Depending on geography and tower placement, you might be able to get a decent signal with a CDMA, PCS or AMPS phone...
Borrow phones from friends and make some test calls to see how they do. I wouldn't just go by number of bars.
Other possibility is to rig up a directional antennna and plug your phone into that (assuming it has an aux antenna jack). You won't be able to walk around the house and talk, but at least you'll have service.
The original nightshot cameras occasionally show up on ebay, and there are mods you can do to the current nightshot cams to get the same functionality...
More seriously, if the flight crew becomes incapacitated there would be no way for anyone on board to get to the cockpit to assist them, call for help, or fly the damned thing.
>> it is apparently possible to sit a table in >> a U.S. cafe without ordering anything.
Café's in the US are usually corporate or franchises and are staffed by low-wage workers who would rather be somewhere else. They aren't really interested in acting as security guards. They don't get tips for throwing people out.
The other thing is, in America, some non-zero percentage of people carry guns. The default is to not confront anyone unless it is your job to do so, and you are also carrying a gun. I'm not exaggerating.
If the 2 hr token code on receipts is too much trouble, one thought would be a wifi router that takes note of the mac address and gradually throttles down bandwidth or puts up a nag message on port 80 requests after the mac address has been in use for too long. Could even be a variable sized time window so that on those busy weekends you only get 1/2 hr but on a slow monday afternoon you get 4 hours.
I don't know of any wifi routers that do this, but it seems like an easy idea to implement and wouldn't require anything from the cash register or employees.
Sure, people can spoof their mac address and get another 2 hrs (or whatever) but it most people wouldn't bother and it sounds like they don't mind a few people hanging out, they just don't want a sea of laptops.
The problem with blocking 25 and moving to other ports is that guess what, several spammers read slashdot and they just added some extra code to their bots to check for SMTP on 80 and 3535 also, and a plain old port scan if they don't find anything right away.
>> [regarding remaster digital recordings] >> How can you get 24 bit sound resolution out >> of 16 bit source material?
a) You cannot; however...
b) A lot of early CD's were recorded with poorly noise-shaped dither, or worse none at all. The first two generations of CD players took this into account and filtered off the high frequencies very aggressively. I have no proof, but I think somewhere around 1990 somebody figured out how to detect the dither quality on a CD and adjust the filters accordingly. My point is that some of the remasters of already-digital material are just to correct the dither problem.
c) The session tapes probably have >16 bits available, plus you get a few bits when you mix stuff down. If the master was rounded down to 16 or 18, it's not a bad idea to remaster before releasing to SACD or DVD-Audio.
d) Compression. In spite of digial recording's increased dynamic range, it's still common (at least for pop/rock) to make your digitally mastered CD's compatible with FM radios and car CD players and walkmen and all the other sub-audiophile gear out there. A digital remaster offers the opportunity to remove this unfortunate but common design choice.
e) Marketing fluff. A remastered CD of a popular recording can list for 1.5x or 2x the price of the original. Even if it's silly idea, as long as they can sell 5000 copies at a higher price w/o having to sober up the artist and pay him another million, it's a go, especially if it looks like the artist isn't going to be putting out any new material for a while.
f) I suspect that there is some behind-the-scenes prayer to the effect high-bit rate recordings will become the new "standard" and the sheep/consumers will shell out again to convert over, like they did from LP's to CD's. I think this is futile in the age of MP3's, but I presume music executives are trying very hard to come up with compelling upgrade paths.
g) Expect a ridiculous music format for blu-ray or HD-DVD discs soon...I'm guessing 32 bits at 96khz with 10+ channels available...and expect another round of remastering to follow.
>> That car, like all other cars to date, has a >> direct mechanical connection between the steering >> wheel and the angle of the front wheels.
I guess I'll correct myself before someone else does it for me...
(a) There have been cars that steer with other methods, i.e. rear wheels instead of the front or by using tank tracks-like differential steering. These are quite uncommon, but it only takes one counterexample to disprove an "all" statement.
(b) Steering is not direct in the sense of one single rod all the way to the wheel hub. Most cars (I won't say all this time) have a steering gearbox, rack/pinion or other steering rack arrangement, and the forces from the driver turning the wheel go through a couple of bushings/shock absorber like things for good measure, on top of all that is the power assist. I should have said something more along the lines of there is a complete mechanical path of control connecting the steering wheel to angle of the (usually) front wheels.
This connection is fail-safe in that it functions at least partially even if all other vehicular systems fail.
Steer-by-wire might fail completely if the electrical system fails, or a software glitch might cause uncommanded steering, or a noisy servo feeds bogus information to the controller, all kinds of stuff, and we don't have 100+ calendar years of engineering experience to draw on the way we do with mechanical steering.
>> I have a 1999 Volvo S80 [...] >> The steering is even drive by wire.
I think this is only partially correct. That car, like all other cars to date, has a direct mechanical connection between the steering wheel and the angle of the front wheels. That connection is the primary means of changing the car's direction. You should be able to observe this by turning the ignition to the unlocked (but off) position and observing that the front wheels still budge when you turn the wheel.
Volvo does have drive by wire for their outboard motors and for their line of forklifts, but the technology is not in cars yet; probably for legal reasons it has to be 100% reliable before they'll risk it.
There is one other sense in which you are correct: when the stability management thing intervenes, it is capable of steering the car to some degree, however it does so by activating the brakes individually or changing the throttle position not by turning the wheel. It's also possible that the power steering unit has electronically controlled boost under control of the stability management thing and tones down the boost when you're already oversteering. But these systems (braking and boost reduction) would not come into play in normal driving.
I would add one more: as a major investor, they can push for access to internal and strategic information, then use this information to formulate a counter-strategy much earlier than if they had to wait for press releases about what redhat was up to.
I suspect Bill and/or Steve have basically freaked out over linux, google, apple's ipod, whatever, and hired 100+ managers to just make sure microsoft doesn't get taken by surprise on every new gadget or service.
It's an end-stage strategy--since msft is so big and slow to react, they're trying to (almost literally) buy time.
>> a professional one with some better power
>> management it could be used 24/7 too
With some proper engineering, I believe that a solar system can be more reliable than edison. A top-quality inverter can draw power from solar panels, a battery bank, start up a generator remotely, and if all else fails, THEN draw power from edison as a last resort. If the equipment is reliable and well-maintained it's not out of the question to expect 5 to 10 years between outages. Further, since the owner/user of the system is in control, intentional outages for upgrades/repairs can be timed for convenience.
However, in the 17w mac server situation, it's probably not cost effective. At edison rates in my area, 17w works out to $1.80 per month. The smallest, cheapest starter system at
backwoods solar electric systems is $800.
$800 ÷ $1.80 = 444 months = 37 years for payback of the initial investment. This is beyond the usual 25 year lifespan for solar panels and way beyond the 5 year lifespan for lead-acid batteries.
I've always been intrigued with solar power, but I just can't make the numbers work. I'm hoping to power up some outbuildings with solar, but these will be really cheap DIY jobbies that don't have to be 24/7 more like 5/2. This can actually be cost effective because it saves having to run buried cable, which is very expensive.
People should keep in mind that Mr. Kircaali really doesn't have the option of fully admitting and apologizing for anything. That would just open up him and his company to a giant lawsuit.
He has to forcefully deny any wrongdoing to remove the possibility that at a later trial, a lawyer could just just hand the apology/admission to a jury and say "Here's the evidence, he admitted to it, please make them give PJ $1 (holds pinky to lip) MILLION dollars"
It appears to me that sotfware has finally gotten "componentized" enough that any given employer of a software engineer will probably focus their development efforts into a fairly narrow marketing niche. Sure, msft and google can afford to explore large swaths of interesting technology, but even then a given worker in a given department is intentionally compartmentalized. Anyone that can see outside their little area of expertise is either fired for being difficult or promoted to management if they can at least pretend to drink the kool aid.
There seems to be very little market for highly educated and well-versed software engineers. The trend is for "coders" that just connect other people's code together with VB or PHP. Businesses would rather pay $2K for a COM+ object and 10c/transaction for a web service and $30/hour for an offshore contractor to connect everything up as opposed to $250K/year for a team of knowledgeable software engineers that could do everything in house.
I've spent many afternoons reading and at least partially understanding printed matter that says "proceedings" or "journal" or "SIG" somewhere on the cover. It was fun in and of itself, but I can't exactly put it on my resume.
Fine print says: The TDI-PD is a diesel engine. Models equipped with this engine are not available for purchase
in California, New York, Vermont, Maine and Massachusetts.
Still it's a good car, with good gas mileage and I look forward to the availability of low-sulfur fuel in the US which should allow more diesels on the road in CA etc.
You probably just punched it from stoplights and ran up to your normal speed and backed off. The extra fuel used because you pushed the throttle farther is partially balanced by the fact that you got to your desired cruising speed sooner. As you indicate, if you aren't a traffic jam you're probably spending most of your time cruising at 65 or so.
Also, the acceleration phase might make up only 5% of your driving time (for example) so a 20% increase in fuel usage during that time may not make a measurable difference in economy.
The smooth-vs-jackrabbit thing makes a bigger difference on cars that have powerful engines. While a cr-v is a perfectly good vehicle, the difference between casual driving and flooring each time the light turns green isn't as dramatic as, say an NSX.
The shell script that accomplishes task x and the assignment of a team of people to accomplish task x really should go through the same level of testing and compliance checking.
In some companies there is actually a resentment on the part of management against "clever" people who automate too much. One place I worked at specifically stated that no shell scripts, aliases or utility programs could be worked on or installed w/o management approval. All the developers ignored this of course and did whatever they had to.
It's easy to see this as a simple grudge between different classes of workers, but I think it actually came down to the fact that automating too much prevented middle managers from being able to justify expanding their headcount. They WANTED to hire some warm bodies to just move files around or to manually type in some commands that really belonged in crontab.
Having an old greybeard type "make" once in a while didn't really score his boss any cred.
Not an expert, but I'd guess that the correspondence problem is simpler for satellite images since you can assume that there are few (if any) discontiguities or obstructions other than, say, clouds or the occasional aircraft.
Dirt is kind of slippery (in the long run) and tends not to hold in odd shapes or overhangs. Of course there are a few cliffs, but in general once a feature is identified, nearby features will only parallax by a few pixels as the terrain rises and falls.
Buildings are generally flat surfaces which meet at right angles. Depending daylight conditions when the picture was taken, you don't even need stereo processing to get building heights--just look at the length of shadows.
Anyway, I'm not saying it's easy, but it's not the full-blown machine-vision while navigating through sagebrush and riverbeds DARPA Grand Challenge level either.
Slashdot Mirror
Depending on geography and tower placement, you might be able to get a decent signal with a CDMA, PCS or AMPS phone...
Borrow phones from friends and make some test calls to see how they do. I wouldn't just go by number of bars.
Other possibility is to rig up a directional antennna and plug your phone into that (assuming it has an aux antenna jack). You won't be able to walk around the house and talk, but at least you'll have service.
The original nightshot cameras occasionally show up on ebay, and there are mods you can do to the current nightshot cams to get the same functionality...
More seriously, if the flight crew becomes incapacitated there would be no way for anyone on board to get to the cockpit to assist them, call for help, or fly the damned thing.
>> it is apparently possible to sit a table in
>> a U.S. cafe without ordering anything.
Café's in the US are usually corporate or franchises and are staffed by low-wage workers who would rather be somewhere else. They aren't really interested in acting as security guards. They don't get tips for throwing people out.
The other thing is, in America, some non-zero percentage of people carry guns. The default is to not confront anyone unless it is your job to do so, and you are also carrying a gun. I'm not exaggerating.
If the 2 hr token code on receipts is too much trouble, one thought would be a wifi router that takes note of the mac address and gradually throttles down bandwidth or puts up a nag message on port 80 requests after the mac address has been in use for too long. Could even be a variable sized time window so that on those busy weekends you only get 1/2 hr but on a slow monday afternoon you get 4 hours.
I don't know of any wifi routers that do this, but it seems like an easy idea to implement and wouldn't require anything from the cash register or employees.
Sure, people can spoof their mac address and get another 2 hrs (or whatever) but it most people wouldn't bother and it sounds like they don't mind a few people hanging out, they just don't want a sea of laptops.
Didn't see groklaw mentioned.
PJ is certainly a celebrity at this point--she even has her own stalker (Maureen O'Gara).
For the record, direcway's main competitor, starband, is also bidirectional through the satellite; no telephone-system connection required.
The problem with blocking 25 and moving to other ports is that guess what, several spammers read slashdot and they just added some extra code to their bots to check for SMTP on 80 and 3535 also, and a plain old port scan if they don't find anything right away.
>> [regarding remaster digital recordings]
>> How can you get 24 bit sound resolution out
>> of 16 bit source material?
a) You cannot; however...
b) A lot of early CD's were recorded with poorly noise-shaped dither, or worse none at all. The first two generations of CD players took this into account and filtered off the high frequencies very aggressively. I have no proof, but I think somewhere around 1990 somebody figured out how to detect the dither quality on a CD and adjust the filters accordingly. My point is that some of the remasters of already-digital material are just to correct the dither problem.
c) The session tapes probably have >16 bits available, plus you get a few bits when you mix stuff down. If the master was rounded down to 16 or 18, it's not a bad idea to remaster before releasing to SACD or DVD-Audio.
d) Compression. In spite of digial recording's increased dynamic range, it's still common (at least for pop/rock) to make your digitally mastered CD's compatible with FM radios and car CD players and walkmen and all the other sub-audiophile gear out there. A digital remaster offers the opportunity to remove this unfortunate but common design choice.
e) Marketing fluff. A remastered CD of a popular recording can list for 1.5x or 2x the price of the original. Even if it's silly idea, as long as they can sell 5000 copies at a higher price w/o having to sober up the artist and pay him another million, it's a go, especially if it looks like the artist isn't going to be putting out any new material for a while.
f) I suspect that there is some behind-the-scenes prayer to the effect high-bit rate recordings will become the new "standard" and the sheep/consumers will shell out again to convert over, like they did from LP's to CD's. I think this is futile in the age of MP3's, but I presume music executives are trying very hard to come up with compelling upgrade paths.
g) Expect a ridiculous music format for blu-ray or HD-DVD discs soon...I'm guessing 32 bits at 96khz with 10+ channels available...and expect another round of remastering to follow.
What I'd like to see is a company that restructures by laying off the most highly paid workers first.
engadget version of story
bbc version of story
>> MSNing for something just seems wrong.
Yeah, it's pretty balmered up.
>> That car, like all other cars to date, has a
>> direct mechanical connection between the steering
>> wheel and the angle of the front wheels.
I guess I'll correct myself before someone else does it for me...
(a) There have been cars that steer with other methods, i.e. rear wheels instead of the front or by using tank tracks-like differential steering. These are quite uncommon, but it only takes one counterexample to disprove an "all" statement.
(b) Steering is not direct in the sense of one single rod all the way to the wheel hub. Most cars (I won't say all this time) have a steering gearbox, rack/pinion or other steering rack arrangement, and the forces from the driver turning the wheel go through a couple of bushings/shock absorber like things for good measure, on top of all that is the power assist. I should have said something more along the lines of there is a complete mechanical path of control connecting the steering wheel to angle of the (usually) front wheels.
This connection is fail-safe in that it functions at least partially even if all other vehicular systems fail.
Steer-by-wire might fail completely if the electrical system fails, or a software glitch might cause uncommanded steering, or a noisy servo feeds bogus information to the controller, all kinds of stuff, and we don't have 100+ calendar years of engineering experience to draw on the way we do with mechanical steering.
>> I have a 1999 Volvo S80 [...]
>> The steering is even drive by wire.
I think this is only partially correct. That car, like all other cars to date, has a direct mechanical connection between the steering wheel and the angle of the front wheels. That connection is the primary means of changing the car's direction. You should be able to observe this by turning the ignition to the unlocked (but off) position and observing that the front wheels still budge when you turn the wheel.
Volvo does have drive by wire for their outboard motors and for their line of forklifts, but the technology is not in cars yet; probably for legal reasons it has to be 100% reliable before they'll risk it.
There is one other sense in which you are correct: when the stability management thing intervenes, it is capable of steering the car to some degree, however it does so by activating the brakes individually or changing the throttle position not by turning the wheel. It's also possible that the power steering unit has electronically controlled boost under control of the stability management thing and tones down the boost when you're already oversteering. But these systems (braking and boost reduction) would not come into play in normal driving.
I would add one more: as a major investor, they can push for access to internal and strategic information, then use this information to formulate a counter-strategy much earlier than if they had to wait for press releases about what redhat was up to.
I suspect Bill and/or Steve have basically freaked out over linux, google, apple's ipod, whatever, and hired 100+ managers to just make sure microsoft doesn't get taken by surprise on every new gadget or service.
It's an end-stage strategy--since msft is so big and slow to react, they're trying to (almost literally) buy time.
>> management it could be used 24/7 too
With some proper engineering, I believe that a solar system can be more reliable than edison. A top-quality inverter can draw power from solar panels, a battery bank, start up a generator remotely, and if all else fails, THEN draw power from edison as a last resort. If the equipment is reliable and well-maintained it's not out of the question to expect 5 to 10 years between outages. Further, since the owner/user of the system is in control, intentional outages for upgrades/repairs can be timed for convenience.
However, in the 17w mac server situation, it's probably not cost effective. At edison rates in my area, 17w works out to $1.80 per month. The smallest, cheapest starter system at backwoods solar electric systems is $800.
$800 ÷ $1.80 = 444 months = 37 years for payback of the initial investment. This is beyond the usual 25 year lifespan for solar panels and way beyond the 5 year lifespan for lead-acid batteries.
I've always been intrigued with solar power, but I just can't make the numbers work. I'm hoping to power up some outbuildings with solar, but these will be really cheap DIY jobbies that don't have to be 24/7 more like 5/2. This can actually be cost effective because it saves having to run buried cable, which is very expensive.
People should keep in mind that Mr. Kircaali really doesn't have the option of fully admitting and apologizing for anything. That would just open up him and his company to a giant lawsuit.
He has to forcefully deny any wrongdoing to remove the possibility that at a later trial, a lawyer could just just hand the apology/admission to a jury and say "Here's the evidence, he admitted to it, please make them give PJ $1 (holds pinky to lip) MILLION dollars"
I get the humor, but just for the record, that would be a Bill of Attainder and unconstitutional.
It appears to me that sotfware has finally gotten "componentized" enough that any given employer of a software engineer will probably focus their development efforts into a fairly narrow marketing niche. Sure, msft and google can afford to explore large swaths of interesting technology, but even then a given worker in a given department is intentionally compartmentalized. Anyone that can see outside their little area of expertise is either fired for being difficult or promoted to management if they can at least pretend to drink the kool aid.
There seems to be very little market for highly educated and well-versed software engineers. The trend is for "coders" that just connect other people's code together with VB or PHP. Businesses would rather pay $2K for a COM+ object and 10c/transaction for a web service and $30/hour for an offshore contractor to connect everything up as opposed to $250K/year for a team of knowledgeable software engineers that could do everything in house.
I've spent many afternoons reading and at least partially understanding printed matter that says "proceedings" or "journal" or "SIG" somewhere on the cover. It was fun in and of itself, but I can't exactly put it on my resume.
Fine print says:
The TDI-PD is a diesel engine. Models equipped with this engine are not available for purchase in California, New York, Vermont, Maine and Massachusetts.
Still it's a good car, with good gas mileage and I look forward to the availability of low-sulfur fuel in the US which should allow more diesels on the road in CA etc.
You probably just punched it from stoplights and ran up to your normal speed and backed off. The extra fuel used because you pushed the throttle farther is partially balanced by the fact that you got to your desired cruising speed sooner. As you indicate, if you aren't a traffic jam you're probably spending most of your time cruising at 65 or so.
Also, the acceleration phase might make up only 5% of your driving time (for example) so a 20% increase in fuel usage during that time may not make a measurable difference in economy.
The smooth-vs-jackrabbit thing makes a bigger difference on cars that have powerful engines. While a cr-v is a perfectly good vehicle, the difference between casual driving and flooring each time the light turns green isn't as dramatic as, say an NSX.
The shell script that accomplishes task x and the assignment of a team of people to accomplish task x really should go through the same level of testing and compliance checking.
In some companies there is actually a resentment on the part of management against "clever" people who automate too much. One place I worked at specifically stated that no shell scripts, aliases or utility programs could be worked on or installed w/o management approval. All the developers ignored this of course and did whatever they had to.
It's easy to see this as a simple grudge between different classes of workers, but I think it actually came down to the fact that automating too much prevented middle managers from being able to justify expanding their headcount. They WANTED to hire some warm bodies to just move files around or to manually type in some commands that really belonged in crontab.
Having an old greybeard type "make" once in a while didn't really score his boss any cred.
Not an expert, but I'd guess that the correspondence problem is simpler for satellite images since you can assume that there are few (if any) discontiguities or obstructions other than, say, clouds or the occasional aircraft.
Dirt is kind of slippery (in the long run) and tends not to hold in odd shapes or overhangs. Of course there are a few cliffs, but in general once a feature is identified, nearby features will only parallax by a few pixels as the terrain rises and falls.
Buildings are generally flat surfaces which meet at right angles. Depending daylight conditions when the picture was taken, you don't even need stereo processing to get building heights--just look at the length of shadows.
Anyway, I'm not saying it's easy, but it's not the full-blown machine-vision while navigating through sagebrush and riverbeds DARPA Grand Challenge level either.
Furthermore, don't businesses have to pay sales tax on the retail value of their own products that they use internally?