This is fascinating. As someone who's worked with automatic theorem proving and proof of correctness techniques, I'd never thought of using them in this way.
What they're doing works like a proof of correctness system in reverse. They difference executables before and after the patch (which in itself is impressive), then, having isolated the patch, analyze it automatically. Security patches usually consist of adding a test which constrains the valid inputs at some point. So they use a symbolic decision procedure, which is part of a theorem prover, to work back through the code and automatically derive a set of inputs that would be caught by the new test.
This is more than just an attack on Windows Update. It's true automated exploit generation.
This is potentially applicable to any security-critical code that changes over time. One could, for example, have something that watched check-ins to the Linux kernel tree and developed new exploits to current stable releases from them.
Cute. Read the actual paper from IBM, not the blogodreck. The same attack works on IE and Firefox, on XP and Vista. An attack for Firefox on Linux is probably possible.
This is easy to fix; it's a one line bug in Flash. But it will take years to replace all the bad versions of Flash out there.
If this is for real, there's a whole new class of electronic devices waiting to be developed.
The Zeno effect has been observed experimentally, but only down near absolute zero. If it can be observed at room temperature, it could be useful.
At SiteTruth, we divide certificates into three categories, rather than the usual two:
"Extended Validation" certificates.
"Organization validated" certificates, which must have an L (location) field and must not have a domain name in the O (organization) field.
"Domain control only validated" or "Quick SSL" certificates, which say nothing about who's at the other end of the connection.
Browsers normally lump category 2 and 3 together. This is not a good thing.
Category 3 certs, the "Instant SSL" certs, have no value in identifying the business. A category 1 or 2 cert increases the site's SiteTruth legitimacy rating, since we have a third party which has vouched for the ownership of the site. A category 3 cert does not.
Browsers should make this distinction. You never want to enter a credit card number into a site that only has a class 3 cert. You have no idea where your money is going.
Foster-Miller has several robot product lines. The TALON line is widely deployed,
mostly for bomb squad use and such. There are a number of variants. The SWORDS robots appear to have been a low-volume prototype design.
The new MAARS line is more modular, with field-replaceable options ranging from a manipulator arm to a gun turret, and different configurations can replace the TALON and SWORDS units.
If you look at the pictures, SWORDS looks like a prototype, with stuff bolted onto a TALON track base, but MAARS looks like a finished product.
They're all rather dumb teleoperators.
Foster-Miller also makes the Big Dog legged robot for Boston Dynamics. That thing actually has some autonomy, using Boston Dynamics software.
The use of "social networking" data for search has been discussed before in the search technology community, where it's not well thought of. "Inertia" in search, where your search history affects your later results, turns out to be a pain. Search becomes nonrepeatable, both for the individual and for others. This adds more hassle than the gain provided by "inertia".
Reading both the article and the interview with the Google VP, it's clear that the article exaggerates Google's interest in this area.
Social networking data is taken seriously on the advertising side, where using social networking data for ad selection is already being done by Myspace and their ilk. Amazon and Netflix already have rather good systems for deciding what to recommend to their customers. That's where this really works, where the seller has a big product selection and the user is already prepped to buy something. Myspace isn't doing as well, but then, as we've pointed out before, their advertisers are mostly bottom feeders. Ad rates on Myspace are very low, and it shows.
A key question is who controls the use of the social networking data for ad selection. Not the user, of course; the disagreement is between the social networking sites and the search engines. Look for a battle in that area, perhaps followed by mergers.
There are now 35535 entries in the Netflix competition. If they all used roughly the same algorithm, with some randomness in the tuning variables, we'd expect to see results about like what we've seen. I think we're looking at noise here.
The same phenomenon shows up with mutual funds. Some outperform the market, some don't, but prior year results are not good predictors of future results.
The idea shortage in Hollywood continues. As Harper's pointed out, more than half of the top-grossing movies of 2007 were sequels where N > 2.
Cartoon (not comic) to live action translation hasn't been that great. "Boris and Natasha: The Movie" (1992) was something of a flop, as was "Dudly Do-Right" (1999). A third try, "The Adventures of Rocky and Bullwinkle" (2000) was a dud, too, although it was at least funny. "Underdog" (2007) is the most recent dud.
"The Flintstones" (1994) was one of the few successes. "Casper" (1995) was a success, mainly because CG animation had become good enough to be used convincingly with live actors. Those had the novelty of a cartoon as live action. But that's been done now, and the novelty has worn off.
Comic books have been a more fruitful source of material, enough so that Marvel now has its own movie studio.
Redirection services have to be vigilant about phishing, and should be tied into the major phishing databases so they can block misuses of their service very rapidly. Otherwise, their URLs are likely to be identified as spam.
Phishing sites like to create URLs that will get through spam filters. So any mechanism which allows a phishing site to create a URL within a well-known domain is an attack vector.
What's striking is how few sites are on that list. It's simply a list of second level domains that are in both PhishTank and Open Directory. When the list was first created, I was expecting to find hundreds or even thousands of entries. But no. Initially, there were around 150, and we've been able to bring that down.
Does the "Pickering Institute" even exist? Their home page is a WordPress blog. They have no contact information other than an e-mail address.
Their domain registration has an address of "2 Cityplace Drive, Suite 200, St. Louis, MO",
which is also the address of Bin95.com, which does industrial equipment maintenance training.
Public embarrassment can be useful. We publish a list of major domains being exploited by active phishing scams. These are major domains where an attacker has found a security hole allowing them to exploit the site for phishing purposes. There are 65 sites on the list. There used to be about 140, but by nagging and publicity, we've been able to get most big-name sites to tighten up. Now and then some big site makes the list, but it often disappears within hours as the hole is plugged.
So it actually is possible to get big companies to tighten up security, if you do it right.
Amusingly, for many years Monster stayed away from the types of cables where quality matters, like VGA cables. VGA cables have a high-bandwidth analog signal, and long (10m or so) VGA cables have transmission-line type problems, where mismatches or crosstalk result in ghosting or blur at the monitor. For short tables, it's not a big deal, but as length increases, it matters. There are lots of crap VGA cables out there. Still, above $15 for 10m, you're overpaying.
HDMI cables have to carry 340MHz, so they're transmission lines. There's a certification process, and if the cable passed it, it should be OK. There are phony HDMI cables out there that don't pass the spec, but all certified cables should work equally well.
There's something to be said for gold-plated connectors, especially for something that's frequently unplugged, but the cost of the gold is trivial.
The basic requirements that must be met for tying to be per se illegal are as follows:
There must be two separate products or services.
There must be a sale or an agreement to sell one product (or service) on the condition that the buyer purchase another product or service (or the buyer agrees not to purchase the product or service from another supplier).
The seller must have sufficient economic power with respect to the tying product to appreciably restrain free competition in the market for the tied product.
The tying arrangement must affect a "not insubstantial" amount of commerce.
Apple would have to try to enforce their EULA in court against an antitrust claim that their EULA is an illegal tying arrangement, which, on its face, it is.
Apple was able to put the previous generation of clone-makers out of business because some key portions of the original MacOS were in ROM, shipped with the machine. So they could make copyright arguments against cloning the Mac ROMs. But for today's machines, the OS isn't built onto the motherboard, so there's no copyright claim.
IBM lost this issue a long time ago, when Compaq made the first PC clone. That's why there's a PC industry.
Apple could put DRM hardware in future Macs and encrypt future OS releases, like a game console. Not having done that, they're stuck.
They should test physics systems with spheres on irregular ground, with uneven (coefficient of kinetic friction), and changing wind.
I always liked to test on the hard cases (see my web site) but, in fact, what game developers and animators want is not realistic physics. The game people want "infinitely destructible environments", where you can blow up anything. And, of course, eye candy. The animation people want physics they can direct, where the starting and ending positions are specified but what happens in the middle might be automated. So most character motion in video games is spliced-together motion capture bits, hand animation, and a bit of physics now and then.
Most game physics engines today cheat on the hard cases, but by now, they usually cheat
in ways that don't do something blatantly nonphysical.
My original goal, back in 1996 when I was working on this, was to get to a two-person fighting game with real martial arts. The physics engine we had was good enough, although in the era of 200MHz CPUs, a bit slow. But
a control system that can do a judo throw against an uncooperative opponent is still out of reach. Worse, controlling a physically realistic martial arts character through a game pad is just hopeless. Play would be like a typical day at the dojo: "You lost your center. Your left foot should have been further forward when you tried that throw." "Yes, sensi". Work out three times a week for a year, and you get to be halfway decent. As a game, that sucks.
So I created the "ragdoll falling downstairs" cliche as a demo, licensed the technology to a middleware provider, and went on to other things.
The Golden Gate Bridge has emergency phones, which used to be marked "Roadside Assistance".
For a while, they were marked "Roadside Assistance and Psychological Counseling".
("And how does having a flat tire make you feel?")
The SWORDS robot isn't autonomous; it has the autonomy level of an R/C car.
Something like this happened in the 1980s with the Sgt. York Air Division Air Defense Gun, which was an automated antiaircraft weapon. During a demo, it pointed its guns at the reviewing stand. The project was canceled. (Arguably, it was canceled for other reasons. The DIVAD was built as a response to the USSR's ZSU, their radar-directed anti-aircraft gun. This class of weapon is useful if you're being attacked by a squadron of helicopters, but it can't hit fast-movers like fighter-bombers. Only the US attacks with large numbers of helicopters, because you have to have both a big budget and air superiority to do that. So it wasn't something the U.S. Army needed to defend against. A few guys with Stingers could stop any small scale helicopter assaults.)
The point, though, is that the U.S. military has a very low tolerance for this class of mistake, and sizable projects have been canceled for it.
This was the very first deployment of an armed ground combat robot to a war zone. Three units went to Iraq. The cancellation of the project is a sizable blow to the future of armed combat robots.
Why do they need "power conditioning" at all? Every unit specified, with the possible exception of the "Receiver"'s amplifier section, has a switching power supply, and should be robust in the face of spikes and sags.
It's a Roland the Plogger story, so you know there has to be something wrong with it.
There have been a few previous attempts at 2D treadmills. They're usually just a big endless chain of belt-type treadmills, like this one. The problem is that the things are big, heavy, expensive, and may have pinch points.
Here's a
paper which lists most of the previous omnidirectional treadmills. The Darken Omnidirectional Treadmill (1997) was a belt made of rollers. The Torus Treadmill was an array of 12 small treadmills on an endless chain. This new treadmill looks like a bigger version of the Torus Treadmill.
The Sarcos Treadport is a more aggressive approach to the problem. It's only a 1D treadmill, but it's on a tilt base, so it can simulate hills. The user is in a suspended harness which can yank the user around to simulate collision with an obstacle. (Now there's force feedback.)
So this isn't that novel; it's just a bit bigger. Typical Roland the Plogger; press release goes in, blogspam comes out.
Slashdot Mirror
Blockbuster's user agreement includes a wavier of your rights under the Video Privacy Protection Act. That's why I don't shop there.
This is fascinating. As someone who's worked with automatic theorem proving and proof of correctness techniques, I'd never thought of using them in this way.
What they're doing works like a proof of correctness system in reverse. They difference executables before and after the patch (which in itself is impressive), then, having isolated the patch, analyze it automatically. Security patches usually consist of adding a test which constrains the valid inputs at some point. So they use a symbolic decision procedure, which is part of a theorem prover, to work back through the code and automatically derive a set of inputs that would be caught by the new test.
This is more than just an attack on Windows Update. It's true automated exploit generation.
This is potentially applicable to any security-critical code that changes over time. One could, for example, have something that watched check-ins to the Linux kernel tree and developed new exploits to current stable releases from them.
Cute. Read the actual paper from IBM, not the blogodreck. The same attack works on IE and Firefox, on XP and Vista. An attack for Firefox on Linux is probably possible.
This is easy to fix; it's a one line bug in Flash. But it will take years to replace all the bad versions of Flash out there.
Wow. I want to see the full paper.
If this is for real, there's a whole new class of electronic devices waiting to be developed. The Zeno effect has been observed experimentally, but only down near absolute zero. If it can be observed at room temperature, it could be useful.
(On a stylistic note, would someone at Slashdot please fix the bogus CSS that makes ordered lists ("ol") appear without numbers? Thank you.)
At SiteTruth, we divide certificates into three categories, rather than the usual two:
Browsers normally lump category 2 and 3 together. This is not a good thing.
Category 3 certs, the "Instant SSL" certs, have no value in identifying the business. A category 1 or 2 cert increases the site's SiteTruth legitimacy rating, since we have a third party which has vouched for the ownership of the site. A category 3 cert does not.
Browsers should make this distinction. You never want to enter a credit card number into a site that only has a class 3 cert. You have no idea where your money is going.
2GB? Really. This business of storing full resolution images of pages is silly.
Sure, you can buy 2GB of RAM, but perhaps you might like to use it for something useful while the web browser is running.
We're approaching the point where web browsers won't run on a 32-bit machine.
Foster-Miller has several robot product lines. The TALON line is widely deployed, mostly for bomb squad use and such. There are a number of variants. The SWORDS robots appear to have been a low-volume prototype design. The new MAARS line is more modular, with field-replaceable options ranging from a manipulator arm to a gun turret, and different configurations can replace the TALON and SWORDS units. If you look at the pictures, SWORDS looks like a prototype, with stuff bolted onto a TALON track base, but MAARS looks like a finished product.
They're all rather dumb teleoperators.
Foster-Miller also makes the Big Dog legged robot for Boston Dynamics. That thing actually has some autonomy, using Boston Dynamics software.
The use of "social networking" data for search has been discussed before in the search technology community, where it's not well thought of. "Inertia" in search, where your search history affects your later results, turns out to be a pain. Search becomes nonrepeatable, both for the individual and for others. This adds more hassle than the gain provided by "inertia".
Reading both the article and the interview with the Google VP, it's clear that the article exaggerates Google's interest in this area.
Social networking data is taken seriously on the advertising side, where using social networking data for ad selection is already being done by Myspace and their ilk. Amazon and Netflix already have rather good systems for deciding what to recommend to their customers. That's where this really works, where the seller has a big product selection and the user is already prepped to buy something. Myspace isn't doing as well, but then, as we've pointed out before, their advertisers are mostly bottom feeders. Ad rates on Myspace are very low, and it shows.
A key question is who controls the use of the social networking data for ad selection. Not the user, of course; the disagreement is between the social networking sites and the search engines. Look for a battle in that area, perhaps followed by mergers.
There are now 35535 entries in the Netflix competition. If they all used roughly the same algorithm, with some randomness in the tuning variables, we'd expect to see results about like what we've seen. I think we're looking at noise here.
The same phenomenon shows up with mutual funds. Some outperform the market, some don't, but prior year results are not good predictors of future results.
The idea shortage in Hollywood continues. As Harper's pointed out, more than half of the top-grossing movies of 2007 were sequels where N > 2.
Cartoon (not comic) to live action translation hasn't been that great. "Boris and Natasha: The Movie" (1992) was something of a flop, as was "Dudly Do-Right" (1999). A third try, "The Adventures of Rocky and Bullwinkle" (2000) was a dud, too, although it was at least funny. "Underdog" (2007) is the most recent dud.
"The Flintstones" (1994) was one of the few successes. "Casper" (1995) was a success, mainly because CG animation had become good enough to be used convincingly with live actors. Those had the novelty of a cartoon as live action. But that's been done now, and the novelty has worn off.
Comic books have been a more fruitful source of material, enough so that Marvel now has its own movie studio.
I noticed tinyurl.com on your list...
Redirection services have to be vigilant about phishing, and should be tied into the major phishing databases so they can block misuses of their service very rapidly. Otherwise, their URLs are likely to be identified as spam.
Phishing sites like to create URLs that will get through spam filters. So any mechanism which allows a phishing site to create a URL within a well-known domain is an attack vector. What's striking is how few sites are on that list. It's simply a list of second level domains that are in both PhishTank and Open Directory. When the list was first created, I was expecting to find hundreds or even thousands of entries. But no. Initially, there were around 150, and we've been able to bring that down.
Does the "Pickering Institute" even exist? Their home page is a WordPress blog. They have no contact information other than an e-mail address.
Their domain registration has an address of "2 Cityplace Drive, Suite 200, St. Louis, MO", which is also the address of Bin95.com, which does industrial equipment maintenance training.
Public embarrassment can be useful. We publish a list of major domains being exploited by active phishing scams. These are major domains where an attacker has found a security hole allowing them to exploit the site for phishing purposes. There are 65 sites on the list. There used to be about 140, but by nagging and publicity, we've been able to get most big-name sites to tighten up. Now and then some big site makes the list, but it often disappears within hours as the hole is plugged.
So it actually is possible to get big companies to tighten up security, if you do it right.
Amusingly, for many years Monster stayed away from the types of cables where quality matters, like VGA cables. VGA cables have a high-bandwidth analog signal, and long (10m or so) VGA cables have transmission-line type problems, where mismatches or crosstalk result in ghosting or blur at the monitor. For short tables, it's not a big deal, but as length increases, it matters. There are lots of crap VGA cables out there. Still, above $15 for 10m, you're overpaying.
HDMI cables have to carry 340MHz, so they're transmission lines. There's a certification process, and if the cable passed it, it should be OK. There are phony HDMI cables out there that don't pass the spec, but all certified cables should work equally well.
There's something to be said for gold-plated connectors, especially for something that's frequently unplugged, but the cost of the gold is trivial.
Apple is going to lose this one. It's a illegal tying arrangement under the Clayton Act:
The basic requirements that must be met for tying to be per se illegal are as follows:
Apple would have to try to enforce their EULA in court against an antitrust claim that their EULA is an illegal tying arrangement, which, on its face, it is.
Apple was able to put the previous generation of clone-makers out of business because some key portions of the original MacOS were in ROM, shipped with the machine. So they could make copyright arguments against cloning the Mac ROMs. But for today's machines, the OS isn't built onto the motherboard, so there's no copyright claim.
IBM lost this issue a long time ago, when Compaq made the first PC clone. That's why there's a PC industry.
Apple could put DRM hardware in future Macs and encrypt future OS releases, like a game console. Not having done that, they're stuck.
They should test physics systems with spheres on irregular ground, with uneven (coefficient of kinetic friction), and changing wind.
I always liked to test on the hard cases (see my web site) but, in fact, what game developers and animators want is not realistic physics. The game people want "infinitely destructible environments", where you can blow up anything. And, of course, eye candy. The animation people want physics they can direct, where the starting and ending positions are specified but what happens in the middle might be automated. So most character motion in video games is spliced-together motion capture bits, hand animation, and a bit of physics now and then.
Most game physics engines today cheat on the hard cases, but by now, they usually cheat in ways that don't do something blatantly nonphysical.
My original goal, back in 1996 when I was working on this, was to get to a two-person fighting game with real martial arts. The physics engine we had was good enough, although in the era of 200MHz CPUs, a bit slow. But a control system that can do a judo throw against an uncooperative opponent is still out of reach. Worse, controlling a physically realistic martial arts character through a game pad is just hopeless. Play would be like a typical day at the dojo: "You lost your center. Your left foot should have been further forward when you tried that throw." "Yes, sensi". Work out three times a week for a year, and you get to be halfway decent. As a game, that sucks.
So I created the "ragdoll falling downstairs" cliche as a demo, licensed the technology to a middleware provider, and went on to other things.
This is just a particle system. Smoke, snow, rain, and maybe water look better, but it apparently has no influence on gameplay.
explain to me what issues are there for which in 2008 we still have to resort to sub-sonic air flights?
Because when you go supersonic, fuel consumption triples.
The Golden Gate Bridge has emergency phones, which used to be marked "Roadside Assistance". For a while, they were marked "Roadside Assistance and Psychological Counseling". ("And how does having a flat tire make you feel?")
Now they're marked Emergency Phone and Crisis Counseling. The suicide rate varies from year to year, but nothing so far has affected it much one way or the other.
Shuttle PCs have had a heat-pipe and heat exchanger liquid cooling system for years. This made possible their little "breadbox" systems.
SQL Maestro is an administrative tool, not a report generator.
PHP Generator for MySQL is free and useful for generating simple database-driven web sites.
Admittedly, the MySQL Query Browser is clunky, but at least it finally works. For several releases, it was badly broken.
Stupid article. Real problem.
The SWORDS robot isn't autonomous; it has the autonomy level of an R/C car.
Something like this happened in the 1980s with the Sgt. York Air Division Air Defense Gun, which was an automated antiaircraft weapon. During a demo, it pointed its guns at the reviewing stand. The project was canceled. (Arguably, it was canceled for other reasons. The DIVAD was built as a response to the USSR's ZSU, their radar-directed anti-aircraft gun. This class of weapon is useful if you're being attacked by a squadron of helicopters, but it can't hit fast-movers like fighter-bombers. Only the US attacks with large numbers of helicopters, because you have to have both a big budget and air superiority to do that. So it wasn't something the U.S. Army needed to defend against. A few guys with Stingers could stop any small scale helicopter assaults.)
The point, though, is that the U.S. military has a very low tolerance for this class of mistake, and sizable projects have been canceled for it. This was the very first deployment of an armed ground combat robot to a war zone. Three units went to Iraq. The cancellation of the project is a sizable blow to the future of armed combat robots.
Why do they need "power conditioning" at all? Every unit specified, with the possible exception of the "Receiver"'s amplifier section, has a switching power supply, and should be robust in the face of spikes and sags.
It's a Roland the Plogger story, so you know there has to be something wrong with it.
There have been a few previous attempts at 2D treadmills. They're usually just a big endless chain of belt-type treadmills, like this one. The problem is that the things are big, heavy, expensive, and may have pinch points.
Here's a paper which lists most of the previous omnidirectional treadmills. The Darken Omnidirectional Treadmill (1997) was a belt made of rollers. The Torus Treadmill was an array of 12 small treadmills on an endless chain. This new treadmill looks like a bigger version of the Torus Treadmill.
The Sarcos Treadport is a more aggressive approach to the problem. It's only a 1D treadmill, but it's on a tilt base, so it can simulate hills. The user is in a suspended harness which can yank the user around to simulate collision with an obstacle. (Now there's force feedback.)
So this isn't that novel; it's just a bit bigger. Typical Roland the Plogger; press release goes in, blogspam comes out.