It's usually not. Crystal oscillators have some temperature dependency, and this is a major reason PC clocks drift in an inconsistent way. A 10 degree C temperature change typically results in about a minute of clock error per month. See
"Crystal considerations for Dallas real time clocks
I used to be. And I knew Dave Mills, who designed NTP.
There are quite a number of problems in time synchronization. Here are a few of them.
Authoritative source You have to get time from a better source than you have locally. So you need some scheme to identify time source quality. And it has to detect and prevent loops, so you don't get a group of machines all synchronizing to each other and behaving as if they're synchronized to a primary standard.
Overhead You don't want to be hammering some primary time standard with requests at a huge rate. So the number of requests must be held down.
Drift Clocks, other than primary time standards, drift. That drift has to be measured, so you can tell how often a correction is necessary.
Lag There's propagation delay on time requests, so there's lag. Less than there used to be, but still, lag. That has to be compensated for.
Monotonicity Many programs behave badly if time moves backwards. So correction shouldn't move time backwards unless the change needed is so big that it's unavoidable.
Leap seconds Leap seconds have real consequences. When a leap second is inserted, every AC generator on the grid has to make 60 extra turns. And they do. It takes about four hours for them to catch up.
important areas (downtown, the haight, the fillmore, noe valley, the castro.)
Nah. The "important areas" are SOMA, the Marina, Pacific Heights, China Basin (the new biotech area), Telegraph Hill/Union Square, and Fisherman's Wharf.
Makes you wonder if Microsoft had a role in encouraging its release, doesn't it?
It's striking how nice the virus writers are to the antivirus companies. Most viruses do just enough damage to require ongoing spending for antivirus tools and upgrades, but not enough to make users switch to, say, Linux. There are exceptions, like the virus that encrypts data on the hard drive and demands payment in E-gold, but those are very rare. Few viruses erase data. Few do things that would make removal impossible without physically opening the computer, like modifying the BIOS so it can only boot from the hard drive.
The mainstream viruses seem to be carefully tuned to optimize the revenue stream of antivirus and upgrade vendors.
Many older US cities do have private libraries.
The Mechanics Institute in San Francisco is a very pleasant little library. It's well kept, and the books, almost all on open shelves, are in excellent condition. The collection is broad, and the older material is heavily engineering-oriented. They have the engineering drawings for the Panama Canal, and bound volumes of most of the journals devoted to heavy engineering.
If you spend much time in SF near the Financial District, it's worth buying a membership. It's only $95 per year.
Right. It's quite possible to build a system, such as QNX, where there's a small microkernel that changes very little. All it really does is manage memory, timers, interprocess communication, and CPU dispatching. It can be placed in ROM if desired, and in embedded systems, it often is.
Everything else, including drivers, file systems, and networking, is a user program.
This doesn't limit extensibility. USB and FireWire support were added without kernel changes. (Hardware drivers do need some privileges, but they really are user programs. You can kill them without a reboot and start another copy. I've written and debugged drivers without rebooting.)
Even hardware drivers work through the standard POSIX API; there's no special "driver mode".
If Microsoft designed a system that modular, removing, say, Internet Exploder or Media and DRM Player would be trivial. Microsoft needs a tangled system to lock in users. Ballmer has admitted this.
A secure microkernel is quite possible, but, as Ballmer once said, "If we stopped adding features to Windows, it would become a commodity, like a BIOS. And Microsoft is not in the BIOS businees".
That's so Hollywood. You can waste vast amounts of time doing pitches or being pitched at. You can have lunch at good restaurants. None of this necessarily means anything is actually going to happen.
In some circles, this is called "development hell".
I used to sell software to Hollywood companies. It's amusing. Projects in development have trouble coming up with a valid credit card number. Projects in production want new features yesterday.
Me too. I run Mozilla with "Ask me before accepting a cookie", and usually click "Deny", with "Apply to all cookies from this site". Unless I'm planning to log into a site, or using its shopping cart, I don't click "Accept". All the cross-site tracking services were thus blocked long ago.
This throws some sites into annoying infinite redirect loops, but that's rare.
This isn't quite as bad as it sounds. First, the basic requirement is that VoIP services which interconnect to the wireline network must forward 911 calls into the wireline 911 network, along with enough caller location data to route the call. This only applies to 911 calls. The caller location data is just "the original location at which service was provided". If the system allows the caller to move around, the end user must have the option to update their location information.
But, as yet, the VoIP service is not required to track their users.
Automatic routing of cellular 911 calls was introduced because manual routing worked very badly. California used to route all 911 calls from cell phones to the California Highway Patrol.
As cell phones became more common, CHP dispatch was overwhelmed. By 2002, the CHP was getting over 8 million calls a year, most of which didn't involve freeway incidents, which is most of what the CHP handles. Call hold times on 911 were reaching 10 minutes during peak periods. The CHP was running a huge call center, which basically asked where callers were and forwarded their calls to some local 911 dispatch center.
That's the background for cellular 911. It's convenient that the dispatcher gets the location of the caller, but the real benefit is that the call gets sent to the right dispatcher.
If 911 routing isn't automated for VoIP, where should the calls go? Some call center in Bangalore? If the VoIP provider doesn't have some clue where the caller is, that's about all they can do.
There's worse stuff than this going on.
The extension of the "Commmunications Assistance to Law Enforcement Act" rules to VoIP is much more of a Big Brother thing. If you aren't aware of how this works, the basic concept is that wiretapping has been built into the phone system, and wiretaps are now delivered to law enforcement over T1 lines. The US wiretapping system is run by Verisign.
That's being extended to VoIP.
What basically went wrong is that we've gone half a century with no major improvement in power sources.
That was unexpected.
In 1905, we had steam, coal, and electricity in major cities. Electricity was DC and was generated locally.
Internal combustion engines were just starting to work, Cities were still full of horses. Cars were rare.
The first big power plants were only a few years old. The Wright Brothers had just flown their first aircraft.
In 1955, jet aircraft were zooming around.
The sound barrier had been broken. Hundreds of thousands of airplanes were flying. Cars were everywere. Electricity was everywhere in the developed world. Power grids were nationwide, and
high-tension lines carried gigawatts.
Gas turbines worked. The first nuclear reactors were just starting up. The A-bomb and H-bomb had been conclusively demonstrated, and fusion power seemed within reach. Oil and natural gas had been discovered in many places, and coal was starting to be phased out. Rockets were reaching into space, and the ICBM was very real. The first solar cells had been made.
In 2005, it's about like 1955 in the high-power world. We have all the stuff we had in 1955. in slightly better versions, and that's about it. Fusion didn't work. Nuclear power was more hazardous than expected. Solar cells got only slightly better. Nothing really new came along.
And now, the oil and natural gas is running out.
Science fiction from the 1950s and 1960s assumed that new power sources would be developed, as they had been in the previous century. Didn't happen.
That's why we're not in space. Chemically fueled rockets just barely have enough energy density to do it at all, and not enough to do it well. We need more power!
Right. It's called "remotely determined gaming". Often this is done for regulatory reasons. A slot machine is considered "class 3 gaming", while a lottery terminal is considered "class 2 gaming", and is allowed in more places. So there's a trend towards packaging up remote lottery terminals to look and act like slot machines. You're actually playing against the other players active at the moment, not against the house.
Bally has
a line of Class 2 games.. They look like slots. But they're remotely determined pari-mutual systems.
Stirling Energy has never done anything but engineering prototypes. They sell no products. They've been in business for a decade. They bought the technology after McDonnell dumped it in the 1980s; they didn't develop it.
From their web site:
"The Stirling Genset product lines are still under development and projected for market introduction in late 2003." They can't even keep their vaporware hype up to date.
If they actually sold a mirror dish/engine system and had a few real installations (not DoE-funded demos) they might be worth taking seriously.
There's no problem building a solar powered Stirling cycle engine. It was first done over a century ago. Toy sized ones are available. Getting out enough power to make it profitable, though, is hard.
Reality is that they've installed six dishes at Sandia, with a peak output of 150KW. Southern California Edison is talking about installing 40 dishes, with a peak output of 1MW. If those work, maybe the 20,000 dish gigawatt facility might happen.
There have already been two big solar projects in Southern California Edison territory, called Solar One and Solar Two. Both were so expensive to operate that they couldn't even cover their operating costs, let alone their construction costs.
VitalSource Technologies has been selling electronic textbooks since 2000. But almost all their titles are in the dental field, with some legal titles.
They do have a deal with IBM to preload ThinkPads with their software, and they offer 2000 public-domain titles, presumably equipped with DRM.
But it never caught on.
It doesn't have to be great, it doesn't even have to be good, it only has to be good enough.
Unfortunately, no. Chemically fueled rockets are just barely capable of making it to orbit. They're mostly fuel tankage. Single stage to orbit craft must have at least a 90% fuel fraction. At least. Any serious inefficiency or weight growth kills the design, as happened for Rotary Rocket.
Staging helps. Two stages will get you to low earth orbit. Beyond low orbit usually requires three. This reduces the fuel fraction, but by less than one would hope. The Shuttle's fuel fraction is around 89%.
So space flight is all about weight reduction. Which is why everything is so fragile and unreliable. If you could build a launch system with a fuel fraction of 50%, which is roughly where most aircraft live, it would be a straightforward job.
Because I'm a VA board member, under SEC regulations there's a six-month lockout on the shares (a regulation designed to keep people from floating bogus offerings, cashing out, and skipping to Argentina before the share price crashes). So it's not strictly true that I'm wealthy right now. I will be wealthy in six months, unless VA or the U.S. economy craters before then. I'll bet on VA; I'm not so sure about the U.S. economy:-).
Ah, yes. Right after the SEC cut the holding period. Until 1997, you usually had to hold restricted stock for two years. The reduction from two years to six months fueled the dot-com bubble.
LNUX launched at $239.00. Six months later it was at $34.00. At two years, it was at $3.21.
So the insiders made money, but nobody else did.
There's a little-known book, Japan's Longest Day, which describes the decision-making process in the Japanese government that led to the surrender.
The Tokyo government was in turmoil. One of the key decision makers was in the process of committing suicide, but hadn't yet done so, paralyzing the War Ministry. Some troops tried to steal the Emperor's recording of the surrender message. One faction of the military wanted to fight on. There was a coup attempt. Other troops tried to take over the radio station. It wasn't clear who was in charge.
But respect for the Emperor carried the day, and finally ended the war.
Re:Serious Question - Number of Nukes in 1945?
on
60 Years Since Hiroshima
·
· Score: 3, Interesting
The next bomb was ready to go and was about to be shipped off to the Pacific theater. Gen. Leslie Groves, head of the Manhattan Project, decided to hold off on shipping it. The production rate at that point was one every three weeks or so, and production was picking up.
The big uranium and plutonium extraction plants were up and running by the end of WWII. Those plants were way overdesigned; over several decades, they produced the materials for about 20,000 bombs.
Neither Groves nor Marshall expected to win the war with just two bombs. The plan was to use about thirteen just to "soften up" the landing zones for the invasion of Japan.
After the war, there was a short period during which the US didn't have any working A-bombs in inventory. The original ones were really prototypes, with no shelf life, no safeguards, and a need for an expert to tend and arm them. It took a while to develop a ruggedized, safe to handle "GI-proof" A-bomb.
Microsoft breaks old code all the time. Sometimes they provide backwards compatibility hacks, like ".pif" files. But it's not like, say, Unisys mainframes, where 30-year old executables still run.
Slashdot Mirror
I went to the site, but couldn't find the content. Just an introduction and ads. Lots of ads.
It's usually not. Crystal oscillators have some temperature dependency, and this is a major reason PC clocks drift in an inconsistent way. A 10 degree C temperature change typically results in about a minute of clock error per month. See "Crystal considerations for Dallas real time clocks
I used to be. And I knew Dave Mills, who designed NTP.
There are quite a number of problems in time synchronization. Here are a few of them.
Nah. The "important areas" are SOMA, the Marina, Pacific Heights, China Basin (the new biotech area), Telegraph Hill/Union Square, and Fisherman's Wharf.
Enjoy.
It's striking how nice the virus writers are to the antivirus companies. Most viruses do just enough damage to require ongoing spending for antivirus tools and upgrades, but not enough to make users switch to, say, Linux. There are exceptions, like the virus that encrypts data on the hard drive and demands payment in E-gold, but those are very rare. Few viruses erase data. Few do things that would make removal impossible without physically opening the computer, like modifying the BIOS so it can only boot from the hard drive. The mainstream viruses seem to be carefully tuned to optimize the revenue stream of antivirus and upgrade vendors.
Somewhere there's a reason for this.
If you spend much time in SF near the Financial District, it's worth buying a membership. It's only $95 per year.
Everything else, including drivers, file systems, and networking, is a user program. This doesn't limit extensibility. USB and FireWire support were added without kernel changes. (Hardware drivers do need some privileges, but they really are user programs. You can kill them without a reboot and start another copy. I've written and debugged drivers without rebooting.) Even hardware drivers work through the standard POSIX API; there's no special "driver mode".
If Microsoft designed a system that modular, removing, say, Internet Exploder or Media and DRM Player would be trivial. Microsoft needs a tangled system to lock in users. Ballmer has admitted this.
A secure microkernel is quite possible, but, as Ballmer once said, "If we stopped adding features to Windows, it would become a commodity, like a BIOS. And Microsoft is not in the BIOS businees".
In some circles, this is called "development hell".
I used to sell software to Hollywood companies. It's amusing. Projects in development have trouble coming up with a valid credit card number. Projects in production want new features yesterday.
This throws some sites into annoying infinite redirect loops, but that's rare.
Automatic routing of cellular 911 calls was introduced because manual routing worked very badly. California used to route all 911 calls from cell phones to the California Highway Patrol. As cell phones became more common, CHP dispatch was overwhelmed. By 2002, the CHP was getting over 8 million calls a year, most of which didn't involve freeway incidents, which is most of what the CHP handles. Call hold times on 911 were reaching 10 minutes during peak periods. The CHP was running a huge call center, which basically asked where callers were and forwarded their calls to some local 911 dispatch center.
That's the background for cellular 911. It's convenient that the dispatcher gets the location of the caller, but the real benefit is that the call gets sent to the right dispatcher.
If 911 routing isn't automated for VoIP, where should the calls go? Some call center in Bangalore? If the VoIP provider doesn't have some clue where the caller is, that's about all they can do.
There's worse stuff than this going on. The extension of the "Commmunications Assistance to Law Enforcement Act" rules to VoIP is much more of a Big Brother thing. If you aren't aware of how this works, the basic concept is that wiretapping has been built into the phone system, and wiretaps are now delivered to law enforcement over T1 lines. The US wiretapping system is run by Verisign. That's being extended to VoIP.
In 1905, we had steam, coal, and electricity in major cities. Electricity was DC and was generated locally. Internal combustion engines were just starting to work, Cities were still full of horses. Cars were rare. The first big power plants were only a few years old. The Wright Brothers had just flown their first aircraft.
In 1955, jet aircraft were zooming around. The sound barrier had been broken. Hundreds of thousands of airplanes were flying. Cars were everywere. Electricity was everywhere in the developed world. Power grids were nationwide, and high-tension lines carried gigawatts. Gas turbines worked. The first nuclear reactors were just starting up. The A-bomb and H-bomb had been conclusively demonstrated, and fusion power seemed within reach. Oil and natural gas had been discovered in many places, and coal was starting to be phased out. Rockets were reaching into space, and the ICBM was very real. The first solar cells had been made.
In 2005, it's about like 1955 in the high-power world. We have all the stuff we had in 1955. in slightly better versions, and that's about it. Fusion didn't work. Nuclear power was more hazardous than expected. Solar cells got only slightly better. Nothing really new came along. And now, the oil and natural gas is running out.
Science fiction from the 1950s and 1960s assumed that new power sources would be developed, as they had been in the previous century. Didn't happen.
That's why we're not in space. Chemically fueled rockets just barely have enough energy density to do it at all, and not enough to do it well. We need more power!
Symantec copied the DOD definition of "force protection condition (formerly called THREATCON, now called FPCON) and hacked it into their ThreatCon definitions. It's not really relevant, but it sounds official and impressive.
Bally has a line of Class 2 games.. They look like slots. But they're remotely determined pari-mutual systems.
They even run Linux on the latest models.
If they actually sold a mirror dish/engine system and had a few real installations (not DoE-funded demos) they might be worth taking seriously.
There's no problem building a solar powered Stirling cycle engine. It was first done over a century ago. Toy sized ones are available. Getting out enough power to make it profitable, though, is hard.
There have already been two big solar projects in Southern California Edison territory, called Solar One and Solar Two. Both were so expensive to operate that they couldn't even cover their operating costs, let alone their construction costs.
NXSYS is interesting for historical reasons - the user interface it emulates is the very first "intelligent user interface". From 1937.
They do have a deal with IBM to preload ThinkPads with their software, and they offer 2000 public-domain titles, presumably equipped with DRM. But it never caught on.
Hundreds of tourists must take similar photos every day.
Unfortunately, no. Chemically fueled rockets are just barely capable of making it to orbit. They're mostly fuel tankage. Single stage to orbit craft must have at least a 90% fuel fraction. At least. Any serious inefficiency or weight growth kills the design, as happened for Rotary Rocket.
Staging helps. Two stages will get you to low earth orbit. Beyond low orbit usually requires three. This reduces the fuel fraction, but by less than one would hope. The Shuttle's fuel fraction is around 89%.
So space flight is all about weight reduction. Which is why everything is so fragile and unreliable. If you could build a launch system with a fuel fraction of 50%, which is roughly where most aircraft live, it would be a straightforward job.
Ah, yes. Right after the SEC cut the holding period. Until 1997, you usually had to hold restricted stock for two years. The reduction from two years to six months fueled the dot-com bubble. LNUX launched at $239.00. Six months later it was at $34.00. At two years, it was at $3.21. So the insiders made money, but nobody else did.
There's a little-known book, Japan's Longest Day, which describes the decision-making process in the Japanese government that led to the surrender.
The Tokyo government was in turmoil. One of the key decision makers was in the process of committing suicide, but hadn't yet done so, paralyzing the War Ministry. Some troops tried to steal the Emperor's recording of the surrender message. One faction of the military wanted to fight on. There was a coup attempt. Other troops tried to take over the radio station. It wasn't clear who was in charge.
But respect for the Emperor carried the day, and finally ended the war.
The big uranium and plutonium extraction plants were up and running by the end of WWII. Those plants were way overdesigned; over several decades, they produced the materials for about 20,000 bombs. Neither Groves nor Marshall expected to win the war with just two bombs. The plan was to use about thirteen just to "soften up" the landing zones for the invasion of Japan.
After the war, there was a short period during which the US didn't have any working A-bombs in inventory. The original ones were really prototypes, with no shelf life, no safeguards, and a need for an expert to tend and arm them. It took a while to develop a ruggedized, safe to handle "GI-proof" A-bomb.
Two decades after DOS, the cruft is too deep.