One of the assumptions of the paper is that some sources, like wind and tidal power, can be expanded as desired. That's not the case.
Tidal power is very limited; the geography has to be just right. There are about four really good sites in the world, the Bay of Fundy being the best. Unfortunately, it's nowhere near a big electrical load.
Wind power sites are more limited than most people realize. Look at the wind maps of the United States. The high wind areas are mostly far from the populated areas. The best wind areas are from the Texas panhandle north to Canada, the big empty space in the US. Illinois looks very promising.The Northeast and South, not much. California has four really good on-shore wind sites, and all four already have wind farms.
Wind power is also more variable than its enthusiasts realize. Check out the current California wind output graph. There's been a 3:1 variation in output just today, and that's with wind farms spread over an area 400 miles across.
Hydroelectric power is great, but all the good dam sites were gone by 1940. The ideal dam is Hoover Dam - plug up a gorge in a useless desert and fill the desert with water. Few other sites are that good.
Hydrogen is, of course, a joke. It's a terrible way to store electric power. Inefficient to make, and dangerous to handle. Electric cars get that job done just fine with batteries.
Ethanol from cellulose looks promising. That works now, although it's still kind of expensive. It runs on agricultural waste and other unwanted cellulose, so it's a good renewable source. That's probably the liquid fuel of the future. Ethanol from food crops is a tax gimmick.
Solar power is very effective in the right climate. Realistically, that means the southern half of the US and points south, Spain, most of India, Africa, etc. There's also the nice property that peak electrical load in places that need air conditioning is guaranteed to coincide with peak solar power output.
The world will end up running mostly on renewable energy, though. The fossil fuels are running out. Oil at $100/bbl is the new normal. It's not going to be pleasant.
If it's doing searches in bulk like that, it's a search spam program. It's exploiting a vulnerability in Google.
Google Trends lists "hot searches", what's being searched for in Google in recent hours. Google Trends drives Google Suggest, the hinting system for Google. That in turn drives Google Instant. Which, in turn, aims users at the target sites. Which are probably full of ads. Profit!
Spamming of Google Trends has been around for a while. It used to be easier, and you'd see things like the name of some mattress discounter at the top of Google Trends for 15 minutes or so. (I ran a program to follow the trends in Google Trends for a while. It was amusing.)
Google seems to now be averaging over more hours, so the spammers have to up their game and use a distributed attack to push their keywords up.
This is the trouble with "crowdsourcing" recommendations. It's too easy to fake a crowd. Yelp, CitySearch, Google Places - they're all choked with recommendation spam. Anonymous recommendations are junk information. And no, requiring a Facebook account won't help. There's an app for that.
Google is now trying a "mark as spam" button in Chrome to identify "content farms". If that starts mattering, it will be spammed. The same applies to Blekko's "slashtags".
The market for R-rated titillation has been reduced. If you want porn, you can get all the porn you want. Of course, as Ozzy Osborne complains, in porn, there are no surprise endings.
Today, if there's sex, it needs to be fully integrated into the plot. "Dollhouse" had some of that, but their better episodes were more like action-adventure bits. "Aeon Flux" tried, but despite a nicely put together production and a very hot female lead, didn't quite succeed. "Serenity" had the Companions, but in a supporting role.
Suppose the Companions had a political agenda, and they were trying to influence events through sex. Now that has potential.
Apple's #2 manager is Timothy Cook, Apple's chief operating officer. He's a cost control and outsourcing guy. He came from Compaq, where he turned them from a manufacturer into a distributor of products made offshore. That's what he did at Apple, too - closed all the Apple factories and outsourced manufacturing to FoxConn. He's good at managing low-cost outsourced manufacturing. He's running Apple now, and will probably succeed Jobs.
Apple will survive. Cook will hire some low-cost design firm in Beijing to do the next products. There's good design coming out of Beijing now. Check out PER design group's work. FoxConn in Shenzen already does the manufacturing. All Apple US has to do is manage the deals with other parties. Cupertino will probably be downsized to a marketing and IP rights organization.
My big hope for the future of movies is that technology is making movies less expensive to produce.
About a decade ago, when I was working on animation software, I had that discussion with a Hollywood director. That's what he wanted - the technology to make an A picture for $10 million or so. He'd just made a $100 million picture with both live action and CGI characters, and, because of the high costs, it had to be pre-planned to death and micromanaged by the front office. He hated that. If the costs came down, the director would be in charge again.
He was impressed by ReBoot - the first whole season CGI cartoon. Not that it was good, but that about 15 to 20 people were cranking out an episode a week. That level of productivity made more interesting projects affordable.
Ten years later, we're still not there. People have tried. "Sky Captain" was originally supposed to be a low-budget film with all CGI sets. If no actor touches something in "Sky Captain", it's CGI. The first few minutes were originally made on a Mac by one guy. (You can find that version on the DVD.) But it didn't look good enough. The budget ballooned to $70 million. The number of set carpenters is way down, but there are buildings full of animators carefully placing CGI props instead.
What we're getting today at the low end is green screen inserts for background details. That's cheap. (Note how many TV scenes seem to take place inside moving cars.) But there's no magic world-building bullet yet.
That will probably come, and from the game industry, which needs to do massive world-building. A city generator which could generate a whole city in fine detail, down to apartment and shop interiors for all buildings, without replicating the same items, is something the game industry needs, and understands they need.
The article just compares the feature lists. It's not clear if either is better from a bug standpoint. A big problem with OpenOffice is that it tends to crash too much. (Especially, for some reason, when exiting.) Also, OpenOffice had some features written in Java, but they were optional. Did LibreOffice get rid of the Oracle Java parts, replace them with something, or what?
It's encouraging that LibreOffice is around. I've been using OpenOffice since 1.0, and haven't used a version of Microsoft Word later than Word 97. OpenOffice in its later incarnations isn't bad, although it still, after ten years, has an amateurish feel to it.
The US military has had practice towns for years. They use the term "MOUT" (Military Operations in Urban Terrain) sites. There's even a VRML file for the one at Fort Benning. There's a bigger one at Fort Irwin, and most major infantry bases have at least a modest mockup town.
The FBI has an elaborate one at Quantico, VA. The Secret Service, though, doesn't seem to have one at their training center. They do, though, have a really big skid pad, for driving practice.
It's not clear from the article whether "Tiny Town" is a planning aid or a training aid. That is, do they match real-world areas where they plan to operate, or just use it for training exercises?
What does work is Bing's approach of using actual user data.
No, that's spammable, too. See "click fraud". Anonymous crowdsourcing in competitive environments only works if you're a little player and nobody cares enough to spam you. If Blekko gets enough market share to attract SEO efforts, their "slashtags" will be overwhelmed by junk.
Read how Craigslist lost the battle against spam. They tried CAPTCHAs. They tried requiring unique email accounts. They tried phone verification. Nothing worked. There are power tools for defeating each of those. Most of the recommendation systems have similar problems. To check this out, read Citysearch recommendations for some category like carpet cleaning or locksmiths, cut out some unique phrase from a recommendation, and search for it to see in how many other recommendations it appears.
The only recommendation systems that really work are ones where either the number of recommendations per item is huge (as with movies and TV), or recommendations are tied to transactions (as with eBay or Amazon.)
Note that the person saying that the women want this is a man. Typical.
For many young Saudi women, an education in the US is their one time of freedom in life. Some years ago, I was chatting with a Saudi woman about to finish Stanford, and she mentioned that she was going to drive across the United States, then fly back to Saudi Arabia from the East Coast. I asked her why the long drive, and she said it was the only time in her life she'd be allowed to do something like that. (In Saudi Arabia, women aren't allowed to drive.)
I like the red-state solution to this problem. Someone at a Texas company wrote that an Islamic female co-worker was being harassed by an Islamic male employee who just assumed that, since he was male. he had the right to her if he wanted. So the Texan took the woman to a shooting range and taught her how to use a 9mm pistol. "You taught her how to shoot?" the annoying guy said when he found out about this. "Yes, and she's good at it, too". No more problems.
There was once a real-time OS company that gave you a Bug, a Volkswagen Beetle, if you found a bug in their OS. They gave away about two cars a year, and it was worth it.
This is one of those press events which gives CEOs nightmares.
There's been press criticism of Google before, but it's mostly been anecdotal - blogs, op-eds, and other commentary. This time, there's real reporting, with the New York Times naming names.
Usually, after criticism, Google says nothing, or perhaps replies in a blog posting. Google people rarely speak in forums that they don't control. This time, Matt Cutts had to sit down with New York Times reporters for an hour long interrogation.
Google's vaunted claims that they can detect link spam were shown to be false. Google didn't catch the spam, the New York Times did. Then Google made an algorithm change and claimed that fixed most of the problem. The Times tracked Google's results and showed that it didn't. Only a "manual action" moved J.C. Penny down.
Now the rest of the business press is going to take a hard look at search. Expect follow-up articles in Bloomberg, Fortune, etc. Google management has weeks of pain ahead. After their feud with Microsoft last week, their troubles with European antitrust regulators, and Blekko nipping at their heels, they didn't need this.
Attention may be focused on those "manual actions". Should those be published? The European Union has specifically asked for that data, and Google can no longer deny that it exists.
I've been critical of Google's anti-spam efforts, mostly on the Places side.I thought they were better at detecting link farms of junk sites, though. That's old-school SEO. If they missed this, they have worse problems than I'd thought.
The people I want to hear most from are not the IP intellectual discussions about moot points of IP policy, but from actual patent holders who have innovative technologies that have been blocked from innovating by the patent system.
I hold four issued patents: First raster to vector conversion on PCs, first anti-slip control for rough terrain legged robots, first ragdoll physics that didn't go unstable, and first anti-web-spam system that eliminates most of the low-end junk.
I've only run into a patent as an obstacle once. I was working on a program to parse financial statements automatically, and ran into a Price Waterhouse patent on how to resolve ambiguities about what rows combine to totals in other rows. (Published financial statements are often ambiguous in this regard.) So I gave up on that project; I wasn't the first to have the idea.
There are legitimate problems with patents on de-facto standards established by a party with monopoly power, but that's properly an antitrust issue. Examples are the Hayes modem escape sequence patent, the Microsoft long filename in FAT file systems patent, and the MPEG-LA patent pool. For all of those, the patents are narrow, and only valuable because there's a de-facto standard on a specific way of doing something. It's easy to invent around any of those narrow patents, but then there's a compatibility problem. If antitrust enforcement in the US hadn't been gutted, existing law could deal with that.
We need a backup plan for a "slow exchange". Something where the price of each stock changes every hour, after an hourly auction. The classic London Gold Fixing and the PJM electricity market work something like that. Such a market wouldn't need that much of an infrastructure, and it would be just enough to maintain liquidity.
This would be hard on exchange-traded funds, so we'd need a rule that, in emergencies, ETFs would be priced daily, like regular funds, and you can redeem them only on a daily basis.
The rules and infrastructure for this should be put in place, with dedicated, hardened computers in several locations idling on standby.
One of the best ones was the DEC Extended Static Checker for Java in the late 1990s. That project died after Compaq acquired DEC and shut down research. But it formed part of Microsoft's efforts for Spec#, which has a formal verification system.
Microsoft has had a huge success with formal proof of correctness - the Static Driver Verifier. This is a system which tries to formally prove that a Windows driver can't crash the operating system. All drivers for Windows 7 have to pass this verifier to be signed, It's an impressive system which works by symbolic case analysis. It yields a proof, a counterexample, or times out analyzing too many cases. About 3% of programs time out. (If your kernel driver has undecidable behavior, it needs work.)
I headed a team that did a proof of correctness system for Pascal back in the early 1980s. That was pushing the limits of what you could do back then; verifying a small program took 45 minutes on a 1 MIPS VAX. Today, with about four orders of magnitude more CPU power, it's a practical technology.
There have been many attempts to bolt some "design by contract" features onto various languages, but they're usually not very good. The problem is saying things about collections. The hacks that try to do design by contract at run time tend to avoid expressions which examine big data structures, since they have to run them over the whole data structure every time. Real verifiers prove or disprove such things at compile time. It turns out, though, that a few standard cases for collections cover many of the usual things you want to say.
Another big issue, which Spec# handles and this Google hack apparently doesn't, is the "insideness" issue for object invariants. Object invariants are supposed to be true whenever control enters and exits an object. But what if an object calls one of its own public methods, perhaps through a long chain of calls? Now, control is inside the same object twice. That violates the invariant rule.
This class of bugs is common in window systems, where all the widget, pane, window, and event objects are frantically calling each other. It also comes up in concurrency, which I don't see Google's hack addressing either.
Because OPEC members production is capped at a ratio of their reserves. So the higher they claim their reserves are the more oil they are allowed to sell.
Right. Mod parent up.
The claimed reserved of most of the OPEC countries went up when OPEC put in that rule in the 1980s. That didn't reflect any new discoveries. Saudi reserves have been widely believed to be exaggerated, but no one is clear on how much.
There's another side to that - intelligence. Almost nobody expected the uprising in Egypt two weeks in advance, not even its current leadership. Few saw the downfall of the Soviet Union coming. More insight into noticing when something big like that is coming is useful.
This is just an OS reinstall. It's not like going in through the JTAG port and loading the firmware.
These devices exist in the space between hard embedded systems and OSs with a user interface. With hard embedded systems, you do all development on another machine and download an entire image. With user-oriented OSs, you can interact with the machine in some reasonable way. These things live in a limbo between those two points - smart enough not to be total slaves, but not smart enough for standalone development.
There's nothing wrong with that; that's normal life in the embedded world. But it's not something that end users previously had to deal with.
New TLDs are mostly a racket for registrars, who try to get companies to purchase the same domain in multiple TLD's for "protection". This is close to an extortion racket. Since ICANN got rid of those annoying "public" board members, it's just been a trade group for registrars.
Most of the newer TLDs are duds, anyway. ".museum" has so few domains that the whole list is a few pages. ".aero" has an entry for each airport code, but those are mostly redirects put up by the registrar. ".biz" is a strip mall in a bad neighborhood with grates over the windows and graffiti on the walls. ".name", for individuals, never caught on. Nor did ".travel", ".pro", or ".jobs". ".tel" is a non-Web service for telephony, and has a legit use. The rest could probably be closed to new registrations and phased out as names drop.
Won't work. The Shuttle required a huge infrastructure, employing about 15,000 people as late as 2009. Layoffs have been underway for years. Manufacturing and repair facilities have been closed down. The parts stock has been depleted. It's over.
Buran is not a clone of the US space shuttle. It looks similar, but it's not. Buran is launched on a huge booster; it has no main engines of its own, unlike the US Shuttle. It's more rugged than the Shuttle, which can't handle rain or cold weather. In many ways, it's a better design. T
I don't want four browser versions in a year. I want one that works right. I'm tired of their browser hanging, taking minutes of compute time to exit, and unable to close some obnoxious hostile pages.
When the Mozilla crowd gets a bug-free 4.x out the door, with all reported bugs fixed, then they can talk about later versions.
(I'm underwhelmed with the Mozilla crowd. They come across as a bunch of Javascript hackers. There was a problem with duplicate entries in the internal SQLite databases, and I suggested making the fields a unique key, so the database would reject bogus updates. But that would have caused broken code to abort. Instead, they bolted on a Javascript hack to try to clean up the mess.)
Linux still has the antiquated "user, group, everyone" security model from the 1970s. By now, we know that outside data can't be given all the privileges of the user. But Linux's legacy security model is so deeply embedded in the UNIX/Linux world that it's almost impossible to get beyond that.
Yes, there's SELinux. But there isn't a whole distribution with a full range of applications which can run under a mandatory security model.
Slashdot Mirror
The paper is disappointing.
One of the assumptions of the paper is that some sources, like wind and tidal power, can be expanded as desired. That's not the case.
Tidal power is very limited; the geography has to be just right. There are about four really good sites in the world, the Bay of Fundy being the best. Unfortunately, it's nowhere near a big electrical load.
Wind power sites are more limited than most people realize. Look at the wind maps of the United States. The high wind areas are mostly far from the populated areas. The best wind areas are from the Texas panhandle north to Canada, the big empty space in the US. Illinois looks very promising.The Northeast and South, not much. California has four really good on-shore wind sites, and all four already have wind farms.
Wind power is also more variable than its enthusiasts realize. Check out the current California wind output graph. There's been a 3:1 variation in output just today, and that's with wind farms spread over an area 400 miles across.
Hydroelectric power is great, but all the good dam sites were gone by 1940. The ideal dam is Hoover Dam - plug up a gorge in a useless desert and fill the desert with water. Few other sites are that good.
Hydrogen is, of course, a joke. It's a terrible way to store electric power. Inefficient to make, and dangerous to handle. Electric cars get that job done just fine with batteries.
Ethanol from cellulose looks promising. That works now, although it's still kind of expensive. It runs on agricultural waste and other unwanted cellulose, so it's a good renewable source. That's probably the liquid fuel of the future. Ethanol from food crops is a tax gimmick.
Solar power is very effective in the right climate. Realistically, that means the southern half of the US and points south, Spain, most of India, Africa, etc. There's also the nice property that peak electrical load in places that need air conditioning is guaranteed to coincide with peak solar power output.
The world will end up running mostly on renewable energy, though. The fossil fuels are running out. Oil at $100/bbl is the new normal. It's not going to be pleasant.
At least they saw it hitting Baidu. But it gets its target list from the botmaster's server. What it hits may change depending on current orders.
If it's doing searches in bulk like that, it's a search spam program. It's exploiting a vulnerability in Google.
Google Trends lists "hot searches", what's being searched for in Google in recent hours. Google Trends drives Google Suggest, the hinting system for Google. That in turn drives Google Instant. Which, in turn, aims users at the target sites. Which are probably full of ads. Profit!
Spamming of Google Trends has been around for a while. It used to be easier, and you'd see things like the name of some mattress discounter at the top of Google Trends for 15 minutes or so. (I ran a program to follow the trends in Google Trends for a while. It was amusing.) Google seems to now be averaging over more hours, so the spammers have to up their game and use a distributed attack to push their keywords up.
This is the trouble with "crowdsourcing" recommendations. It's too easy to fake a crowd. Yelp, CitySearch, Google Places - they're all choked with recommendation spam. Anonymous recommendations are junk information. And no, requiring a Facebook account won't help. There's an app for that.
Google is now trying a "mark as spam" button in Chrome to identify "content farms". If that starts mattering, it will be spammed. The same applies to Blekko's "slashtags".
The market for R-rated titillation has been reduced. If you want porn, you can get all the porn you want. Of course, as Ozzy Osborne complains, in porn, there are no surprise endings.
Today, if there's sex, it needs to be fully integrated into the plot. "Dollhouse" had some of that, but their better episodes were more like action-adventure bits. "Aeon Flux" tried, but despite a nicely put together production and a very hot female lead, didn't quite succeed. "Serenity" had the Companions, but in a supporting role.
Suppose the Companions had a political agenda, and they were trying to influence events through sex. Now that has potential.
Apple's #2 manager is Timothy Cook, Apple's chief operating officer. He's a cost control and outsourcing guy. He came from Compaq, where he turned them from a manufacturer into a distributor of products made offshore. That's what he did at Apple, too - closed all the Apple factories and outsourced manufacturing to FoxConn. He's good at managing low-cost outsourced manufacturing. He's running Apple now, and will probably succeed Jobs.
Apple will survive. Cook will hire some low-cost design firm in Beijing to do the next products. There's good design coming out of Beijing now. Check out PER design group's work. FoxConn in Shenzen already does the manufacturing. All Apple US has to do is manage the deals with other parties. Cupertino will probably be downsized to a marketing and IP rights organization.
My big hope for the future of movies is that technology is making movies less expensive to produce.
About a decade ago, when I was working on animation software, I had that discussion with a Hollywood director. That's what he wanted - the technology to make an A picture for $10 million or so. He'd just made a $100 million picture with both live action and CGI characters, and, because of the high costs, it had to be pre-planned to death and micromanaged by the front office. He hated that. If the costs came down, the director would be in charge again.
He was impressed by ReBoot - the first whole season CGI cartoon. Not that it was good, but that about 15 to 20 people were cranking out an episode a week. That level of productivity made more interesting projects affordable.
Ten years later, we're still not there. People have tried. "Sky Captain" was originally supposed to be a low-budget film with all CGI sets. If no actor touches something in "Sky Captain", it's CGI. The first few minutes were originally made on a Mac by one guy. (You can find that version on the DVD.) But it didn't look good enough. The budget ballooned to $70 million. The number of set carpenters is way down, but there are buildings full of animators carefully placing CGI props instead.
What we're getting today at the low end is green screen inserts for background details. That's cheap. (Note how many TV scenes seem to take place inside moving cars.) But there's no magic world-building bullet yet.
That will probably come, and from the game industry, which needs to do massive world-building. A city generator which could generate a whole city in fine detail, down to apartment and shop interiors for all buildings, without replicating the same items, is something the game industry needs, and understands they need.
(Read the print version of the article on one page. It's one of those "short article spread across many ad-heavy pages" crap sites.)
The article just compares the feature lists. It's not clear if either is better from a bug standpoint. A big problem with OpenOffice is that it tends to crash too much. (Especially, for some reason, when exiting.) Also, OpenOffice had some features written in Java, but they were optional. Did LibreOffice get rid of the Oracle Java parts, replace them with something, or what?
It's encouraging that LibreOffice is around. I've been using OpenOffice since 1.0, and haven't used a version of Microsoft Word later than Word 97. OpenOffice in its later incarnations isn't bad, although it still, after ten years, has an amateurish feel to it.
The US military has had practice towns for years. They use the term "MOUT" (Military Operations in Urban Terrain) sites. There's even a VRML file for the one at Fort Benning. There's a bigger one at Fort Irwin, and most major infantry bases have at least a modest mockup town.
The FBI has an elaborate one at Quantico, VA. The Secret Service, though, doesn't seem to have one at their training center. They do, though, have a really big skid pad, for driving practice.
It's not clear from the article whether "Tiny Town" is a planning aid or a training aid. That is, do they match real-world areas where they plan to operate, or just use it for training exercises?
What does work is Bing's approach of using actual user data.
No, that's spammable, too. See "click fraud". Anonymous crowdsourcing in competitive environments only works if you're a little player and nobody cares enough to spam you. If Blekko gets enough market share to attract SEO efforts, their "slashtags" will be overwhelmed by junk.
Read how Craigslist lost the battle against spam. They tried CAPTCHAs. They tried requiring unique email accounts. They tried phone verification. Nothing worked. There are power tools for defeating each of those. Most of the recommendation systems have similar problems. To check this out, read Citysearch recommendations for some category like carpet cleaning or locksmiths, cut out some unique phrase from a recommendation, and search for it to see in how many other recommendations it appears.
The only recommendation systems that really work are ones where either the number of recommendations per item is huge (as with movies and TV), or recommendations are tied to transactions (as with eBay or Amazon.)
Note that the person saying that the women want this is a man. Typical.
For many young Saudi women, an education in the US is their one time of freedom in life. Some years ago, I was chatting with a Saudi woman about to finish Stanford, and she mentioned that she was going to drive across the United States, then fly back to Saudi Arabia from the East Coast. I asked her why the long drive, and she said it was the only time in her life she'd be allowed to do something like that. (In Saudi Arabia, women aren't allowed to drive.)
I like the red-state solution to this problem. Someone at a Texas company wrote that an Islamic female co-worker was being harassed by an Islamic male employee who just assumed that, since he was male. he had the right to her if he wanted. So the Texan took the woman to a shooting range and taught her how to use a 9mm pistol. "You taught her how to shoot?" the annoying guy said when he found out about this. "Yes, and she's good at it, too". No more problems.
There was once a real-time OS company that gave you a Bug, a Volkswagen Beetle, if you found a bug in their OS. They gave away about two cars a year, and it was worth it.
This is one of those press events which gives CEOs nightmares.
There's been press criticism of Google before, but it's mostly been anecdotal - blogs, op-eds, and other commentary. This time, there's real reporting, with the New York Times naming names.
Usually, after criticism, Google says nothing, or perhaps replies in a blog posting. Google people rarely speak in forums that they don't control. This time, Matt Cutts had to sit down with New York Times reporters for an hour long interrogation.
Google's vaunted claims that they can detect link spam were shown to be false. Google didn't catch the spam, the New York Times did. Then Google made an algorithm change and claimed that fixed most of the problem. The Times tracked Google's results and showed that it didn't. Only a "manual action" moved J.C. Penny down.
Now the rest of the business press is going to take a hard look at search. Expect follow-up articles in Bloomberg, Fortune, etc. Google management has weeks of pain ahead. After their feud with Microsoft last week, their troubles with European antitrust regulators, and Blekko nipping at their heels, they didn't need this. Attention may be focused on those "manual actions". Should those be published? The European Union has specifically asked for that data, and Google can no longer deny that it exists.
I've been critical of Google's anti-spam efforts, mostly on the Places side.I thought they were better at detecting link farms of junk sites, though. That's old-school SEO. If they missed this, they have worse problems than I'd thought.
"Many indie New York City cafes now heavily restrict, or ban outright, the use of Kindles, Nooks and iPads."
That sentence should have been followed by the names of some cafes, and an interview with an owner. But no. So it's just blithering.
The Times is slipping.
The people I want to hear most from are not the IP intellectual discussions about moot points of IP policy, but from actual patent holders who have innovative technologies that have been blocked from innovating by the patent system.
I hold four issued patents: First raster to vector conversion on PCs, first anti-slip control for rough terrain legged robots, first ragdoll physics that didn't go unstable, and first anti-web-spam system that eliminates most of the low-end junk.
I've only run into a patent as an obstacle once. I was working on a program to parse financial statements automatically, and ran into a Price Waterhouse patent on how to resolve ambiguities about what rows combine to totals in other rows. (Published financial statements are often ambiguous in this regard.) So I gave up on that project; I wasn't the first to have the idea.
There are legitimate problems with patents on de-facto standards established by a party with monopoly power, but that's properly an antitrust issue. Examples are the Hayes modem escape sequence patent, the Microsoft long filename in FAT file systems patent, and the MPEG-LA patent pool. For all of those, the patents are narrow, and only valuable because there's a de-facto standard on a specific way of doing something. It's easy to invent around any of those narrow patents, but then there's a compatibility problem. If antitrust enforcement in the US hadn't been gutted, existing law could deal with that.
We need a backup plan for a "slow exchange". Something where the price of each stock changes every hour, after an hourly auction. The classic London Gold Fixing and the PJM electricity market work something like that. Such a market wouldn't need that much of an infrastructure, and it would be just enough to maintain liquidity.
This would be hard on exchange-traded funds, so we'd need a rule that, in emergencies, ETFs would be priced daily, like regular funds, and you can redeem them only on a daily basis.
The rules and infrastructure for this should be put in place, with dedicated, hardened computers in several locations idling on standby.
One of the best ones was the DEC Extended Static Checker for Java in the late 1990s. That project died after Compaq acquired DEC and shut down research. But it formed part of Microsoft's efforts for Spec#, which has a formal verification system.
Microsoft has had a huge success with formal proof of correctness - the Static Driver Verifier. This is a system which tries to formally prove that a Windows driver can't crash the operating system. All drivers for Windows 7 have to pass this verifier to be signed, It's an impressive system which works by symbolic case analysis. It yields a proof, a counterexample, or times out analyzing too many cases. About 3% of programs time out. (If your kernel driver has undecidable behavior, it needs work.)
I headed a team that did a proof of correctness system for Pascal back in the early 1980s. That was pushing the limits of what you could do back then; verifying a small program took 45 minutes on a 1 MIPS VAX. Today, with about four orders of magnitude more CPU power, it's a practical technology.
There have been many attempts to bolt some "design by contract" features onto various languages, but they're usually not very good. The problem is saying things about collections. The hacks that try to do design by contract at run time tend to avoid expressions which examine big data structures, since they have to run them over the whole data structure every time. Real verifiers prove or disprove such things at compile time. It turns out, though, that a few standard cases for collections cover many of the usual things you want to say.
Another big issue, which Spec# handles and this Google hack apparently doesn't, is the "insideness" issue for object invariants. Object invariants are supposed to be true whenever control enters and exits an object. But what if an object calls one of its own public methods, perhaps through a long chain of calls? Now, control is inside the same object twice. That violates the invariant rule.
This class of bugs is common in window systems, where all the widget, pane, window, and event objects are frantically calling each other. It also comes up in concurrency, which I don't see Google's hack addressing either.
Because OPEC members production is capped at a ratio of their reserves. So the higher they claim their reserves are the more oil they are allowed to sell.
Right. Mod parent up. The claimed reserved of most of the OPEC countries went up when OPEC put in that rule in the 1980s. That didn't reflect any new discoveries. Saudi reserves have been widely believed to be exaggerated, but no one is clear on how much.
World oil production peaked in 2005.
There's another side to that - intelligence. Almost nobody expected the uprising in Egypt two weeks in advance, not even its current leadership. Few saw the downfall of the Soviet Union coming. More insight into noticing when something big like that is coming is useful.
This is just an OS reinstall. It's not like going in through the JTAG port and loading the firmware.
These devices exist in the space between hard embedded systems and OSs with a user interface. With hard embedded systems, you do all development on another machine and download an entire image. With user-oriented OSs, you can interact with the machine in some reasonable way. These things live in a limbo between those two points - smart enough not to be total slaves, but not smart enough for standalone development.
There's nothing wrong with that; that's normal life in the embedded world. But it's not something that end users previously had to deal with.
Just monitor police cars and garbage trucks. They tend to cover most streets every few days.
New TLDs are mostly a racket for registrars, who try to get companies to purchase the same domain in multiple TLD's for "protection". This is close to an extortion racket. Since ICANN got rid of those annoying "public" board members, it's just been a trade group for registrars.
Most of the newer TLDs are duds, anyway. ".museum" has so few domains that the whole list is a few pages. ".aero" has an entry for each airport code, but those are mostly redirects put up by the registrar. ".biz" is a strip mall in a bad neighborhood with grates over the windows and graffiti on the walls. ".name", for individuals, never caught on. Nor did ".travel", ".pro", or ".jobs". ".tel" is a non-Web service for telephony, and has a legit use. The rest could probably be closed to new registrations and phased out as names drop.
Won't work. The Shuttle required a huge infrastructure, employing about 15,000 people as late as 2009. Layoffs have been underway for years. Manufacturing and repair facilities have been closed down. The parts stock has been depleted. It's over.
Buran is not a clone of the US space shuttle. It looks similar, but it's not. Buran is launched on a huge booster; it has no main engines of its own, unlike the US Shuttle. It's more rugged than the Shuttle, which can't handle rain or cold weather. In many ways, it's a better design. T
I don't want four browser versions in a year. I want one that works right. I'm tired of their browser hanging, taking minutes of compute time to exit, and unable to close some obnoxious hostile pages.
When the Mozilla crowd gets a bug-free 4.x out the door, with all reported bugs fixed, then they can talk about later versions.
(I'm underwhelmed with the Mozilla crowd. They come across as a bunch of Javascript hackers. There was a problem with duplicate entries in the internal SQLite databases, and I suggested making the fields a unique key, so the database would reject bogus updates. But that would have caused broken code to abort. Instead, they bolted on a Javascript hack to try to clean up the mess.)
Linux still has the antiquated "user, group, everyone" security model from the 1970s. By now, we know that outside data can't be given all the privileges of the user. But Linux's legacy security model is so deeply embedded in the UNIX/Linux world that it's almost impossible to get beyond that.
Yes, there's SELinux. But there isn't a whole distribution with a full range of applications which can run under a mandatory security model.