old so their batteries are still close to top-shape. In two or three years from now, performance will start to go down and we will see reports of people complaining about the cost of replacement batteries and how the reseller did not "properly" educate them about this unavoidable issue before closing the deal.
As much as people may not fancy having a tank of liquid hydrogen in their cars, I still think this is the most viable option until major battery technology break-throughs enable higher capacity (kWh/weight/volume), higher performance (higher currents, lower impedances) and longer-lived - or at the very least cheaper - cells.
I proposed 100 years on the basis that should I ever publish something of cultural value, I would like to have at least some level of control on how it gets used at least throughout my lifetime.
If I wrote a book, I want to still be entitled to my fair share of the profits should a movie be made of it 50 years later - only a very small minority of books become overnight successes that may attract the immediate attention of movie studios, less spectacular titles are even less likely to get that attention. For movies, if I distributed a popular movie on Beta/VHS I would most likely want to remaster it if the cost can be justified and sell it again on DVD, HD-DVD, Blu-Ray and anything else that may come in the future for as long as there is a public for it - that's pretty much what Lucas has been doing with Star Wars.
Making a living off writing is extremely tough for those whose names are still unknown - I can easily imagine them going suicidal/homicidal should a movie studio make blockbuster movie off books authors actually lost money on nearly immediately after the copyrights expired. Life + X years at least guarantees that authors will never have to face this situation and it guarantees that the author's heirs will not have to face the insult while still mourning the author.
Your vision of copyright appears to be far too short-sighted. As far as books, movie, music and other similar entertainments are concerned, the content retains tangible immediate (as-is) economic value (without impeding the work of others like perpetual patents would) for several years beyond initial publication.
Well, I'm in Canada... fatal cyclist accidents are currently around 150/year so that makes it more like at least 1.3% of all country-wide fatal bicycle accidents last year happened on that boulevard. Since that boulevard is the only way to cross the few highways and rail tracks (without doing a 5+ km detour only to end up on yet another high-traffic boulevard or highway service lanes) that stand in the way of me getting to work, avoiding it is a practical impossibility. Taking the subway would "work" but the detours (due to lousy PT service outside the core downtown area) it imposes make bicycle-only faster by about 30 minutes. Depending on traffic, bicycle is even competitive with going by car during rush hours: PT(sub+bus)=1h00-2h00 (1h20 typ.), bicycle=0h50, car=0h30-1h00 (0h40 typ.).
Lets continue the argument of perpetual copyright compared to ownership of physical objects... physical objects become damaged, obsolete and eventually disposed of. Works of art aside, physical objects have very much finite existences.
Also, even if copyrights were perpetual, it is very likely that descendants would eventually get rid or forget about their rights to their ancestors's work.
IMO, life (up to the last living actor/director/writer/etc. involved in a team production) + 25 years should be plenty sufficient... or maybe a flat 100 years from initial publications.
Most IR remote controls do not have IR receivers... and even if they do, they are there only to record IR patterns from other remote controls and this function is disabled unless the user triggers a key programming event. They most likely have no signal suitable for reliable remote triggering.
Network signal meters are also not really suitable as remote triggers since you get no control over the trigger mechanism. The whole point of the cell-phone, pager or any remote-controlled approach is to simply wire the speaker/vibrator/whatever to the trigger circuit and call the phone/pager/whatever to trigger. Network (non-)detection would be more like a backup self-destruct trigger in case the signal gets unexpectedly jammed.
Using alternative RF bands will work if the deployed jamming is targeted at specific service frequencies... but if the jamming device is a high-power Tesla coil or better broadband noise sources, the reliability of dumb RF triggers (accidental or failed triggering) would be substantially undermined.
Unregistered prepaid cell phones were convenient off-the-shelf addressable remote triggers. New procurement requirements mean terrorists will have to seek similarly convenient and reliable replacements from other sources - but they certainly could fall back to less reliable previous remote trigger schemes in the meantime like tuned high-Q LC oscillators.
I just find the precious irony and total lack of perspective here delicious. Kind of funny that you should write that... I am starting to think you are too frosty to read properly.
Swap alcohol and pot, swap drinkers and pot-heads, swap legality issues, the picture would not be much/any prettier.
Last time I checked, pot was not classified as a hard drug.
BTW, drug addicts rarely commit crimes while they are stoned... they commit them when they are nearly "sober", have run out of stock and start craving their next fix. They start with petty theft and depending on the addiction's severity, it can escalate all the way to home invasions and homicides - alcoholics rarely go that far.
Cases of lethal pot DUIs are not unheard of
If you consider the totals for alcohol incidents on one side and junkies on the other, the tallies of crimes and fatal incidents sides against the junkies when proportions are considered.
Alcohol-related incidents get more media attention in large part because they are more frequent and sometimes spectacular. Narcotics&all get less attention because fewer addicts/abusers means fewer (sometimes spectacular) cases to report so the whole issue fades into the collective background until the next big junkie case related to serial home invasions.
Do you seriously think the world would be better off if we replaced alcoholics by an equal number of pot addicts?
I do not hear about alcoholics committing robberies to get something to drink remotely as often as I see stories of junkies doing so for crack, heroine and other hard drugs... even though there probably are well over 100 alcoholics (to various degrees) for every hard-drugs junkie out there.
These guys often have a toe or more in organized crime and are far more likely to sell your banking and credit information to get their fix than your average alcoholic.
If I was a terrorist and I had to work around RF jamming, one possibility would be to use IR sensors and IR lasers/diodes. Not quite off-the-shelf like cell phones but it would be unscramblable by anything short of an EMP. The detector itself would be a wired remote to the actual detonator so the payload would not have to lie in plain sight. Add a small microcontroller and you get the ability to remotely program detonation delays so you can dispose of the triggering device before the fireworks start.
The major shortcoming would be the limited triggering range and suitable transmitter locations due to the LoS requirement.
Ultrasonic and spread-spectrum triggers would be more complex but definitely doable with DSP or FPGA+DAC... but the bomb would require somewhat large batteries since they would most likely have to be set several days ahead of the target's arrival.