The Wall Street Journal writes of "Battling the Cyber Warmongers". The big event in this area seems to have been the Cyber Shock Wave TV special, in which various former senior Federal officials participated in a simulation of a "cyber attack" on the power and phone system. Big names were involved: Michael Chertoff (former head of Homeland Security) and John Negroponte (former acting CIA director).
The show was embarrassing. It was clear that the participants not only had no clue what to do, they didn't know who to call who did. The person representing Energy kept harping on the thousands of energy companies there are in the US, and how they needed more authority over them. The level of cluelessness makes it clear how Hurricane Katrina (for which they had three days warning, and which happened on the watch of many of the participants) was botched.
In reality, the US power grid is divided operationally into seven regional grids, each with a control center and a backup control center. The supervisors at those control centers are the ones who really run things. If someone in the U.S. Government is dealing with an attack on the power grid, they need to have those supervisors on speed dial.
In an emergency, the best thing representatives of the Government can do is call up each one and ask "What do you need right now". They're likely to get an answer like "We need troops protecting these key substations", or "We need a heavy-lift helicopter to move a spare transformer." Actions that would help fix the problem. One of the listed duties of the shift supervisor at the PJM Interconnect is to talk to Government officials.
The Government officials in that simulated emergency didn't have that basic info. They didn't know that there were seven people who were really running things. But they wanted to be in charge. That's the problem.
Few high government officials have a background as first responders or in incident management. If anything, the military officers are more likely to have a clue; their training teaches them to prioritize in a crisis and to deal with confusing, conflicting information.
The ACLU has taken at least two cases in that area.
The Maryland motorcycle case: "This prosecution by the Maryland State Police and Harford County State's Attorney is profoundly dangerous, and seems meant to intimidate people from making a record of what public officials do," said David Rocah, Staff Attorney at the ACLU of Maryland. "It is hard to imagine anything more antithetical to a democracy than for the government to tell its citizens that they do not have the right to record what government officials say or do or how they behave."
I'm on record as having called the dot-com bubble, the oil spike, and the mortgage crisis. It's not hard to predict bubbles. If you look at historical ratios, it's usually clear when assets are overpriced. Historically, the median house in the US sells for 2x to 2.5x the median income. That's about what people can pay for. That ratio hit 4 nationally, and 10 in some states. It was blindingly obvious that there was a housing bubble.
Dot-com predictions were easy. I had a program which read SEC filings for cash and burn rate, and simply projected when the money would run out. This was far more successful than one would expect. I used to get hate mail from CFOs for that.
The problem is figuring out when a bubble will pop. I expected the mortgage bubble to pop about two years earlier than it did. (Arguably, the Fed's cheap-money policy under Greenspan postponed the inevitable.)
Predictions on the debt side are harder than on the equity side. Public policy dominates the debt world. I don't make political predictions, so I can't say much about the current situation. I've been expecting an interest rate spike for years, but instead we've had a Federal deficit spike as money is pumped into the system. Eventually something will give there, as with Iceland, Greece, and other debtor countries. I'm not sure how that will unwind. We may get an interest rate spike and hyperinflation, which is what usually happens when a currency gets into trouble.
Flash itself is really very clever. The player packs an incredible amount of functionality into a very tiny executable. It's only 1.83MB. There's an animation engine, a JIT compiler, a video player, an audio system, and a multichannel download manager.
The problem is what people use it for. Which is mostly either ads or lame web sites.
Nobody really bothers doing Flash animations as entertainment much. If you've never seen one, check out Thugs on Film. Flash games remain popular, although Shockwave, which has full 3D capability, is a far better game platform. Many console games use Flash for 2D interface elements, typically using Adobe's authoring tools but a non-Adobe player. (Yes, there are non-Adobe Flash players.)
But it's not Adobe's fault that the content sucks.
It's sad. C++ has major problems, which are well known. The fundamental problem is that C++ has hiding without memory safety. C++ is unique among major languages in that respect. There are languages with no hiding, (C, assembler) and languages with both hiding and memory safety (Java, Delphi, Go, D, Ada, Modula, Erlang, Eiffel, and all the "scripting" languages).
The C++ committee, and Strostrup, are in denial about this. They're off in template la-la land, trying to abuse the template system into a compile-time interpreter.
C has its own problems. The basic one is that you lie to the language about how big things are. A string parameter is declared as "char *", not "char[n]". So the language has no idea how big objects are. It doesn't require garbage collection or array descriptors to fix this. It just requires a declaration syntax where you can talk about the size of the arrays you're passing around. This fundamental defect in C is the cause of most buffer overflows. (Technical discussion: C should have had array parameter declarations like int read(int fd, char[n]& buf, size_t n); rather than int read(int fd, char* buf, size_t n); Then the language knows how big the array is. This could be extended to structs, etc. so that whenever you have an array, there's an expression known to the compiler that defines its size. This allows subscript checking.)
C++ tries to paper this over with collection classes, but the mold always seeps through the wallpaper. Too often, you need a raw pointer, which breaks what little protection the collection classes provide.
A long time ago, when Project Gutenberg texts were really the only "ebooks" one could find, I had the idea of creating a separate data file that would accompany the.txt files. My idea was to leave the actual content of the book in plaintext for maximum portability, but allow fancy formatting (pagination, font, links, etc) via a separate binary file which would reference the.txt by character position.
"Bravo" for the Xerox Alto, the first multi-font WYSIWYG editor, worked that way. The text was stored as ASCII and terminated with a control-Z. Following the control-Z was the formatting information. Text-only utilities, like compilers, could read the files as plaintext. Late 1970s technology.
It's a real issue if a map system leads people into a trap.
There was a problem in England where a map system led truck drivers down a narrow dead-end lane from which they could not turn around. This was an ongoing problem, and the local property owners raised hell.
The problem is serious in situations where the map system leads people onto an initially plausible route which gets them into trouble. Looking at the pictures for this walking problem, there's a dirt trail alongside the road. But does that dirt trail end before the destination?
Because Google actually drives out roads and takes pictures, people may be led to have more confidence in Google's mapping than in systems which just use ordinary map data.
"Cloud" companies are really just hosting providers. Hosting providers have their ups and downs; quite often they start out OK and get worse. You need to have a second source and a migration strategy.
In 12 years I've been through four hosting companies for one site. The first one started small, was acquired, was acquired again, and was eventually spun off to Earthlink. The second one was a good hosting company until they got into "permission-based email marketing" (i.e. spamming) and went downhill from there. The third one offered both dedicated servers and shared servers, then spun off the dedicated server business, leaving the shared business in bad shape. The fourth one is doing reasonably well right now.
I have a few things that use "cloud" type services. One uses a search API, and I have both Yahoo and Google versions. Another uses an SMS gateway, and I have both Google Voice and Twilio versions.
I own all my own domains, and the domain registrars are in no way affiliated with the hosting providers.
For the important domains, I have registered U.S. trademarks. I've had to switch registrars on one occasion.
All servers are Linux, and all necessary tools are open source.
You have to assume that your suppliers can fail. Stay in this business for a few years, and some of them will fail.
The Tesla Roadster has a 245 mile range. And basic stuff like bumpers.
The student car looks like it has about a 3 inch ground clearance. If that. That's not going to get very far on anything less than a perfect road. And they want to drive it down from Prudhoe Bay? Right.
It's amusing that one of the linked articles mentions an "iPad killer" from Foxconn. Foxconn makes the iPad.
Foxconn's 2008 revenue was $62 billion. They're the "largest exporter in Greater China" and the world's largest maker of phone handsets. They have 486,000 employees. (Apple: 35,000. General Motors: 245,000.)
To see what it was like to use these fonts on the web, I created a test page. This uses dynamically downloaded fonts, which work in most current web browsers. (Firefox users need Firefox 3.6 or later.)
This sample is sized at 16 point. Smaller than that, many of the symbols are unreadable. That's something to be careful about. When you have a huge symbol set, the symbols need to be bigger. However, some of the symbols don't scale up well. If you scale up that page, the integral symbols look great, but the arrows become pixilated. Some of the symbols seem to have been were badly encoded.
This is just a raw demonstration of the font; for formulas, you'd use MathML. I'm not sure if MathML, the W3C names for math characters in HTML, and the STIX fonts are all synchronized yet. But at least you don't have to tell people "to display this page, install all these fonts first."
"Rather than sell electricity or water, though, operators will be taking data to measure impact on sea life, the generator's performance, and the cost of operation, said Douglas Sandberg, the vice president of the privately funded company."
So it's just a demo. Only generates 60KW. Not clear if that's average or max power; probably max. On days with low surf, not much will happen.
They've been hyping this since 2004. There are better wave powered generation devices, and even the best ones are commercial flops.
What idiots. They're shipping an "application" that consists of over 4,000 image files and some XML. You have to download 500MB of stuff to read the magazine. How long is that going to take?
Then there's the content problem. Wired, at this point, is basically a product catalog. Yet they didn't put a shopping cart system in the online version. That's just dumb. The demographic that buys both iPads and Wired would definitely click on "buy now" buttons.
Law enforcement works best when it's complaint driven. When cops are responding to crimes, they're performing a service function.
The quality of that service can be measured. Crime rates, clearance rates, arrest rates, conviction rates, and recidivism rates can all be tallied, and are are politically important.
When law enforcement sets its own agenda, it does less well. It's harder to measure whether a vice squad or a drug squad is doing a good job. Those are the units that tend to be corrupt and/or incompetent. It's also much harder to measure their cost-effectiveness.
This problem shows up in the rhetoric. We have the War on Drugs, the War on Terror, and the War on Kiddie Porn. We don't have the War on Financial Fraud, the War on Child Labor, or the War on Bribery. Which is a problem.
If we were serious about Protecting the Children, child labor laws would be much tougher.
I thought the unpublished speed of the SR-71 was around Mach 6?
No, top speed of the SR-71 was around Mach 3.2, which it could sustain for hours. The absolute top speed ever achieved was 2,193.2 mph. The Museum of Flight at Boeing Field has had a SR-71 on display for years, and they have all the documentation.
The SR-71 powerplant is a turbojet that converts to a ramjet at high speeds.
When Kelly Johnson was designing the SR-71, they considered a liquid hydrogen propulsion system that might have made it to Mach 6 and beyond. Work on liquid hydrogen engines was conducted at Burbank. (That helped out NASA later on.) But they figured out how to do
the necessary job with far less volatile fuels.
The original article is almost devoid of facts. What training is required to speak for "Advice Line"?
It's not at all clear what training is required for IT today? The Cisco "Rack Test"? How to fix broken Windows systems?
J2EE programming? Linux server administration?
CS is even tougher. Robotics? AI? Machine learning? Graphics? Digital logic? "Cloud" programming? There are too many narrow niches. Pick the wrong one and you're toast.
Here's the bottom of the apparel food chain. "Bulk Bale Clothing". $0.25/lb. Minimum order 55,000 pounds. Supply availability 1,000,000 pounds per month. Bulk in 1000 pound bales.
That's just one of a hundred similar suppliers. "We currently have 28 containers of brand name clothing acquired in a bank deal." "250,000 lbs baled used clothing. 25% coats,sweaters, heavy clothing. $0.84/lb."
That's life in the no-IP world of apparel. The wastage is enormous.
Remember a company who dragged their entire OS to Trash, emptied it and restarted with a fresh and open source OS instead of trying to "fix" it?
Apple did that to bail out Jobs, when NeXt was in the tank. Apple had developed a new OS, MacOS 8 ("Copeland"), which was a reasonably good rewrite. The claim was that a warmed-over NextStep could be on the market sooner than Copeland. It wasn't, but the deal saved Jobs' personal wealth.
Fashion depends on churn. There aren't many original ideas. If you look at this year's runway fashion, and are familiar with the history of fashion, you can usually find a matching piece from decades ago. The Fashion Institute of Technology in New York has a large library of clothing against which new designs can be compared. This year, we have khaki (again), Gautier is trying green spandex shorts (80s aerobic wear), and Issey Miyake is over-pleating again (he did that better in the '80s) and using pink accents (so last year.) The jeans industry keeps fussing with various levels of fading, but they've been doing that for so long that nobody is paying much attention.
There's some technological progress, and it gets IP protection. Gore-Tex was patented, and for a long time, had a monopoly over waterproof fabrics that breathe well. Progress in materials, in sewing technology, and in cleaning has led to new ranges of clothing.
Jeans, for example, depend on a sewing technology for strong corners that's only about fifty years old. (Today, rivets in jeans are decorative, not structural.) Sportswear, which was invented by Coco Chanel, wasn't really feasible before washing machines.
Elastic fabrics opened up many new options. Not much new has come along in the last two decades, ("pleather" made a small splash) and fashion technology has somewhat stagnated.
As the TED talk points out, the big thing today is trademark protection via "designer labels". This is a relatively new concept. Until the late 1970s, no respectable garment maker would have the designer's mark visible, let alone a prominent feature of the design. Logos were associated with cheap T-shirts. The interlocking double C now associated with Chanel did not appear on Chanel products until the 1980s.
The apparel distribution pipeline is incredibly inefficient.
Over 60% of apparel is eventually sold on sale. There's a hierarchy. First there's the initial sale, with a big markup. Then there's the sale rack at the original retailer, with the original tags still attached. Then there's the discount retailer who buys from the original retailer and resells from their own store. Finally, the unsold apparel is rolled into big balls about eight feet in diameter, which are rolled into shipping containers and shipped to third world countries for final sale, or recycled into nonwoven fabrics like cleaning cloths. There was an attempt during the dot-com era, called "Tradeweave", to create a secondary trading market in unsold apparel, but it only lasted from 1999-2001.
The watch industry is a branch of the fashion industry now. "We are not in the watch industry, we are are in the luxury industry" says the CEO of Rolex. They had the basic problem that their overpriced machinery is less accurate than a midrange quartz watch, and they now have the worse problem that people who grew up with cell phones see no need for watches.
There have been attempts by the phone industry to do "designer cell phones", but so far, that's mostly a joke. Apple tried to position themselves as a high-end product, but you can now get an iPhone at WalMart for $97. Emulating the fashion industry hasn't worked for technology industries.
True, but 'so what?'. The key problem with GPS location often isn't CPU power, but antenna design.
If antenna design were the limitation, phone GPS wouldn't work at all. All the GPS bits have to come in through the phone's antenna, after all.
"Assisted GPS" could be done by having cell sites give out their location, ephemeris data, and GPS propagation corrections to the handsets, instead of sending the phone's GPS data to the cell site. If the phone gets a rough location and time from the cell site, it then knows what almanac and ephemeris data to request (or cell sites could just broadcast the current, local ephemeris extract), and can cold-start its GPS algorithm quickly. Once basic lock has been achieved, the propagation corrections could be used to improve position, as with WAIS, only better.
(In standard GPS, the almanac and ephemeris data, which gives the orbital elements for all the satellites, is sent out with the GPS signal, but at a very slow data rate. A full download cycle requires 12.5 minutes of uninterrupted reception. But you don't need all that data if you have some idea of where you are. So a local node could broadcast the excepts you do need right now.)
The real reason "assisted GPS" is overcentralized is so that "Enhanced 911" and other police/surveillance functions (CALEA) will work.
You could do this with a <meta refresh> tag, have the server recognize the refererring page, and load the new, hostile page. This attack doesn't need JavaScript.
Some of the concurrency stuff needs a complete rewrite - acquiring synchronization primitives is painful, the new 'amazingly fast' locking that they use for GCD is marginally better than a FreeBSD mutex, and between one and three orders of magnitude (depending on load) faster than a Darwin mutex. Part of this is a userspace problem (not optimising for the uncontended case, which is the most common in good code), but a lot of it comes from the route down through the myriad kernel layers when sleeping a thread.
That problem in Mach is part of what gave microkernels a bad name. QNX, which is a real microkernel (about 65K of code) does thread dispatching, locking, and message passing very fast, in constant time, and without long interrupt lockouts. Those are the functions which must go fast in a microkernel, because they're used so much. In QNX, locking a mutex in the uncontested case is about three instructions in-line, with no system call. Those three functions are most of what the QNX kernel really does. In Mach, they were an afterthought, written on top of BSD.
This really belongs in the "when is it time to rewrite" thread.
Slashdot Mirror
The Wall Street Journal writes of "Battling the Cyber Warmongers". The big event in this area seems to have been the Cyber Shock Wave TV special, in which various former senior Federal officials participated in a simulation of a "cyber attack" on the power and phone system. Big names were involved: Michael Chertoff (former head of Homeland Security) and John Negroponte (former acting CIA director).
The show was embarrassing. It was clear that the participants not only had no clue what to do, they didn't know who to call who did. The person representing Energy kept harping on the thousands of energy companies there are in the US, and how they needed more authority over them. The level of cluelessness makes it clear how Hurricane Katrina (for which they had three days warning, and which happened on the watch of many of the participants) was botched.
In reality, the US power grid is divided operationally into seven regional grids, each with a control center and a backup control center. The supervisors at those control centers are the ones who really run things. If someone in the U.S. Government is dealing with an attack on the power grid, they need to have those supervisors on speed dial. In an emergency, the best thing representatives of the Government can do is call up each one and ask "What do you need right now". They're likely to get an answer like "We need troops protecting these key substations", or "We need a heavy-lift helicopter to move a spare transformer." Actions that would help fix the problem. One of the listed duties of the shift supervisor at the PJM Interconnect is to talk to Government officials.
The Government officials in that simulated emergency didn't have that basic info. They didn't know that there were seven people who were really running things. But they wanted to be in charge. That's the problem.
Few high government officials have a background as first responders or in incident management. If anything, the military officers are more likely to have a clue; their training teaches them to prioritize in a crisis and to deal with confusing, conflicting information.
The ACLU has taken at least two cases in that area.
The Maryland motorcycle case: "This prosecution by the Maryland State Police and Harford County State's Attorney is profoundly dangerous, and seems meant to intimidate people from making a record of what public officials do," said David Rocah, Staff Attorney at the ACLU of Maryland. "It is hard to imagine anything more antithetical to a democracy than for the government to tell its citizens that they do not have the right to record what government officials say or do or how they behave."
The video is on YouTube.
I'm on record as having called the dot-com bubble, the oil spike, and the mortgage crisis. It's not hard to predict bubbles. If you look at historical ratios, it's usually clear when assets are overpriced. Historically, the median house in the US sells for 2x to 2.5x the median income. That's about what people can pay for. That ratio hit 4 nationally, and 10 in some states. It was blindingly obvious that there was a housing bubble.
Dot-com predictions were easy. I had a program which read SEC filings for cash and burn rate, and simply projected when the money would run out. This was far more successful than one would expect. I used to get hate mail from CFOs for that.
The problem is figuring out when a bubble will pop. I expected the mortgage bubble to pop about two years earlier than it did. (Arguably, the Fed's cheap-money policy under Greenspan postponed the inevitable.)
Predictions on the debt side are harder than on the equity side. Public policy dominates the debt world. I don't make political predictions, so I can't say much about the current situation. I've been expecting an interest rate spike for years, but instead we've had a Federal deficit spike as money is pumped into the system. Eventually something will give there, as with Iceland, Greece, and other debtor countries. I'm not sure how that will unwind. We may get an interest rate spike and hyperinflation, which is what usually happens when a currency gets into trouble.
Flash itself is really very clever. The player packs an incredible amount of functionality into a very tiny executable. It's only 1.83MB. There's an animation engine, a JIT compiler, a video player, an audio system, and a multichannel download manager.
The problem is what people use it for. Which is mostly either ads or lame web sites.
Nobody really bothers doing Flash animations as entertainment much. If you've never seen one, check out Thugs on Film. Flash games remain popular, although Shockwave, which has full 3D capability, is a far better game platform. Many console games use Flash for 2D interface elements, typically using Adobe's authoring tools but a non-Adobe player. (Yes, there are non-Adobe Flash players.)
But it's not Adobe's fault that the content sucks.
It's sad. C++ has major problems, which are well known. The fundamental problem is that C++ has hiding without memory safety. C++ is unique among major languages in that respect. There are languages with no hiding, (C, assembler) and languages with both hiding and memory safety (Java, Delphi, Go, D, Ada, Modula, Erlang, Eiffel, and all the "scripting" languages).
The C++ committee, and Strostrup, are in denial about this. They're off in template la-la land, trying to abuse the template system into a compile-time interpreter.
C has its own problems. The basic one is that you lie to the language about how big things are. A string parameter is declared as "char *", not "char[n]". So the language has no idea how big objects are. It doesn't require garbage collection or array descriptors to fix this. It just requires a declaration syntax where you can talk about the size of the arrays you're passing around. This fundamental defect in C is the cause of most buffer overflows. (Technical discussion: C should have had array parameter declarations like int read(int fd, char[n]& buf, size_t n); rather than int read(int fd, char* buf, size_t n); Then the language knows how big the array is. This could be extended to structs, etc. so that whenever you have an array, there's an expression known to the compiler that defines its size. This allows subscript checking.)
C++ tries to paper this over with collection classes, but the mold always seeps through the wallpaper. Too often, you need a raw pointer, which breaks what little protection the collection classes provide.
A long time ago, when Project Gutenberg texts were really the only "ebooks" one could find, I had the idea of creating a separate data file that would accompany the .txt files. My idea was to leave the actual content of the book in plaintext for maximum portability, but allow fancy formatting (pagination, font, links, etc) via a separate binary file which would reference the .txt by character position.
"Bravo" for the Xerox Alto, the first multi-font WYSIWYG editor, worked that way. The text was stored as ASCII and terminated with a control-Z. Following the control-Z was the formatting information. Text-only utilities, like compilers, could read the files as plaintext. Late 1970s technology.
You can do anything you set your mind to when you have vision, determination, and an endless supply of expendable labor.
It's a real issue if a map system leads people into a trap.
There was a problem in England where a map system led truck drivers down a narrow dead-end lane from which they could not turn around. This was an ongoing problem, and the local property owners raised hell.
An Oregon couple was sent down a Forest Service road in a remote area and was stuck for three days before they were rescued.
The problem is serious in situations where the map system leads people onto an initially plausible route which gets them into trouble. Looking at the pictures for this walking problem, there's a dirt trail alongside the road. But does that dirt trail end before the destination?
Because Google actually drives out roads and takes pictures, people may be led to have more confidence in Google's mapping than in systems which just use ordinary map data.
"Cloud" companies are really just hosting providers. Hosting providers have their ups and downs; quite often they start out OK and get worse. You need to have a second source and a migration strategy.
In 12 years I've been through four hosting companies for one site. The first one started small, was acquired, was acquired again, and was eventually spun off to Earthlink. The second one was a good hosting company until they got into "permission-based email marketing" (i.e. spamming) and went downhill from there. The third one offered both dedicated servers and shared servers, then spun off the dedicated server business, leaving the shared business in bad shape. The fourth one is doing reasonably well right now.
I have a few things that use "cloud" type services. One uses a search API, and I have both Yahoo and Google versions. Another uses an SMS gateway, and I have both Google Voice and Twilio versions.
I own all my own domains, and the domain registrars are in no way affiliated with the hosting providers. For the important domains, I have registered U.S. trademarks. I've had to switch registrars on one occasion.
All servers are Linux, and all necessary tools are open source.
You have to assume that your suppliers can fail. Stay in this business for a few years, and some of them will fail.
The Tesla Roadster has a 245 mile range. And basic stuff like bumpers.
The student car looks like it has about a 3 inch ground clearance. If that. That's not going to get very far on anything less than a perfect road. And they want to drive it down from Prudhoe Bay? Right.
It's amusing that one of the linked articles mentions an "iPad killer" from Foxconn. Foxconn makes the iPad.
Foxconn's 2008 revenue was $62 billion. They're the "largest exporter in Greater China" and the world's largest maker of phone handsets. They have 486,000 employees. (Apple: 35,000. General Motors: 245,000.)
To see what it was like to use these fonts on the web, I created a test page. This uses dynamically downloaded fonts, which work in most current web browsers. (Firefox users need Firefox 3.6 or later.)
This sample is sized at 16 point. Smaller than that, many of the symbols are unreadable. That's something to be careful about. When you have a huge symbol set, the symbols need to be bigger. However, some of the symbols don't scale up well. If you scale up that page, the integral symbols look great, but the arrows become pixilated. Some of the symbols seem to have been were badly encoded.
This is just a raw demonstration of the font; for formulas, you'd use MathML. I'm not sure if MathML, the W3C names for math characters in HTML, and the STIX fonts are all synchronized yet. But at least you don't have to tell people "to display this page, install all these fonts first."
"Rather than sell electricity or water, though, operators will be taking data to measure impact on sea life, the generator's performance, and the cost of operation, said Douglas Sandberg, the vice president of the privately funded company."
So it's just a demo. Only generates 60KW. Not clear if that's average or max power; probably max. On days with low surf, not much will happen.
They've been hyping this since 2004. There are better wave powered generation devices, and even the best ones are commercial flops.
What idiots. They're shipping an "application" that consists of over 4,000 image files and some XML. You have to download 500MB of stuff to read the magazine. How long is that going to take?
Then there's the content problem. Wired, at this point, is basically a product catalog. Yet they didn't put a shopping cart system in the online version. That's just dumb. The demographic that buys both iPads and Wired would definitely click on "buy now" buttons.
Law enforcement works best when it's complaint driven. When cops are responding to crimes, they're performing a service function. The quality of that service can be measured. Crime rates, clearance rates, arrest rates, conviction rates, and recidivism rates can all be tallied, and are are politically important.
When law enforcement sets its own agenda, it does less well. It's harder to measure whether a vice squad or a drug squad is doing a good job. Those are the units that tend to be corrupt and/or incompetent. It's also much harder to measure their cost-effectiveness.
This problem shows up in the rhetoric. We have the War on Drugs, the War on Terror, and the War on Kiddie Porn. We don't have the War on Financial Fraud, the War on Child Labor, or the War on Bribery. Which is a problem.
If we were serious about Protecting the Children, child labor laws would be much tougher.
I thought the unpublished speed of the SR-71 was around Mach 6?
No, top speed of the SR-71 was around Mach 3.2, which it could sustain for hours. The absolute top speed ever achieved was 2,193.2 mph. The Museum of Flight at Boeing Field has had a SR-71 on display for years, and they have all the documentation.
The SR-71 powerplant is a turbojet that converts to a ramjet at high speeds.
When Kelly Johnson was designing the SR-71, they considered a liquid hydrogen propulsion system that might have made it to Mach 6 and beyond. Work on liquid hydrogen engines was conducted at Burbank. (That helped out NASA later on.) But they figured out how to do the necessary job with far less volatile fuels.
Now it's plugged with mud. With the flow much reduced, concrete can be put in.
One hurricane season and the mess will be gone. 8 to 14 hurricanes are expected in the Atlantic region by the end of the year.
Relief wells will be drilled; after all, there's definitely oil down there. The reservoir will be pumped out.
Everybody will be a lot more serious about blowout preventers.
More equipment for dealing with such problems will be on standby in some Gulf port for decades to come.
No big deal.
The original article is almost devoid of facts. What training is required to speak for "Advice Line"?
It's not at all clear what training is required for IT today? The Cisco "Rack Test"? How to fix broken Windows systems? J2EE programming? Linux server administration?
CS is even tougher. Robotics? AI? Machine learning? Graphics? Digital logic? "Cloud" programming? There are too many narrow niches. Pick the wrong one and you're toast.
Here's the bottom of the apparel food chain. "Bulk Bale Clothing". $0.25/lb. Minimum order 55,000 pounds. Supply availability 1,000,000 pounds per month. Bulk in 1000 pound bales.
That's just one of a hundred similar suppliers. "We currently have 28 containers of brand name clothing acquired in a bank deal." "250,000 lbs baled used clothing. 25% coats,sweaters, heavy clothing. $0.84/lb."
That's life in the no-IP world of apparel. The wastage is enormous.
Remember a company who dragged their entire OS to Trash, emptied it and restarted with a fresh and open source OS instead of trying to "fix" it?
Apple did that to bail out Jobs, when NeXt was in the tank. Apple had developed a new OS, MacOS 8 ("Copeland"), which was a reasonably good rewrite. The claim was that a warmed-over NextStep could be on the market sooner than Copeland. It wasn't, but the deal saved Jobs' personal wealth.
Fashion depends on churn. There aren't many original ideas. If you look at this year's runway fashion, and are familiar with the history of fashion, you can usually find a matching piece from decades ago. The Fashion Institute of Technology in New York has a large library of clothing against which new designs can be compared. This year, we have khaki (again), Gautier is trying green spandex shorts (80s aerobic wear), and Issey Miyake is over-pleating again (he did that better in the '80s) and using pink accents (so last year.) The jeans industry keeps fussing with various levels of fading, but they've been doing that for so long that nobody is paying much attention.
There's some technological progress, and it gets IP protection. Gore-Tex was patented, and for a long time, had a monopoly over waterproof fabrics that breathe well. Progress in materials, in sewing technology, and in cleaning has led to new ranges of clothing. Jeans, for example, depend on a sewing technology for strong corners that's only about fifty years old. (Today, rivets in jeans are decorative, not structural.) Sportswear, which was invented by Coco Chanel, wasn't really feasible before washing machines. Elastic fabrics opened up many new options. Not much new has come along in the last two decades, ("pleather" made a small splash) and fashion technology has somewhat stagnated.
As the TED talk points out, the big thing today is trademark protection via "designer labels". This is a relatively new concept. Until the late 1970s, no respectable garment maker would have the designer's mark visible, let alone a prominent feature of the design. Logos were associated with cheap T-shirts. The interlocking double C now associated with Chanel did not appear on Chanel products until the 1980s.
The apparel distribution pipeline is incredibly inefficient. Over 60% of apparel is eventually sold on sale. There's a hierarchy. First there's the initial sale, with a big markup. Then there's the sale rack at the original retailer, with the original tags still attached. Then there's the discount retailer who buys from the original retailer and resells from their own store. Finally, the unsold apparel is rolled into big balls about eight feet in diameter, which are rolled into shipping containers and shipped to third world countries for final sale, or recycled into nonwoven fabrics like cleaning cloths. There was an attempt during the dot-com era, called "Tradeweave", to create a secondary trading market in unsold apparel, but it only lasted from 1999-2001.
The watch industry is a branch of the fashion industry now. "We are not in the watch industry, we are are in the luxury industry" says the CEO of Rolex. They had the basic problem that their overpriced machinery is less accurate than a midrange quartz watch, and they now have the worse problem that people who grew up with cell phones see no need for watches. There have been attempts by the phone industry to do "designer cell phones", but so far, that's mostly a joke. Apple tried to position themselves as a high-end product, but you can now get an iPhone at WalMart for $97. Emulating the fashion industry hasn't worked for technology industries.
there's no assurance that the target isn't viewing the tab when it refreshes.
True, but phishing, like spamming, is a high-volume low-success-rate scam. It doesn't have to work all the time.
True, but 'so what?'. The key problem with GPS location often isn't CPU power, but antenna design.
If antenna design were the limitation, phone GPS wouldn't work at all. All the GPS bits have to come in through the phone's antenna, after all.
"Assisted GPS" could be done by having cell sites give out their location, ephemeris data, and GPS propagation corrections to the handsets, instead of sending the phone's GPS data to the cell site. If the phone gets a rough location and time from the cell site, it then knows what almanac and ephemeris data to request (or cell sites could just broadcast the current, local ephemeris extract), and can cold-start its GPS algorithm quickly. Once basic lock has been achieved, the propagation corrections could be used to improve position, as with WAIS, only better.
(In standard GPS, the almanac and ephemeris data, which gives the orbital elements for all the satellites, is sent out with the GPS signal, but at a very slow data rate. A full download cycle requires 12.5 minutes of uninterrupted reception. But you don't need all that data if you have some idea of where you are. So a local node could broadcast the excepts you do need right now.)
The real reason "assisted GPS" is overcentralized is so that "Enhanced 911" and other police/surveillance functions (CALEA) will work.
This attack only works if you allow Javascript...
You could do this with a <meta refresh> tag, have the server recognize the refererring page, and load the new, hostile page. This attack doesn't need JavaScript.
Some of the concurrency stuff needs a complete rewrite - acquiring synchronization primitives is painful, the new 'amazingly fast' locking that they use for GCD is marginally better than a FreeBSD mutex, and between one and three orders of magnitude (depending on load) faster than a Darwin mutex. Part of this is a userspace problem (not optimising for the uncontended case, which is the most common in good code), but a lot of it comes from the route down through the myriad kernel layers when sleeping a thread.
That problem in Mach is part of what gave microkernels a bad name. QNX, which is a real microkernel (about 65K of code) does thread dispatching, locking, and message passing very fast, in constant time, and without long interrupt lockouts. Those are the functions which must go fast in a microkernel, because they're used so much. In QNX, locking a mutex in the uncontested case is about three instructions in-line, with no system call. Those three functions are most of what the QNX kernel really does. In Mach, they were an afterthought, written on top of BSD.
This really belongs in the "when is it time to rewrite" thread.